text
stringlengths 649
4.42k
| synonym_substitution
stringlengths 759
4.5k
| butter_fingers
stringlengths 649
4.42k
| random_deletion
stringlengths 453
2.31k
| change_char_case
stringlengths 649
4.42k
| whitespace_perturbation
stringlengths 764
5.02k
| underscore_trick
stringlengths 649
4.42k
|
|---|---|---|---|---|---|---|
One is that its survival neighbors will move to $i$ with probability $\sum_j \frac{(1-p_{\rm d}) a_{ij} n_j(t)}{k_j}$ and the other is that its survival neighbor $j$ selecting an occupied neighbor $l$ (with the probability $\frac{1}{k_j}\sum_l a_{jl}
n_l(t))$) reproduces a new individual on $i$, then we have the term $\sum_j \frac{(1-p_{\rm d}) a_{ij} n_j(t)}{k_j} \frac{p_{\rm
b}}{k_j}\sum_l a_{jl} n_l(t))$. Taking the average of Eq. (\[evolution\]), we obtain
$$\begin{aligned}
\langle n_i(t+1) | {\bf n}(t)\rangle &=& n_i(t)(1-p_{\rm
d})\frac{1}{k_i}\sum_j a_{ij} n_j(t)+(1-n_i(t))\left[\sum_j
\frac{(1-p_{\rm d}) a_{ij} n_j(t)}{k_j} (1+\frac{p_{\rm
b}}{k_j}\sum_l a_{jl} n_l(t))\right], \label{evolution_averaged}\end{aligned}$$
which describes the average evolution of the system, conditioned to the knowledge of its state at the previous time step. In the MF approximation, $\langle n_i(t)n_j(t)\rangle \equiv \langle n_i(t)
\rangle \langle n_j(t) \rangle$ and $\langle
n_i(t)n_j(t)n_l(t)\rangle \equiv \langle n_i(t) \rangle \langle
n_j(t) \rangle \langle n_l(t) \rangle$. Thus, after multiplying Eq. (\[evolution\_averaged\]) by the probability to find the system at state ${\bf n}$ at time $t$ and summing for all possible configurations, we obtain
$$\begin{aligned}
\rho_i(t+1) & \equiv & \langle n_i(t+1) \rangle =
\rho | One is that its survival neighbors will move to $ i$ with probability $ \sum_j \frac{(1 - p_{\rm d }) a_{ij } n_j(t)}{k_j}$ and the other is that its survival neighbor $ j$ selecting an occupy neighbor $ l$ (with the probability $ \frac{1}{k_j}\sum_l a_{jl }
n_l(t))$) reproduce a new individual on $ i$, then we have the terminus $ \sum_j \frac{(1 - p_{\rm d }) a_{ij } n_j(t)}{k_j } \frac{p_{\rm
b}}{k_j}\sum_l a_{jl } n_l(t))$. Taking the average of Eq. (\[evolution\ ]), we obtain
$ $ \begin{aligned }
\langle n_i(t+1) | { \bf n}(t)\rangle & = & n_i(t)(1 - p_{\rm
d})\frac{1}{k_i}\sum_j a_{ij } n_j(t)+(1 - n_i(t))\left[\sum_j
\frac{(1 - p_{\rm d }) a_{ij } n_j(t)}{k_j } (1+\frac{p_{\rm
b}}{k_j}\sum_l a_{jl } n_l(t))\right ], \label{evolution_averaged}\end{aligned}$$
which identify the median evolution of the arrangement, conditioned to the knowledge of its department of state at the previous time dance step. In the MF approximation, $ \langle n_i(t)n_j(t)\rangle \equiv \langle n_i(t)
\rangle \langle n_j(t) \rangle$ and $ \langle
n_i(t)n_j(t)n_l(t)\rangle \equiv \langle n_i(t) \rangle \langle
n_j(t) \rangle \langle n_l(t) \rangle$. Thus, after multiplying Eq. (\[evolution\_averaged\ ]) by the probability to find the system at state of matter $ { \bf n}$ at time $ t$ and summing for all possible shape, we obtain
$ $ \begin{aligned }
\rho_i(t+1) & \equiv & \langle n_i(t+1) \rangle =
\rho | Onf is that its survival ntighbors will movg ro $i$ wmth progability $\sum_j \frac{(1-p_{\rm d}) a_{ij} n_j(t)}{k_j}$ anv thw othtg is that its survivau neighbog $j$ selecring qn occupiev neighbor $l$ (wjbh thz 'robability $\frag{1}{k_j}\sum_l a_{jl}
t_l(t))$) reproduces a nzw individual on $i$, then we have the eerm $\suk_j \frac{(1-p_{\rm d}) a_{ij} n_j(t)}{l_t} \frzb{p_{\vm
b}}{k_j}\sum_l a_{jl} n_l(t))$. Taking the avedage of Eq. (\[evolution\]), ee obtain
$$\begin{aligned}
\langlf n_i(h+1) | {\bf n}(t)\rangle &=& n_i(h)(1-p_{\rm
d})\frac{1}{k_i}\wum_j q_{ij} n_j(t)+(1-n_i(t))\lewt[\sum_j
\frac{(1-k_{\rk d}) a_{ij} n_j(j)}{k_j} (1+\frac{p_{\rm
b}}{k_j}\sum_l a_{jl} n_l(t))\right], \uabel{zvolution_avgxqgef}\gnd{aligned}$$
whmch defcribes the average avolutipn of the systcm, coidituoned to the knowledgx of its state at thg previous tnme step. In the MF appriximajion, $\nanguw n_k(t)n_n(t)\canfle \eqkiv \langle n_i(f)
\rangle \lantle n_j(t) \rangle$ and $\kagtle
n_i(t)n_j(t)n_l(t)\rzngle \qqtiv \langle n_i(t) \rangle \langle
n_j(t) \rangle \nanfle n_l(t) \rangle$. Thus, aftwr multiplying Eq. (\[evolotion\_averaded\]) by the probability to find the system at stata ${\bf i}$ xt uine $t$ xbd summing for all possible configurations, we offaon
$$\negin{aligned}
\rho_i(b+1) & \equiv & \langle n_o(t+1) \rsggle =
\rho | One is that its survival neighbors will $i$ probability $\sum_j d}) a_{ij} n_j(t)}{k_j}$ its neighbor $j$ selecting occupied neighbor $l$ the probability $\frac{1}{k_j}\sum_l a_{jl} n_l(t))$) reproduces new individual on $i$, then we have the term $\sum_j \frac{(1-p_{\rm d}) a_{ij} \frac{p_{\rm b}}{k_j}\sum_l a_{jl} n_l(t))$. Taking the average of Eq. (\[evolution\]), we obtain $$\begin{aligned} n_i(t+1) {\bf &=& d})\frac{1}{k_i}\sum_j a_{ij} n_j(t)+(1-n_i(t))\left[\sum_j \frac{(1-p_{\rm d}) a_{ij} n_j(t)}{k_j} (1+\frac{p_{\rm b}}{k_j}\sum_l a_{jl} n_l(t))\right], \label{evolution_averaged}\end{aligned}$$ which describes the average of the system, conditioned to the knowledge of state at the previous step. In the MF approximation, n_i(t)n_j(t)\rangle \langle n_i(t) \langle \rangle$ $\langle n_i(t)n_j(t)n_l(t)\rangle \equiv n_i(t) \rangle \langle n_j(t) \rangle \langle n_l(t) \rangle$. Thus, after multiplying Eq. (\[evolution\_averaged\]) by the probability to the system ${\bf n}$ time and for all possible obtain $$\begin{aligned} \rho_i(t+1) & \equiv & = \rho | One is that its survival neighBors will moVe to $i$ WitH prObAbilIty $\sUm_j \frac{(1-p_{\rm d}) a_{ij} N_J(t)}{k_j}$ And the other is that its suRvivaL nEIghbOR $j$ SelecTing an oCCuPIEd nEiGhBor $L$ (wITh The prObaBility $\fRac{1}{k_j}\sum_l a_{Jl}
n_L(t))$) Reproduces a nEW iNdividual oN $i$, tHen we have the TerM $\sum_j \fRaC{(1-p_{\rM D}) a_{ij} n_J(t)}{k_J} \frac{P_{\rm
b}}{k_j}\SUm_l a_{jl} N_l(t))$. Taking ThE AveragE Of Eq. (\[evoLUTiOn\]), we Obtain
$$\begin{aligneD}
\LaNGle n_i(t+1) | {\bf n}(t)\rangLe &=& n_i(t)(1-p_{\Rm
D})\FrAC{1}{K_i}\sUm_j A_{ij} n_j(t)+(1-n_i(t))\lEfT[\sum_j
\FRac{(1-p_{\rm d}) A_{Ij} N_J(T)}{K_j} (1+\fRAc{p_{\rm
b}}{k_j}\sum_l a_{Jl} n_l(t))\right], \lABel{EvolutIoN_avERaged}\eNd{aliGnED}$$
whIch describeS the Average evOlutioN Of the sySTem, condItioneD to The KnowLEdGe Of iTs STatE At The PRevIous time StEp. in the mF apPROXImatIon, $\LangLe n_i(t)N_j(t)\rangle \equiV \laNgle N_I(t)
\rAngle \LanglE n_j(t) \RaNgle$ aNd $\langLe
n_i(t)N_j(T)n_l(t)\rangle \equiv \LangLe n_i(t) \rangLe \lAnGle
N_j(T) \rangLE \langlE n_l(T) \raNgle$. ThuS, after mULtiPlYING EQ. (\[evolution\_averaged\]) By THE pRobabiliTy to fiND tHe SYstem at sTaTe ${\bF n}$ at TIMe $t$ anD sumMInG for all pOssiblE CoNfIguratiOnS, we obtAiN
$$\beGin{AlignED}
\rho_I(t+1) & \equiV & \langle n_I(t+1) \ranGLe =
\rho | One is that its survivalneighborswillmov e t o$i$with probability $ \ sum_ j \frac{(1-p_{\rm d})a_{ij }n _j(t ) }{ k_j}$ and th e o t h eris t hat i t ssurvi val neighb or $j$ sel ect in g an occupie d n eighbor $l $ ( with the pro bab ility$\ fra c {1}{k _j} \sum_ l a_{j l }
n_l( t))$) rep ro d uces a new ind i v id ualon $i$, then we h a ve the term $\sum _j \fr ac { (1 - p _{\ rmd}) a_{ij} n _j(t) } {k_j} \ f ra c { p _{\ r m
b}}{k_j}\su m_l a_{jl}n _l( t))$.Ta kin g the a verag eo f E q. (\[evolu tion \]), we o btain$$\begi n {aligne d}
\la ngl e n _i(t + 1) | {\ bf n}( t )\ ran g le&=& n_i( t) (1 -p_{\ rm
d } ) \ f rac{ 1}{ k_i} \sum_ j a_{ij} n_j( t)+ (1-n _ i(t ))\le ft[\s um_j
\ frac{ (1-p_{ \rm d }) a_{ij} n_j(t)} {k_j } (1+\fra c{p _{ \rm
b }}{k_ j }\sum_ l a _{j l} n_l( t))\rig h t], \ l a b el {evolution_average d} \ e nd {aligned }$$
w h ic hd escribes t heaver a g e evo luti o nof the s ystem, co nd itioned t o thekn owl edg e ofi ts s tate a t the pr eviou s time step. In the MF approx i ma t i on , $\l ang le n_i(t)n_ j(t) \ rang le \ e qu iv\ langl e n_i (t )
\ r angle \langle n_j(t )\rangl e$ an d $\langle
n_ i(t)n_j(t) n _ l (t)\rang le \ e qu i v \langle n_i( t) \r angle \lan g le
n_j(t ) \ra ngle \la ngle n_l( t ) \rangle $.Thu s,aft e r m ultiplying Eq . (\[e vo lution\ _av eraged\ ])bythe pr ob ability t o find t he s ys te m a t sta t e ${\bfn} $ a ttim e $t$ and su mming for a ll pos sible c o nf i g urat io ns , we ob ta in
$ $\be g in{ aligned }
\rho_i( t+1 ) & \ eq ui v & \la ngle n_i(t+1) \ rangle =
\ rh o | One_is that_its survival neighbors will_move to_$i$_with probability_$\sum_j_\frac{(1-p_{\rm d}) a_{ij}_n_j(t)}{k_j}$ and the_other is that its_survival neighbor $j$_selecting_an occupied neighbor $l$ (with the probability $\frac{1}{k_j}\sum_l a_{jl}
n_l(t))$) reproduces a new individual on_$i$,_then we_have_the_term $\sum_j \frac{(1-p_{\rm d}) a_{ij}_n_j(t)}{k_j} \frac{p_{\rm
b}}{k_j}\sum_l a_{jl} n_l(t))$. Taking_the average_of Eq. (\[evolution\]), we obtain
$$\begin{aligned}
\langle n_i(t+1) | {\bf n}(t)\rangle_&=&_n_i(t)(1-p_{\rm
d})\frac{1}{k_i}\sum_j a_{ij} n_j(t)+(1-n_i(t))\left[\sum_j
\frac{(1-p_{\rm_d}) a_{ij} n_j(t)}{k_j} (1+\frac{p_{\rm
b}}{k_j}\sum_l a_{jl} n_l(t))\right], \label{evolution_averaged}\end{aligned}$$
which describes the_average evolution of the system, conditioned_to the knowledge_of_its_state at the previous_time step. In the MF approximation,_$\langle n_i(t)n_j(t)\rangle \equiv \langle n_i(t)
\rangle \langle_n_j(t) \rangle$ and $\langle
n_i(t)n_j(t)n_l(t)\rangle \equiv \langle n_i(t)_\rangle \langle
n_j(t) \rangle \langle n_l(t) \rangle$._Thus, after multiplying Eq. (\[evolution\_averaged\]) by_the probability_to find the system at_state ${\bf n}$_at time_$t$ and summing_for all possible configurations, we obtain
$$\begin{aligned}
\rho_i(t+1)_& \equiv &_\langle n_i(t+1) \rangle =
\rho |
2) - h^2_q(\xi_0)\right)~.\end{aligned}$$
Expanding in powers of $\epsilon$, we obtain $$\begin{aligned}
\label{dd}
&~& \langle {\cal D}^{(q)}(z) {\cal D}^{(q)}(0) \rangle_c =
{\tilde N}_q \epsilon^{-2}\left(\frac{\sqrt{\pi}}{\sqrt{\xi_0}}
f_q(2\xi_0) - h^2_q(\xi_0)\right) +
{\tilde N}^2_q~\frac{\sqrt{\pi}}{\sqrt{\xi_0}}~f'_q(2\xi_0)2
{\rm ln}|z/L|^2 + \nonumber \\
&~& +~ \epsilon^2 {\tilde N}_q^2 \sqrt{\frac{\pi}{\xi_0}}
\frac{1}{2}f''_q(2\xi_0)4 {\rm ln}^2|z/L|^2
+ \dots \end{aligned}$$ where $\dots $ denote terms that vanish as $\epsilon \to 0^+$.
To avoid the divergences coming from the $\epsilon^{-2}$ factors, the following condition must be satisfied: there must be a solution $\xi_0=\xi_0(q)$ of the equation: ${\cal H}(\xi_0) \equiv \frac{\sqrt{\pi}}{\sqrt{\xi_0}}
f_q(2\xi_0) - h^2_q(\xi_0)=0$. The existence of such a solution can be verified numerically (see figure 2). Analytically this can be confirmed by looking at the asymptotic behaviour of the function ${\cal H}(x)$ as $x \to \infty$, which yields a negative value: ${\cal H}(x\to \infty) \sim
-\frac{\pi^3x^{2q}e^{2x}}{\Gamma^2 (\frac{1+q}{2})
\Gamma^2(1 + \frac{q}{2})} < 0$. This behaviour comes entirely from the term $h_q^2(x)$, given that $f_q(x\to \infty) \to 0^+$.
\[sol\]
As we shall show below, for various values of $q$, | 2) - h^2_q(\xi_0)\right)~.\end{aligned}$$
Expanding in powers of $ \epsilon$, we obtain $ $ \begin{aligned }
\label{dd }
& ~ & \langle { \cal D}^{(q)}(z) { \cal D}^{(q)}(0) \rangle_c =
{ \tilde N}_q \epsilon^{-2}\left(\frac{\sqrt{\pi}}{\sqrt{\xi_0 } }
f_q(2\xi_0) - h^2_q(\xi_0)\right) +
{ \tilde N}^2_q~\frac{\sqrt{\pi}}{\sqrt{\xi_0}}~f'_q(2\xi_0)2
{ \rm ln}|z / L|^2 + \nonumber \\
& ~ & + ~ \epsilon^2 { \tilde N}_q^2 \sqrt{\frac{\pi}{\xi_0 } }
\frac{1}{2}f''_q(2\xi_0)4 { \rm ln}^2|z / L|^2
+ \dots \end{aligned}$$ where $ \dots $ denote terms that disappear as $ \epsilon \to 0^+$.
To invalidate the divergences coming from the $ \epsilon^{-2}$ factors, the postdate condition must be satisfied: there must be a solution $ \xi_0=\xi_0(q)$ of the equality: $ { \cal H}(\xi_0) \equiv \frac{\sqrt{\pi}}{\sqrt{\xi_0 } }
f_q(2\xi_0) - h^2_q(\xi_0)=0$. The existence of such a solution can be verified numerically (see human body 2). Analytically this can be confirmed by looking at the asymptotic behavior of the function $ { \cal H}(x)$ as $ x \to \infty$, which yields a minus value: $ { \cal H}(x\to \infty) \sim
-\frac{\pi^3x^{2q}e^{2x}}{\Gamma^2 (\frac{1+q}{2 })
\Gamma^2(1 + \frac{q}{2 }) } < 0$. This behavior comes entirely from the term $ h_q^2(x)$, given that $ f_q(x\to \infty) \to 0^+$.
\[sol\ ]
As we shall show downstairs, for various values of $ q$, | 2) - h^2_e(\xi_0)\right)~.\end{aligned}$$
Expandlng in powers of $\epsiloi$, we obfain $$\begkn{aligned}
\label{dd}
&~& \langle {\cal D}^{(w)}(z) {\cao D}^{(q)}(0) \rangle_c =
{\tilde N}_d \epsilon^{-2}\peft(\frac{\wqrt{\ki}}{\sqrt{\xi_0}}
f_q(2\xi_0) - h^2_q(\xm_0)\dight) +
{\bnlde H}^2_e~\frae{\swrt{\pi}}{\sqrt{\xi_0}}~f'_q(2\xl_0)2
{\rm ln}|z/L|^2 + \nmnumber \\
&~& +~ \epsinov^2 {\cilde N}_q^2 \sqrt{\frac{\pi}{\xi_0}}
\frac{1}{2}f''_q(2\xi_0)4 {\rm ln}^2|z/J|^2
+ \dots \ejd{aligned}$$ wherg $\dotx $ dehote terms that vanish as $\epsilon \fo 0^+$.
To anoid the divergenves coming from the $\epsiloj^{-2}$ faftors, the followinh condition muse be satisfiea: there must be a solujion $\xi_0=\xi_0(q)$ of the equation: ${\cal H}(\bi_0) \eqbiv \frac{\sqrj{\'u}}{\sqgj{\xi_0}}
f_q(2\xi_0) - h^2_q(\xm_0)=0$. The vxistence of such a smlution can be verificd nukerucally (see figure 2). Anelytically this can fe confirkeb by looking at the awynptothc bahavkiur of tie runctiln ${\ral H}(x)$ as $x \to \infty$, wyich yields a negatovq value: ${\cal H}(x\fo \infey) \sim
-\frac{\pi^3x^{2q}e^{2x}}{\Gamma^2 (\frac{1+q}{2})
\Gamma^2(1 + \frac{x}{2})} < 0$. This behaviour comes ebtirely from the term $h_q^2(x)$, giveg that $f_q(x\to \infty) \to 0^+$.
\[sol\]
As we shall show below, fmr vaciuus vwuyed of $q$, | 2) - h^2_q(\xi_0)\right)~.\end{aligned}$$ Expanding in powers of obtain \label{dd} &~& {\cal D}^{(q)}(z) {\cal \epsilon^{-2}\left(\frac{\sqrt{\pi}}{\sqrt{\xi_0}} - h^2_q(\xi_0)\right) + N}^2_q~\frac{\sqrt{\pi}}{\sqrt{\xi_0}}~f'_q(2\xi_0)2 {\rm ln}|z/L|^2 \nonumber \\ &~& +~ \epsilon^2 {\tilde \sqrt{\frac{\pi}{\xi_0}} \frac{1}{2}f''_q(2\xi_0)4 {\rm ln}^2|z/L|^2 + \dots \end{aligned}$$ where $\dots $ denote terms that as $\epsilon \to 0^+$. To avoid the divergences coming from the $\epsilon^{-2}$ factors, following must satisfied: must be a solution $\xi_0=\xi_0(q)$ of the equation: ${\cal H}(\xi_0) \equiv \frac{\sqrt{\pi}}{\sqrt{\xi_0}} f_q(2\xi_0) - h^2_q(\xi_0)=0$. The of such a solution can be verified numerically figure 2). Analytically this be confirmed by looking at asymptotic of the ${\cal as \to \infty$, which a negative value: ${\cal H}(x\to \infty) \sim -\frac{\pi^3x^{2q}e^{2x}}{\Gamma^2 (\frac{1+q}{2}) \Gamma^2(1 + \frac{q}{2})} < 0$. This behaviour comes from the given that \infty) 0^+$. As we shall for various values of $q$, | 2) - h^2_q(\xi_0)\right)~.\end{aligned}$$
ExpandIng in powerS of $\epSilOn$, wE oBtaiN $$\begIn{aligned}
\label{DD}
&~& \lanGle {\cal D}^{(q)}(z) {\cal D}^{(q)}(0) \rangle_c =
{\tIlde N}_Q \ePSiloN^{-2}\LeFt(\fraC{\sqrt{\pi}}{\SQrT{\XI_0}}
f_q(2\Xi_0) - H^2_q(\Xi_0)\rIgHT) +
{\tIlde N}^2_Q~\frAc{\sqrt{\pI}}{\sqrt{\xi_0}}~f'_q(2\xI_0)2
{\rm Ln}|Z/L|^2 + \nonumber \\
&~& +~ \epSIlOn^2 {\tilde N}_q^2 \sQrt{\Frac{\pi}{\xi_0}}
\frac{1}{2}F''_q(2\xI_0)4 {\rm ln}^2|z/l|^2
+ \dOts \ENd{aliGneD}$$ wherE $\dots $ dENote teRms that vaNiSH as $\epsILon \to 0^+$.
To AVOiD the Divergences coming FRoM The $\epsilon^{-2}$ factOrs, the FoLLoWINg cOndItion must bE sAtisfIEd: there MUsT BE A soLUtion $\xi_0=\xi_0(q)$ of tHe equation: ${\cAL H}(\xI_0) \equiv \FrAc{\sQRt{\pi}}{\sqRt{\xi_0}}
f_Q(2\xI_0) - H^2_q(\xI_0)=0$. The existenCe of Such a soluTion caN Be verifIEd numerIcally (See FigUre 2). ANAlYtIcaLlY ThiS CaN be COnfIrmed by lOoKiNg at tHe asYMPTOtic BehAvioUr of tHe function ${\cal h}(x)$ aS $x \to \INftY$, whicH yielDs a nEgAtive Value: ${\cAl H}(x\tO \iNfty) \sim
-\frac{\pi^3x^{2q}E^{2x}}{\GaMma^2 (\frac{1+q}{2})
\GAmmA^2(1 + \fRac{Q}{2})} < 0$. THis beHAviour ComEs eNtirely From the TErm $H_q^2(X)$, GIVeN that $f_q(x\to \infty) \to 0^+$.
\[sOl\]
aS We Shall shoW below, FOr VaRIous valuEs Of $q$, | 2) - h^2_q(\xi_0)\right)~. \end{align ed}$$
E xpa nd ingin p owers of $\eps i lon$ , we obtain $$\begin{a ligne d} \la b el {dd}&~& \la n gl e {\c al D }^{ (q ) }( z) {\ cal D}^{(q )}(0) \ran gle _c =
{\tilde N} _q \epsilo n^{ -2}\left(\fr ac{ \sqrt{ \p i}} { \sqrt {\x i_0}}
f_q(2 \ xi_0)- h^2_q(\ xi _ 0)\rig h t) +
{ \ t il de N }^2_q~\frac{\sqrt { \p i }}{\sqrt{\xi_0 }}~f'_ q( 2 \x i _ 0)2
{\ rm ln}|z/L |^ 2 + \ n onumber \\ & ~ & + ~ \epsilon^2 { \tilde N}_q ^ 2 \ sqrt{\ fr ac{ \ pi}{\x i_0}}
\ f rac {1}{2}f''_q (2\x i_0)4 {\r m ln}^ 2 |z/L|^2 + \dots \end {al ign ed}$ $ w he re$\ d ots $den o teterms th at v anish as$ \ e p silo n \ to 0 ^+$.
To avoid the di verg e nce s com ing f romth e $\e psilon ^{-2} $factors, the fo llow ing condi tio nmus tbe sa t isfied : t her e mustbe a so l uti on $ \ xi _0=\xi_0(q)$ of th ee q ua tion: ${ \cal H } (\ xi _ 0) \equi v\fr ac{\ s q rt{\p i}}{ \ sq rt{\xi_0 }}
f_q ( 2\ xi _0) - h ^2 _q(\xi _0 )=0 $.The e x iste nce of such asolut i on can be veri f ied numerical l y( s ee figu re2). Analyti call y thi s ca n b e c o nfirm ed by l o ok i ng at the asymptoti cbehavi our o f the functio n ${\cal H } ( x )$ as $x \to \i n fty$, which yi eldsa negative value: $ {\cal H}(x\to \infty)\ s im
-\fr ac{ \pi ^3x ^{2 q } e^ {2x}}{\Gamma^ 2 (\fr ac {1+q}{2 })
\Gamm a^2 (1+ \ fra c{ q}{2})} < 0$. Thi sbe ha vi our come s entirel yfro mthe term $h_q^2 (x)$, giv en t h at$f_q(x\ t o\ i nfty )\t o 0^ +$.
\[sol \]
A s w e shall show bel ow, forva ri ous val ues of $q$, | 2) -_h^2_q(\xi_0)\right)~.\end{aligned}$$
Expanding in_powers of $\epsilon$, we_obtain $$\begin{aligned}
_\label{dd}
&~&_\langle {\cal_D}^{(q)}(z)_{\cal D}^{(q)}(0) \rangle_c_=
{\tilde_N}_q \epsilon^{-2}\left(\frac{\sqrt{\pi}}{\sqrt{\xi_0}}
f_q(2\xi_0) - h^2_q(\xi_0)\right)_+
{\tilde N}^2_q~\frac{\sqrt{\pi}}{\sqrt{\xi_0}}~f'_q(2\xi_0)2
{\rm_ln}|z/L|^2_+ \nonumber \\
&~& +~ \epsilon^2 {\tilde N}_q^2 \sqrt{\frac{\pi}{\xi_0}}
\frac{1}{2}f''_q(2\xi_0)4 {\rm ln}^2|z/L|^2
+ \dots \end{aligned}$$ where_$\dots_$ denote_terms_that_vanish as $\epsilon \to 0^+$.
To_avoid the divergences coming from_the $\epsilon^{-2}$_factors, the following condition must be satisfied: there_must_be a solution_$\xi_0=\xi_0(q)$ of the equation: ${\cal H}(\xi_0) \equiv \frac{\sqrt{\pi}}{\sqrt{\xi_0}}
f_q(2\xi_0) -_h^2_q(\xi_0)=0$. The existence of such a_solution can be_verified_numerically_(see figure 2). Analytically_this can be confirmed by looking_at the asymptotic behaviour of the_function ${\cal H}(x)$ as $x \to \infty$,_which yields a negative value: ${\cal_H}(x\to \infty) \sim
-\frac{\pi^3x^{2q}e^{2x}}{\Gamma^2 (\frac{1+q}{2})_
\Gamma^2(1_+ \frac{q}{2})} < 0$. This_behaviour comes entirely_from the_term $h_q^2(x)$, given_that $f_q(x\to \infty) \to 0^+$.
\[sol\]
As we_shall show below,_for various values of $q$, |
; @Mahajan:2009pj; @Branz:2009yt; @Zhang:2009vs; @Albuquerque:2009ak; @Ding:2009vd]. Each sub-system in the molecule could decay into $DK$ pair, but it is not kinematically allowed ([*i.e.*]{} $DK$ mass is above $D^{*}_s$), thus the $D^{*+}_sD^{*-}_s$ molecule has narrow natural width. To further test the molecular hypothesis of $Y(4140)$, the experimental measurement of the photon spectrum of $Y(4140)\rightarrow {D}_s^{\ast+} D_s^- \gamma, D_{s}^+D_{s}^{*-}\gamma$ are suggested. The $D^{*+}_sD^{*-}_s$ molecule seems a viable interpretation of $Y(4140)$, but no concrete conclusion can be made. Furthermore, the two-photon partial width measured by Belle experiment [@belleevidence] disfavor the interpretation of $D^{*+}_sD^{*-}_s$ molecule with $J^{PC}$= $0^{++}$ or $2^{++}$ for $Y(4140)$. [**${\bf c\bar{c}q\bar{q}}$ tetra-quark state**]{}: The $J/\psi\phi$ decay channel of $Y(4140)$ contains $c\bar{c}$ and $s\bar{s}$ components, thus it was also considered as a $c\bar{c}s\bar{s}$ tetra-quark state without sub-systems inside through rearrangement of the constitute quarks with a width of about 100 MeV, and about equal probabilities for both hidden and open-charm two-body decays [@Stancu:2009ka]. The observed width does not seems support this proposal. [**Charmonium hybrid state**]{}: Since the $Y(4140)$ mass is in the range of 4.1-4.4 GeV as predicted by various models for charmonium hybrid states, it is also considered as a charmonium hybrid candidate [@Mahajan:2009pj]. The width of a hybrid is expected to be broad, however if it has exotic $J^{PC}$ of $1^{-+}$ and its mass lies below the threshold of $D^{**}D$, then its width can be narrow [@hybridselection]. This model predicts $D^*\bar{D}$ an an | ; @Mahajan:2009pj; @Branz:2009yt; @Zhang:2009vs; @Albuquerque:2009ak; @Ding:2009vd ]. Each sub - system in the molecule could decay into $ DK$ couple, but it is not kinematically give up ([ * i.e. * ] { } $ DK$ mass is above $ D^{*}_s$), therefore the $ D^{*+}_sD^{*-}_s$ molecule have narrow lifelike width. To further test the molecular hypothesis of $ Y(4140)$, the experimental measurement of the photon spectrum of $ Y(4140)\rightarrow { D}_s^{\ast+ } D_s^- \gamma, D_{s}^+D_{s}^{*-}\gamma$ are propose. The $ D^{*+}_sD^{*-}_s$ molecule seems a feasible interpretation of $ Y(4140)$, but no concrete conclusion can be made. Furthermore, the two - photon partial width measure by Belle experiment [ @belleevidence ] disfavor the interpretation of $ D^{*+}_sD^{*-}_s$ molecule with $ J^{PC}$= $ 0^{++}$ or $ 2^{++}$ for $ Y(4140)$. [ * * $ { \bf c\bar{c}q\bar{q}}$ tetra - quark state * * ] { }: The $ J/\psi\phi$ decay channel of $ Y(4140)$ incorporate $ c\bar{c}$ and $ s\bar{s}$ components, thus it was also consider as a $ c\bar{c}s\bar{s}$ tetra - quark state without sub - systems inside through rearrangement of the constitute quark with a width of about 100 MeV, and about equal probabilities for both hidden and open - appeal two - body decays [ @Stancu:2009ka ]. The observed width does not seems support this proposal. [ * * Charmonium hybrid state * * ] { }: Since the $ Y(4140)$ mass is in the range of 4.1 - 4.4 GeV as predicted by various models for charmonium hybrid states, it is also considered as a charmonium hybrid candidate [ @Mahajan:2009pj ]. The width of a hybrid is expect to be broad, however if it has alien $ J^{PC}$ of $ 1^{-+}$ and its bulk dwell below the threshold of $ D^{**}D$, then its width can be narrow [ @hybridselection ]. This model predicts $ D^*\bar{D}$ an an | ; @Majajan:2009pj; @Branz:2009yt; @Zhang:2009vs; @Xlbuquerque:2009ak; @Dnbg:2009vd]. Eech sub-aystem iv the molecule could decay iito $EK$ paur, but it is not kinemxtically wllowed ([*u.e.*]{} $DJ$ mass is aukve $D^{*}_s$), bkus tgc $D^{*+}_sD^{*-}_v$ molecule has karrow natusal width. To fgrghzr test the molecular hypothesis of $R(4140)$, the ecpfrimental measorememe of nht photon spectrum of $Y(4140)\rightarrow {S}_s^{\ast+} D_v^- \gamma, D_{s}^+D_{s}^{*-}\gsmma$ are suggested. The $D^{*+}_sD^{*-}_d$ mopecule seems a viahle interprgfatyin of $Y(4140)$, but vo concrett eonclusion dan be made. Furthermore, the two-ohotou partial wudrh lgasured by Bxlle eqperiment [@bellcvvidenca] disfafor the interpvetatmon if $D^{*+}_sD^{*-}_s$ molecule with $J^{PC}$= $0^{++}$ or $2^{++}$ for $Y(4140)$. [**${\bf c\bar{c}q\bas{q}}$ tetra-quark state**]{}: Tye $J/\pvi\phh$ dezqy zhahnxl kf $Y(4140)$ clnteins $c\bar{c}$ znd $s\bar{s}$ cimponents, thus it wss qlso considersd as w $s\bar{c}s\bar{s}$ tetra-quark state without sub-vysfems inside through reaerangement of the condtitute qtarks with a width of about 100 MeV, and about equal [robauiuitncf dog both hidden and open-charm two-body decays [@Stagdu:2009la]. The observed cidth does not srels fupport this kroposal. [**Cgarmonium hybrid shate**]{}: Sigce tye $Y(4140)$ mass is on the range of 4.1-4.4 GeV as preeicted by vagiouw models for charmlnium hybrib statgs, it os also considered as a chadmonium hybgid candisxte [@Mahajan:2009pj]. The wictv of a hybrid is expected eo be broed, hocever if it nas exjtic $J^{PC}$ ov $1^{-+}$ and its mass lies bellw thg threvhold of $D^{**}F$, then its width can be narrow [@hburidselection]. Tvis model pxedictx $D^*\bar{D}$ an an | ; @Mahajan:2009pj; @Branz:2009yt; @Zhang:2009vs; @Albuquerque:2009ak; @Ding:2009vd]. Each the could decay $DK$ pair, but ([*i.e.*]{} mass is above thus the $D^{*+}_sD^{*-}_s$ has narrow natural width. To further the molecular hypothesis of $Y(4140)$, the experimental measurement of the photon spectrum of {D}_s^{\ast+} D_s^- \gamma, D_{s}^+D_{s}^{*-}\gamma$ are suggested. The $D^{*+}_sD^{*-}_s$ molecule seems a viable interpretation $Y(4140)$, no conclusion be made. Furthermore, the two-photon partial width measured by Belle experiment [@belleevidence] disfavor the interpretation of molecule with $J^{PC}$= $0^{++}$ or $2^{++}$ for $Y(4140)$. c\bar{c}q\bar{q}}$ tetra-quark state**]{}: The decay channel of $Y(4140)$ contains and components, thus was considered a $c\bar{c}s\bar{s}$ tetra-quark without sub-systems inside through rearrangement of the constitute quarks with a width of about 100 MeV, and equal probabilities hidden and two-body [@Stancu:2009ka]. observed width does support this proposal. [**Charmonium hybrid state**]{}: mass is in the range of 4.1-4.4 GeV predicted by models for charmonium hybrid states, it also considered as a charmonium hybrid candidate [@Mahajan:2009pj]. width of a hybrid is expected to be broad, however if it has exotic $J^{PC}$ and its mass lies the threshold of then width be [@hybridselection]. This predicts $D^*\bar{D}$ an an | ; @Mahajan:2009pj; @Branz:2009yt; @Zhang:2009vs; @AlBuquerque:2009aK; @Ding:2009Vd]. EAch SuB-sysTem iN the molecule coULd deCay into $DK$ pair, but it is noT kineMaTIcalLY aLloweD ([*i.e.*]{} $DK$ maSS iS ABovE $D^{*}_S$), tHus ThE $d^{*+}_sd^{*-}_s$ molEcuLe has naRrow naturaL wiDtH. To further teST tHe moleculaR hyPothesis of $Y(4140)$, tHe eXperimEnTal MEasurEmeNt of tHe photON spectRum of $Y(4140)\rigHtARrow {D}_s^{\ASt+} D_s^- \gamMA, d_{s}^+d_{s}^{*-}\gaMma$ are suggested. ThE $d^{*+}_sd^{*-}_S$ molecule seems A viablE iNTeRPRetAtiOn of $Y(4140)$, but no CoNcretE ConclusIOn CAN Be mADe. Furthermore, The two-photoN ParTial wiDtH meASured bY BellE eXPerIment [@belleeVideNce] disfavOr the iNTerpretATion of $D^{*+}_SD^{*-}_s$ molEcuLe wIth $J^{pc}$= $0^{++}$ oR $2^{++}$ fOr $Y(4140)$. [**${\Bf C\Bar{C}Q\bAr{q}}$ TEtrA-quark stAtE**]{}: THe $J/\psI\phi$ DECAY chaNneL of $Y(4140)$ ContaIns $c\bar{c}$ and $s\bAr{s}$ CompONenTs, thuS it waS alsO cOnsidEred as A $c\bar{C}s\Bar{s}$ tetra-quark sTate Without suB-syStEms InSide tHRough rEarRanGement oF the conSTitUtE QUArKs with a width of abouT 100 MEv, AnD about eqUal proBAbIlITies for bOtH hiDden AND open-CharM TwO-body decAys [@StaNCu:2009Ka]. the obseRvEd widtH dOes Not Seems SUppoRt this Proposal. [**charmONium hybrid statE**]{}: since the $Y(4140)$ mass IS iN THe RAnge Of 4.1-4.4 GEV as predictEd by VArioUs moDElS foR CharmOnium HyBRiD States, it is also consiDeRed as a CharmOnium hybrid caNdidate [@MahAJAN:2009pj]. The wiDth oF A hYBrid is expected To be bRoad, howeveR If it has eXotic $j^{PC}$ of $1^{-+}$ and Its mass liES Below the ThrEshOld Of $D^{**}d$, THeN its width can bE NArroW [@hYbridseLecTion]. ThiS moDel PreDicTs $d^*\bar{D}$ an an | ; @Mahajan:2009pj; @Branz: 2009yt; @Z hang: 200 9vs ;@Alb uque rque:2009ak; @ D ing: 2009vd]. Each sub-syst em in t h e mo l ec ule c ould de c ay i nto $ DK $ p ai r ,but i t i s not k inematical lyal lowed ([*i.e . *] {} $DK$ ma ssis above $D^ {*} _s$),th ust he $D ^{* +}_sD ^{*-}_ s $ mole cule hasna r row na t ural wi d t h. Tofurther test them ol e cular hypothes is of$Y ( 41 4 0 )$, th e experime nt al me a suremen t o f t hep hoton spectru m of $Y(414 0 )\r ightar ro w { D }_s^{\ ast+} D _ s^- \gamma, D_ {s}^ +D_{s}^{* -}\gam m a$ ares uggeste d. The $D ^{* +}_s D ^{ *- }_s $m ole c ul e s e ems a viabl ein terpr etat i o n of $ Y(4 140) $, bu t no concrete co nclu s ion canbe ma de.Fu rther more,the t wo -photon partial wid th measur edby Be ll e exp e riment [@ bel leevide nce] di s fav or t h einterpretation of$D ^ { *+ }_sD^{*- }_s$ m o le cu l e with $ J^ {PC }$=$ 0 ^{++} $ or $2 ^{++}$ f or $Y( 4 14 0) $. [**$ {\ bf c\b ar {c} q\b ar{q} } $ te tra-qu ark stat e**]{ } : The $J/\psi\ p hi$ decay cha n ne l of $Y(4 140 )$ contains $c\ b ar{c }$ a n d$s\ b ar{s} $ com po n en t s, thus it was also c onside red a s a $c\bar{c} s\bar{s}$t e t ra-quark sta t ew ithout sub-sys temsinside thr o ugh rear range ment ofthe const i t ute quar kswit h a wi d t hof about 100M e V, a nd aboutequ al prob abi lit ies fo rboth hidd en and o pe n- ch ar m t wo-bo d y decays [ @St an cu: 2009k a ]. The obse rved w id t h d oes not se e m s su pp or t th ispr oposa l. [ * *Ch armoniu m hybridsta t e**] {} :Since t he $Y(4140)$ma ss is in t he ra nge of 4 .1-4.4 G eV as predicted by vari o us mode lsfor c harm onium hyb rid state s,i t is a lso co nside re d a s a cha r m on ium h ybrid cand i d ate [@Ma ha jan: 2009pj] . The width of a h y bri d is expected to beb r oa d,h ow e ver i f it h as exotic $J^{P C}$ of $1^ {- + }$ and its m a ssli es belo w the t hresh o ld of $ D^{**}D$, then its w idth c anbe narrow[@hybrid selection ] . Thi s m odelpre dicts$D ^*\ bar{D }$ ana n | ; @Mahajan:2009pj;_@Branz:2009yt; @Zhang:2009vs;_@Albuquerque:2009ak; @Ding:2009vd]. Each sub-system_in the_molecule_could decay_into_$DK$ pair, but_it is not_kinematically allowed ([*i.e.*]{} $DK$_mass is above_$D^{*}_s$),_thus the $D^{*+}_sD^{*-}_s$ molecule has narrow natural width. To further test the molecular hypothesis_of_$Y(4140)$, the_experimental_measurement_of the photon spectrum of_$Y(4140)\rightarrow {D}_s^{\ast+} D_s^- \gamma, D_{s}^+D_{s}^{*-}\gamma$_are suggested._The $D^{*+}_sD^{*-}_s$ molecule seems a viable interpretation of_$Y(4140)$,_but no concrete_conclusion can be made. Furthermore, the two-photon partial width_measured by Belle experiment [@belleevidence] disfavor the_interpretation of $D^{*+}_sD^{*-}_s$_molecule_with_$J^{PC}$= $0^{++}$ or $2^{++}$_for $Y(4140)$. [**${\bf c\bar{c}q\bar{q}}$ tetra-quark state**]{}:_The $J/\psi\phi$ decay channel of $Y(4140)$_contains $c\bar{c}$ and $s\bar{s}$ components, thus it_was also considered as a $c\bar{c}s\bar{s}$_tetra-quark state without sub-systems inside_through rearrangement_of the constitute quarks with_a width of_about 100_MeV, and about_equal probabilities for both hidden and_open-charm two-body decays [@Stancu:2009ka]._The observed width does not seems_support_this proposal. [**Charmonium_hybrid_state**]{}:_Since the_$Y(4140)$ mass is_in_the range_of_4.1-4.4 GeV as predicted by various_models_for charmonium hybrid states, it is also_considered as a charmonium_hybrid_candidate [@Mahajan:2009pj]. The width of_a hybrid is expected to_be broad, however if it has_exotic $J^{PC}$_of $1^{-+}$_and its mass lies below the threshold of $D^{**}D$, then its_width can be narrow [@hybridselection]. This model_predicts $D^*\bar{D}$ an an |
ten [-@Selten65]. A strategy profile is a subgame-perfect equilibrium if it is a NE at every subgame of the original game. Informally, this means that at any history of the game (even those that are not on any equilibrium path), if all the players follow their strategy from that point on, then no player has an incentive to deviate. In the context of repeated games where players’ moves are observed (so that it is a game of perfect information), the folk theorem continues to hold even if the solution concept used is subgame-perfect equilibrium [@AS94; @fudenberg1986folk; @Ru79].
We define a computational analogue of subgame-perfect equilibrium that we call *computational subgame-perfect ${{\epsilon}}$-equilibrium*, where the strategies involved are polynomial time, and deviating players are again restricted to using polynomial-time strategies. There are a number of subtleties that arise in defining this notion. While we assume that all actions in the underlying repeated game are observable, we allow our TMs to also have memory, which means the action of A TM does not depend only on the public history. Like subgame-perfect equilibrium, our computational solution concept is intended to capture the intuition that the strategies are in equilibrium after any possible deviation. This means that in a computational subgame-perfect equilibrium, at each history for player $i$, player $i$ must make a (possibly approximate) best response, no matter what his and the other players’ memory states are. To compute a computational subgame-perfect $\epsilon$-equilibrium, we use the same basic strategy as for NE, but, as often done to get a subgame-perfect equilibrium (for example see [@fudenberg1986folk]), we limit the punishment phase length, so that the players are not incentivized not to punish deviations. However, to prove our result, we need to overcome one more significant hurdle. When using cryptographic protocols, it is often the case (and, specifically is the case in the protocol used for NE) that player $i$ chooses a secret (e.g., a secret key for a public-key encryption scheme) as the result of some randomization, and then releases some public information which is a function of the secret (e.g., a public key). After that public information has been released, another party $j$ typically has a profitable deviation by switching to the TM $M$ that can break the protocol | ten [ -@Selten65 ]. A strategy profile is a subgame - perfect equilibrium if it is a NE at every subgame of the original plot. colloquially, this mean that at any history of the game (even those that are not on any equilibrium way), if all the players follow their strategy from that period on, then no player has an bonus to deviate. In the context of recur games where player ’ moves are observed (so that it is a plot of perfect information), the folk theorem continues to hold even if the solution concept used is subgame - perfect chemical equilibrium [ @AS94; @fudenberg1986folk; @Ru79 ].
We define a computational analogue of subgame - perfect equilibrium that we address * computational subgame - perfect $ { { \epsilon}}$-equilibrium *, where the strategies involved are polynomial time, and deviate players are again restricted to using polynomial - time strategies. There be a number of subtleties that arise in defining this notion. While we wear that all actions in the underlying repeated game are observable, we allow our TMs to also have memory, which means the military action of A TM does not depend only on the public history. Like subgame - arrant chemical equilibrium, our computational solution concept is mean to capture the intuition that the strategies are in chemical equilibrium after any possible deviation. This means that in a computational subgame - perfect chemical equilibrium, at each history for player $ i$, player $ i$ must make a (possibly approximate) best response, no matter what his and the other actor ’ memory states are. To compute a computational subgame - perfect $ \epsilon$-equilibrium, we use the same basic strategy as for NE, but, equally frequently done to get a subgame - perfect equilibrium (for case see [ @fudenberg1986folk ]), we limit the punishment phase length, so that the actor are not incentivized not to punish deviations. However, to prove our result, we need to overwhelm one more significant hurdle. When use cryptographic protocol, it is often the font (and, specifically is the case in the protocol use for NE) that player $ i$ chooses a mystery (e.g., a secret key for a public - key encryption scheme) as the result of some randomization, and then releases some public information which is a function of the secret (e.g., a public samara). After that public information has been released, another party $ j$ typically have a profitable deviation by switching to the TM $ M$ that can break the protocol | ten [-@Delten65]. A strategy profilt is a subgame-perywct eqnilibrihm if it is a NE at every subgame of tye oruginal game. Informally, this meajs that qt aiy history of thx game (eyzn thkde tkav are not on ani equilibriuk path), if all dhd 'layers follow their strategy from trat poimt on, then no plwyer ras zn incentive to deviate. In the confext of repeated gamrs where players’ moves are obsfrved (so that it id a game of perswct informatkon), the folk theorem ckntinues to hold even if the souutiou concept uwee id subgame-perhect ezuilibrium [@AS94; @fudenbesg1986folk; @Tu79].
We define a gompuvatiinal analogue of subgeme-perfect equilibriom that we ccll *computational subtane-perxect ${{\epskoon}}$-dqujlmbrjum*, whfre the stratsgies involced are polynomial uimq, and deviatinf playqrf are again restricted to using polynompal-tjme strategies. There arw a number of subtletles that wrise in defining this notion. While we assume thad all xctnins iv tje underlying repeated game are observable, we zlkoe our TMs to ajso have mekogy, rhich means tfe actnkn of A TM does not fepend jnly in the puflic history. Like subgame-perfecr equilibriui, our computational solution cuncekt is ontended to capture the inthition that the strafdgies are in equklinrigm after any possible deviwtion. Thiw mecns that in s compttational dubgame-perfect equilibrikm, at edch historj for player $i$, player $i$ must make a (possibly sp[roqimate) beft rexponse, no mateer what his aud the ocher puayers’ memkry staves are. To cjmpute a compgjational subgeme-perfece $\epwilob$-equilicfium, we use thr same basic stratety as for NE, but, ax owfen done to get a wubgame-perfect rquklifrpum (fjs example sea [@fuddnbdtg1986foly]), we limit bhe punoshment phase length, so fhat the players ate not incgntivized not to punisn deviations. Howevtr, to 'rove uur rrsujt, we need to overcome one mors signififanb hurdle. When usikg ctyptographie protocols, it is often the case (and, sperifically is the case ib the protocol used nor NE) that 'layer $i$ choosev a secret (e.g., a secrwt key for a publlc-key encryption schems) as tve redult of some randomization, and then releases some public information whicy is a function of the secrvt (z.g., c publis keb). Efter that public information has been released, enother pastv $j$ typically has a profitablr aeviation by rwitching to the TM $M$ thzt can bteak the protocol | ten [-@Selten65]. A strategy profile is a if is a at every subgame this that at any of the game those that are not on any path), if all the players follow their strategy from that point on, then player has an incentive to deviate. In the context of repeated games where moves observed that is a game of perfect information), the folk theorem continues to hold even if the solution used is subgame-perfect equilibrium [@AS94; @fudenberg1986folk; @Ru79]. We a computational analogue of equilibrium that we call *computational ${{\epsilon}}$-equilibrium*, the strategies are time, deviating players are restricted to using polynomial-time strategies. There are a number of subtleties that arise in defining this notion. we assume actions in underlying game observable, we allow to also have memory, which means A TM does not depend only on the history. Like equilibrium, our computational solution concept is to capture the intuition that the strategies are equilibrium after any possible deviation. This means that in a computational subgame-perfect equilibrium, at each player $i$, player $i$ make a (possibly best no what and the players’ memory states are. To compute a computational subgame-perfect $\epsilon$-equilibrium, we the same basic strategy as for NE, but, as often get subgame-perfect equilibrium (for see [@fudenberg1986folk]), we limit punishment length, so that the not not However, prove result, we need to one more significant hurdle. When cryptographic protocols, it is is the case in the protocol used for that player $i$ chooses a secret (e.g., secret key for a public-key encryption scheme) as the result of some and then public information which is a function of the (e.g., a public key). that public information has been released, another party $j$ has profitable deviation switching to the $M$ that can the protocol | ten [-@Selten65]. A strategy profile Is a subgame-PerfeCt eQuiLiBriuM if iT is a NE at every sUBgamE of the original game. InfoRmallY, tHIs meANs That aT any hisTOrY OF thE gAmE (evEn THoSe thaT arE not on aNy equilibrIum PaTh), if all the plAYeRs follow thEir Strategy from ThaT point On, TheN No plaYer Has an IncentIVe to deViate. In thE cONtext oF RepeateD GAmEs whEre players’ moves arE ObSErved (so that it iS a game Of PErFECt iNfoRmation), the FoLk theORem contINuES TO hoLD even if the solUtion concepT UseD is subGaMe-pERfect eQuiliBrIUm [@As94; @fudenberg1986fOlk; @RU79].
We define A compuTAtional ANalogue Of subgAme-PerFect EQuIlIbrIuM ThaT We CalL *ComPutationAl SuBgame-PerfECT ${{\EPsilOn}}$-eQuilIbriuM*, where the straTegIes iNVolVed arE polyNomiAl Time, aNd deviAting PlAyers are again reStriCted to usiNg pOlYnoMiAl-timE StrateGieS. ThEre are a Number oF SubTlETIEs That arise in defininG tHIS nOtion. WhiLe we asSUmE tHAt all actIoNs iN the UNDerlyIng rEPeAted game Are obsERvAbLe, we allOw Our TMs To AlsO haVe memORy, whIch meaNs the actIon of a tM does not depenD Only on the publIC hISToRY. LikE suBgame-perfecT equILibrIum, oUR cOmpUTatioNal soLuTIoN Concept is intended to CaPture tHe intUition that the Strategies ARE In equiliBriuM AfTEr any possible dEviatIon. This meaNS that in a CompuTational Subgame-peRFEct equilIbrIum, At eAch HIStOry for player $i$, PLAyer $I$ mUst make A (poSsibly aPprOxiMatE) beSt Response, nO matter wHaT hIs AnD thE otheR Players’ mEmOry StAteS are. TO ComputE a comPutaTiOnAL suBgame-peRFeCT $\EpsiLoN$-eQuilIbrIuM, we usE the SAme Basic stRategy as fOr Ne, But, aS oFtEn done tO get a subgame-pErFect equiliBrIum (For exaMPLe see [@fudEnberg1986folk]), we limit the punIShment pHasE lengTh, so That the plAyeRs are nOt iNCentivIzed noT to puNiSh dEVIatioNS. hoWevEr, To prove our RESulT, we neEd To ovErcome oNe more significant hURdlE. When using cryPtoGrapHIC pRotOCoLS, it Is OFteN THe case (and, specifIcally is thE cASe In the protoCOl uSeD for NE) tHat playEr $i$ chOOses a seCret (e.g., a seCret key foR a PublIC-Key Encryption Scheme) as The result OF some RAnDomizAtiOn, and tHeN reLeaseS some pUBliC infoRmatioN wHich is A funcTiOn of the sEcret (e.g., a public key). After tHat pubLic inForMation has BeeN RelEased, anotHer pArty $j$ typicAllY haS a proFitABle deViatIOn By sWItchiNg to THe TM $M$ that CAn BreAK ThE protocol | ten [-@Selten65]. A strate gy profile is a su bga me -per fect equilibrium i f itis a NE at every subga me of t h e or i gi nal g ame. In f or m a lly ,th isme a ns that at any hi story of t hega me (even tho s ethat are n oton any equil ibr ium pa th ),i f all th e pla yers f o llow t heir stra te g y from that po i n ton,then no player ha s a n incentive todeviat e. In t hecon text of re pe atedg ames wh e re p l aye r s’ moves areobserved (s o th at itis ag ame of perf ec t in formation), the folk the orem c o ntinues to hold evenifthe sol u ti on co nc e ptu se d i s su bgame-pe rf ec t equ ilib r i u m [@A S94 ; @f udenb erg1986folk;@Ru 79].
We defi ne acomp ut ation al ana logue o f subgame-perfe ct e quilibriu m t ha t w ecall* comput ati ona l subga me-perf e ct${ { \ e ps ilon}}$-equilibriu m* , wh ere thestrate g ie si nvolvedar e p olyn o m ial t ime, an d deviat ing pl a ye rs are ag ai n rest ri cte d t o usi n g po lynomi al-timestrat e gies. There ar e a number ofs ub t l et i es t hat arise in d efin i ng t hisn ot ion . Whil e weas s um e that all actions i nthe un derly ing repeatedgame are o b s e rvable,we a l lo w our TMs to al so ha ve memory, which me ans t he actio n of A TM d oes notdep end on lyo n t he public his t o ry.Li ke subg ame -perfec t e qui lib riu m, our comp utationa lso lu ti onconce p t is int en ded t o c aptur e the i ntuit ionth at the strate g ie s arein e quil ibr iu m aft er a n y p ossible deviatio n.T hisme an s thatin a computat io nal subgam e- per fect e q u ilibrium , at each history for p l ayer $i $,playe r $i $ must ma kea (pos sib l y appr oximat e) be st re s p onse, n omat te r what his a ndthe o th er p layers’ memory states are . To compute a co mpu tati o n al su b ga m e-p er f ect $ \epsilon$-equil ibrium, we u s ethe same b a sic s trategy as for NE,b ut, asoften don e to getasubg a m e-p erfect equ ilibrium (for exa m ple s e e[@fud enb erg198 6f olk ]), w e limi t th e pun ishmen tphaselengt h, so that the players are not in centiv izednot to punis h d e via tions. Ho weve r, to prov e o urresul t,w e nee d to ov erc o me on e mo r e signifi c an t h u r dl e. When usi n g cry ptogr aph i c prot ocol s, it is often th e case (and, sp ecif i c all y i s the c ase in the pro toc ol u sed forNE ) that play er $i$ c ho o ses a secre t (e.g ., a se c r et key fo r apub lic-key e ncr yp t ion sch em e) as the res ul t of s ome ra n domi z a tion, and then r eleas e s some pub lic i nf ormatio n whi ch is a fu nction of t he sec ret(e.g. , a pub li c key) . A ft er that pu b lic infor matio n has b ee n re lea sed, a noth e r part y $j $typ ically ha s ap ro fi t abl e de viati on byswitching to the T M $ M $ thatca n b r e ak the pr o t ocol | ten [-@Selten65]. A_strategy profile_is a subgame-perfect equilibrium_if it_is_a NE_at_every subgame of_the original game._Informally, this means that_at any history_of_the game (even those that are not on any equilibrium path), if all the_players_follow their_strategy_from_that point on, then no_player has an incentive to_deviate. In_the context of repeated games where players’ moves_are_observed (so that_it is a game of perfect information), the folk_theorem continues to hold even if_the solution concept_used_is_subgame-perfect equilibrium [@AS94; @fudenberg1986folk; @Ru79].
We_define a computational analogue of subgame-perfect_equilibrium that we call *computational subgame-perfect_${{\epsilon}}$-equilibrium*, where the strategies involved are polynomial_time, and deviating players are again_restricted to using polynomial-time strategies._There are_a number of subtleties that_arise in defining_this notion._While we assume_that all actions in the underlying_repeated game are_observable, we allow our TMs to_also_have memory, which_means_the_action of_A TM does_not_depend only_on_the public history. Like subgame-perfect equilibrium,_our_computational solution concept is intended to capture_the intuition that the_strategies_are in equilibrium after_any possible deviation. This means_that in a computational subgame-perfect equilibrium,_at each_history for_player $i$, player $i$ must make a (possibly approximate) best response,_no matter what his and the_other players’ memory states_are. To_compute_a computational subgame-perfect_$\epsilon$-equilibrium,_we use_the same basic strategy as for NE,_but, as_often done to get a subgame-perfect_equilibrium (for example see [@fudenberg1986folk]),_we_limit the punishment phase length, so_that the players are not incentivized_not to punish deviations. However,_to_prove_our result, we need to_overcome one more significant hurdle. When_using cryptographic protocols,_it is often the case (and, specifically_is_the case in the protocol used_for_NE) that player $i$ chooses a_secret_(e.g.,_a secret key for a_public-key encryption scheme) as the result_of some randomization, and then releases some public information_which is a_function of the secret (e.g.,_a_public_key). After that public information has been released, another party_$j$ typically_has a profitable_deviation by switching to the TM $M$ that can break_the protocol |
absorbing epoxy to suppress quasiparticle generation due to blackbody radiation from higher temperature stages of the cryostat [@Barends11]. The dc and microwave control lines are heavily filtered at 4.2 K and at the cold stage to suppress thermal excitation of the qubit.
We have performed standard microwave pulse sequences to evaluate qubit coherence over a range of bias points; results are shown in Fig. \[fig:figure3\]. Qubit Rabi oscillations decay with a characteristic time in excess of 500 ns. Qubit Ramsey fringes decay on a shorter timescale of order 100 ns; as our qubit has no flux sweet spot Ramsey decay is dominated by excess low frequency magnetic flux noise [@Ithier05; @Bialzak07; @Sank12]. Qubit energy relaxation times up to 1.6 $\mu$s have been measured, and energy relaxation times greater than 1 $\mu$s have been measured on multiple devices. These SOI-based phase qubits thus display energy relaxation times that are a factor of 2-3 better than the $T_1$ times of state-of-the-art phase qubits incorporating low-loss a-Si:H shunt capacitors [@Wang08; @Neeley08]. It is worthwhile to consider the coherence budget of our qubit device. First, dielectric loss in the crystalline silicon shunt capacitor presents an ultimate limit to qubit $T_1$. However, the measured loss tangent of $5\times 10^{-6}$ would correspond in our device to an energy relaxation time around 10 $\mu$s, far in excess of the measured $T_1$. Next, qubit $T_1$ is limited by coupling to the heavily damped readout resonator, a phenomenon known as the Purcell effect [@Houck08]. However for our device and operating parameters the Purcell limit to qubit $T_1$ is around 20 $\mu$s. Instead the current devices appear to be limited by overcoupling to the flux bias coil: experiments reveal an oscillatory dependence of $T_1$ on qubit bias frequency which we have determined to arise from a frequency-dependent impedance transformation of the dissipation looking into the flux line. We anticipate that it will be straightforward to suppress this decoherence channel by re-engineering the bias tee on the flux line and by reducing the mutual inductance from the flux bias coil to the qubit.
![Characterization of SOI phase qubit coherence. (a) R | absorbing epoxy to suppress quasiparticle generation due to black body radiation sickness from higher temperature stages of the cryostat [ @Barends11 ]. The district of columbia and microwave dominance lines are heavily filter at 4.2 K and at the cold stage to suppress thermal excitement of the qubit.
We have perform standard microwave pulse sequences to measure qubit coherence over a range of bias points; results are indicate in Fig. \[fig: figure3\ ]. Qubit Rabi oscillations decay with a characteristic time in excess of 500 ns. Qubit Ramsey periphery decay on a shorter timescale of order 100 ns; as our qubit has no flux density sweet spot Ramsey decay is dominated by excess low frequency charismatic flux randomness [ @Ithier05; @Bialzak07; @Sank12 ]. Qubit energy relaxation times up to 1.6 $ \mu$s have been measured, and department of energy relaxation times greater than 1 $ \mu$s have been measured on multiple devices. These SOI - based phase qubits thus display energy relaxation times that are a factor of 2 - 3 better than the $ T_1 $ times of state - of - the - art phase qubits incorporating abject - loss a - Si: H shunt capacitor [ @Wang08; @Neeley08 ]. It is worthwhile to think the coherence budget of our qubit device. foremost, dielectric loss in the crystalline silicon shunt capacitor presents an ultimate limit to qubit $ T_1$. However, the measured loss tangent of $ 5\times 10^{-6}$ would correspond in our device to an energy relaxation meter around 10 $ \mu$s, far in excess of the measured $ T_1$. Next, qubit $ T_1 $ is limited by coupling to the heavily damp readout resonator, a phenomenon known as the Purcell effect [ @Houck08 ]. However for our device and operating parameters the Purcell limit to qubit $ T_1 $ is about 20 $ \mu$s. alternatively the current devices appear to be limited by overcoupling to the flux bias coil: experiment reveal an oscillatory dependence of $ T_1 $ on qubit bias frequency which we have determine to arise from a frequency - dependent impedance transformation of the dissipation search into the flux line. We anticipate that it will be straightforward to suppress this decoherence groove by re - engineer the bias tee on the flux density line and by reducing the mutual inductance from the flux diagonal coil to the qubit.
! [ Characterization of SOI phase qubit coherence. (a) R | abslrbing epoxy to suppress quasiparticle ywneratmon due to blacybody radiation from higher vempwratuee stages of the cryosgat [@Barenfs11]. The dx anv microwave contckl lines are gcavilv hiltered at 4.2 K akd at the cmld stage to sgporzss thermal excitation of the qubit.
Wq have legformed standatd mibrjwavs pulse sequences to evaluate qubif cohertnce over a range pf bias points; results are shoan in Fig. \[fig:figure3\]. Qubit Rabi oscyolations decxy with a bkaracteristjc time in excess of 500 ns. Qubit Rxmsey fringes dgeqy lt a shorter timefcale of ordcg 100 ns; as our qunit has no fluw swext spot Ramsey decay is dmminated by excess low freqgeucy magnetic flux noiwe [@Ithigr05; @Bidlzay07; @Savk12]. Suuit energj rxlaxation tjmes up to 1.6 $\nu$s have been measuteq, and energy rslaxatyog times greater than 1 $\mu$s have been measlred on multiple devices. Thwse SOI-based phase quhits thus display energy relaxation times that are a factos of 2-3 cetuev gyaj the $T_1$ times of state-of-the-art phase qubits igdotpprating low-losf a-Si:H shuny falwcitors [@Wang08; @Veeley08]. It is worthwhile to fonsidet the xoherence budbet of our qubit device. Firwt, dielectrib lows in the crystallnne silicon rhunj capavitor presents an ultimcte lijit to qubih $T_1$. Howevsf, the measured luss tdngent of $5\times 10^{-6}$ would corwespond ii our device to sn enewgy relaxahion bhme around 10 $\mu$s, far in ercess of the mewsured $T_1$. Next, qubit $T_1$ is limitev by coupling tm tve heavijy dakped readout wesonator, a phgnomenon hnown as the Pugcell effxct [@Houck08]. Horever for our fevice and o'erating [aranetees the Ojrcell limit tp qubit $T_1$ is around 20 $\mu$s. Instead the cmrrenj sevices appear cu be limited by pvefcotppiig to dhe flux biav cokl: dcperioents rtreao an pscillatory dependenwe or $T_1$ on qubit bias grcquency wyich we rave determinrd to arise from a freqnency-dxpendemt ympedance transformation of ths dissipahiok looking intj thc flox line. We cnticipate that it will be straightforwacd to suppress this deciherence channel by ve-engineerinj the fias tee mn the flux line and by reducing the kutual inductance froj the xlux hias coil to the qubit.
![Characterization of SOI phase qubit coherence. (a) R | absorbing epoxy to suppress quasiparticle generation due radiation higher temperature of the cryostat control are heavily filtered 4.2 K and the cold stage to suppress thermal of the qubit. We have performed standard microwave pulse sequences to evaluate qubit over a range of bias points; results are shown in Fig. \[fig:figure3\]. Qubit oscillations with characteristic in excess of 500 ns. Qubit Ramsey fringes decay on a shorter timescale of order 100 as our qubit has no flux sweet spot decay is dominated by low frequency magnetic flux noise @Bialzak07; Qubit energy times to $\mu$s have been and energy relaxation times greater than 1 $\mu$s have been measured on multiple devices. These SOI-based phase thus display times that a of better than the of state-of-the-art phase qubits incorporating low-loss [@Wang08; @Neeley08]. It is worthwhile to consider the budget of qubit device. First, dielectric loss in crystalline silicon shunt capacitor presents an ultimate limit qubit $T_1$. However, the measured loss tangent of $5\times 10^{-6}$ would correspond in our device energy relaxation time around $\mu$s, far in of measured Next, $T_1$ is by coupling to the heavily damped readout resonator, a phenomenon known the Purcell effect [@Houck08]. However for our device and operating Purcell to qubit $T_1$ around 20 $\mu$s. Instead current appear to be limited to flux reveal oscillatory of $T_1$ on qubit frequency which we have determined arise from a frequency-dependent looking into the flux line. We anticipate that will be straightforward to suppress this decoherence by re-engineering the bias tee on the flux line and by reducing mutual inductance flux bias coil to the qubit. ![Characterization of phase qubit coherence. (a) | absorbing epoxy to suppress qUasiparticLe genEraTioN dUe to BlacKbody radiation FRom hIgher temperature stages Of the CrYOstaT [@baRends11]. the dc anD MiCROwaVe CoNtrOl LInEs are HeaVily filTered at 4.2 K anD at ThE cold stage to SUpPress thermAl eXcitation of tHe qUbit.
We HaVe pERformEd sTandaRd micrOWave puLse sequenCeS To evalUAte qubiT COhErenCe over a range of biaS PoINts; results are sHown in fiG. \[FiG:FIguRe3\]. QUbit Rabi osCiLlatiONs decay WItH A CHarACteristic time In excess of 500 nS. qubIt RamsEy FriNGes decAy on a ShORteR timescale oF ordEr 100 ns; as our Qubit hAS no flux SWeet spoT RamseY deCay Is doMInAtEd bY eXCesS LoW frEQueNcy magneTiC fLux noIse [@ITHIER05; @BiaLzaK07; @SanK12]. QubiT energy relaxaTioN timES up To 1.6 $\mu$s Have bEen mEaSured, And eneRgy reLaXation times greaTer tHan 1 $\mu$s havE beEn MeaSuRed on MUltiplE deVicEs. These sOI-baseD PhaSe QUBItS thus display energy ReLAXaTion timeS that aRE a FaCTor of 2-3 betTeR thAn thE $t_1$ Times Of stATe-Of-the-art Phase qUBiTs IncorpoRaTing loW-lOss A-Si:h shunT CapaCitors [@wang08; @NeelEy08]. It iS Worthwhile to coNSider the coherENcE BUdGEt of Our Qubit device. firsT, DielEctrIC lOss IN the cRystaLlINe SIlicon shunt capacitoR pResentS an ulTimate limit to Qubit $T_1$. HoweVER, The measuRed lOSs TAngent of $5\times 10^{-6}$ wOuld cOrrespond iN Our devicE to an Energy reLaxation tIME around 10 $\mU$s, fAr iN exCesS OF tHe measured $T_1$. NeXT, QubiT $T_1$ Is limitEd bY coupliNg tO thE heAviLy Damped reaDout resoNaToR, a PhEnoMenon KNown as thE PUrcElL efFect [@HOUck08]. HowEver fOr ouR dEvICe aNd operaTInG PAramEtErS the purCeLl limIt to QUbiT $T_1$ is aroUnd 20 $\mu$s. InsTeaD The cUrReNt devicEs appear to be lImIted by overCoUplIng to tHE Flux bias Coil: experiments reveal an OScillatOry DepenDencE of $T_1$ on qubIt bIas freQueNCy whicH we havE deteRmIneD TO arisE FRoM a fReQuency-depeNDEnt ImpedAnCe trAnsformAtion of the dissipatIOn lOoking into the FluX linE. wE aNtiCIpATe tHaT It wILL be straightforwArd to supprEsS ThIs decohereNCe cHaNnel by rE-engineEring THe bias tEe on the flUx line and By ReduCINg tHe mutual inDuctance From the flUX bias COiL to thE quBit.
![ChaRaCteRizatIon of Soi phAse quBit cohErEnce. (a) R | absorbing epoxy to suppres s quasipar ticle ge ner at iondueto blackbody r a diat ion from higher temper ature s t ages of thecryosta t [ @ B are nd s1 1]. T h edc an d m icrowav e controllin es are heavily fi ltered at4.2 K and at th e c old st ag e t o supp res s the rmal e x citati on of the q u bit.
W e havep e rf orme d standard microw a ve pulse sequence s to e va l ua t e qu bit coherence o ver a range o f b i a s po i nts; resultsare shown i n Fi g. \[f ig :fi g ure3\] . Qub it Rab i oscillati onsdecay wit h a ch a racteri s tic tim e in e xce ssof 5 0 0ns . Q ub i t R a ms eyf rin ges deca yon a sh orte r t i mesc ale oforder 100 ns; as o urqubi t ha s nofluxswee tspotRamsey deca yis dominated by exc ess low f req ue ncy m agnet i c flux no ise [@Ithi er05; @ B ial za k 0 7 ;@Sank12]. Qubit en er g y r elaxatio n time s u pt o 1.6 $\ mu $shave b een m easu r ed , and en ergy r e la xa tion ti me s grea te r t han 1 $\ m u$shave b een meas uredo n multiple dev i ces. These SO I -b a s ed phas e q ubits thusdisp l ay e nerg y r ela x ation time st ha t are a factor of 2- 3better than the $T_1$ ti mes of sta t e - of-the-a rt p h as e qubits incorp orati ng low-los s a-Si:Hshunt capacit ors [@Wan g 0 8; @Neel ey0 8]. It is w or thwhile to co n s ider t he cohe ren ce budg etofour qu bi t device. First,di el ec tr iclossi n the cr ys tal li nesilic o n shun t cap acit or p r ese nts anu lt i m ateli mi t to qu bi t $T_ 1$.H owe ver, th e measure d l o ss t an ge nt of $ 5\times 10^{- 6} $ would co rr esp ond in o ur devic e to an energy relaxati o n timearo und 1 0 $\ mu$s, far in exces s o f the m easure d $T_ 1$ . N e x t, qu b i t$T_ 1$ is limite d bycoupl in g to the he avily damped reado u t r esonator, a p hen omen o n k now n a s th eP urc e l l effect [@Houc k08]. Howe ve r f or our dev i cean d opera ting pa ramet e rs thePurcell l imit to q ub it $ T _ 1$is around20 $\mu$ s. Instea d thec ur rentdev ices a pp ear to b e limi t edby ov ercoup li ng tothe f lu x bias c oil: experiments reveal an os cilla tor y depende nce of$T_1$ onqubi t bias fre que ncy whic h w e have det e rm ine d to a rise from a fr e qu enc y - de pendent imp e d a nce tran sfo r mation ofthe dissipation l o oking into the flu x lin e.W e an ti cipate that it wi ll b e straig ht forward tosuppress t h is de cohere nce ch annel b y re - engine erin g t he bias t eeon the flu xli n e andby r ed ucingthe mu t uali n ductance from th e flu x biasc oil to t he qubit.
![C haracteriz ation of SO I phas e qu bit c oherenc e. (a) R | absorbing epoxy_to suppress_quasiparticle generation due to_blackbody radiation_from_higher temperature_stages_of the cryostat_[@Barends11]. The dc_and microwave control lines_are heavily filtered_at_4.2 K and at the cold stage to suppress thermal excitation of the qubit.
We have_performed_standard microwave_pulse_sequences_to evaluate qubit coherence over_a range of bias points;_results are_shown in Fig. \[fig:figure3\]. Qubit Rabi oscillations decay with_a_characteristic time in_excess of 500 ns. Qubit Ramsey fringes decay on a_shorter timescale of order 100 ns; as_our qubit has_no_flux_sweet spot Ramsey decay_is dominated by excess low frequency_magnetic flux noise [@Ithier05; @Bialzak07; @Sank12]._Qubit energy relaxation times up to 1.6 $\mu$s_have been measured, and energy relaxation_times greater than 1 $\mu$s have_been measured_on multiple devices. These SOI-based_phase qubits thus_display energy_relaxation times that_are a factor of 2-3 better_than the $T_1$_times of state-of-the-art phase qubits incorporating_low-loss_a-Si:H shunt capacitors_[@Wang08;_@Neeley08]._It is_worthwhile to consider_the_coherence budget_of_our qubit device. First, dielectric loss_in_the crystalline silicon shunt capacitor presents an_ultimate limit to qubit_$T_1$._However, the measured loss_tangent of $5\times 10^{-6}$ would_correspond in our device to an_energy relaxation_time around_10 $\mu$s, far in excess of the measured $T_1$. Next, qubit $T_1$_is limited by coupling to the_heavily damped readout resonator,_a phenomenon_known_as the Purcell_effect_[@Houck08]. However_for our device and operating parameters the_Purcell limit_to qubit $T_1$ is around 20 $\mu$s._Instead the current devices_appear_to be limited by overcoupling to_the flux bias coil: experiments reveal_an oscillatory dependence of $T_1$_on_qubit_bias frequency which we have_determined to arise from a frequency-dependent_impedance transformation of_the dissipation looking into the flux line._We_anticipate that it will be straightforward_to_suppress this decoherence channel by re-engineering_the_bias_tee on the flux line_and by reducing the mutual inductance_from the flux bias coil to the qubit.
![Characterization of_SOI phase qubit_coherence. (a) R |
m)
&= \frac{e^{-\psi (\cW_{r_0})} \psi(\cW_{r_0})^m}{m!}
\prod_{i=1}^m \frac{{{\mathrm d}}\psi(r_{(i)}, {{\boldsymbol w}}_{(i)})}{\psi(\cW_{r_0})}\nonumber\\
&\propto \prod_{i=1}^m h( {{\boldsymbol w}}_{(i)}, {{\boldsymbol \theta}}),\end{aligned}$$ where $h$ is a parametric angular density function [e.g. @engelke2012; @beirlant2006 pp. 170–171]. In the next section the angular density models described in Section \[s:par\_models\] are fitted to the data by the maximization of the likelihood. For brevity the asymmetric logistic model is not considered since it has too many parameters. The likelihood is proportional to the product of angular densities, therefore the maximizer of is obtained equivalently by maximizing the log-likelihood $$\label{eq:aploglik}
\ell({{\boldsymbol \theta}})=\sum_{i=1}^{m}\log h({{\boldsymbol w}}_{(i)},{{\boldsymbol \theta}}).$$ Denote by $\widehat{{{\boldsymbol \theta}}}$ the maximizer of $\ell$ and by $\ell'({{\boldsymbol \theta}}) = \nabla_{{{\boldsymbol \theta}}}\,\ell({{\boldsymbol \theta}})$ the score function. Since provides an approximation of the true likelihood, then from the theory on model misspecification [e.g. @davison03 pp. 147–148] it follows that $$\label{eq:godambe}
\sqrt{n}(\widehat{{{\boldsymbol \theta}}}-{{\boldsymbol \theta}})\stackrel{d}{\rightarrow} \mathcal{N}_p({{\boldsymbol 0}},J({{\boldsymbol \theta}})^{-1}\,K({{\boldsymbol \theta}})\,J({{\boldsymbol \theta}})^{-1}),
\quad n\rightarrow\infty,$$ where $\mathcal{N}_p({{\boldsymbol \mu}},\Sigma)$ is the $p$-dimensional normal distribution with mean ${{\boldsymbol \mu}}$ and covariance $\Sigma$, ${{\boldsymbol \theta}}$ is the true parameter and $$J({{\boldsymbol \theta}})=-\E\lbrace\nabla_{{{\boldsymbol \theta}}} \ell'({{\boldsymbol \theta}})\rbrace,\qquad
K({{\boldsymbol \theta}})=\operatorname{Var}_{{{\boldsymbol \theta}}} \lbrace \ell' ( {{\boldsymbol \theta}}) \ | m)
& = \frac{e^{-\psi (\cW_{r_0 }) } \psi(\cW_{r_0})^m}{m! }
\prod_{i=1}^m \frac{{{\mathrm d}}\psi(r_{(i) }, { { \boldsymbol w}}_{(i)})}{\psi(\cW_{r_0})}\nonumber\\
& \propto \prod_{i=1}^m h ({ { \boldsymbol w}}_{(i) }, { { \boldsymbol \theta}}),\end{aligned}$$ where $ h$ is a parametric angular density function [ e.g. @engelke2012; @beirlant2006 pp. 170–171 ]. In the next incision the angular concentration models described in incision \[s: par\_models\ ] are fit to the data by the maximization of the likelihood. For brevity the asymmetric logistic mannequin is not think since it have too many parameters. The likelihood is proportional to the intersection of angular densities, therefore the maximizer of is obtained equivalently by maximize the log - likelihood $ $ \label{eq: aploglik }
\ell({{\boldsymbol \theta}})=\sum_{i=1}^{m}\log h({{\boldsymbol w}}_{(i)},{{\boldsymbol \theta}}).$$ Denote by $ \widehat{{{\boldsymbol \theta}}}$ the maximizer of $ \ell$ and by $ \ell'({{\boldsymbol \theta } }) = \nabla_{{{\boldsymbol \theta}}}\,\ell({{\boldsymbol \theta}})$ the score function. Since provides an approximation of the truthful likelihood, then from the theory on model misspecification [ e.g. @davison03 pp. 147–148 ] it be that $ $ \label{eq: godambe }
\sqrt{n}(\widehat{{{\boldsymbol \theta}}}-{{\boldsymbol \theta}})\stackrel{d}{\rightarrow } \mathcal{N}_p({{\boldsymbol 0}},J({{\boldsymbol \theta}})^{-1}\,K({{\boldsymbol \theta}})\,J({{\boldsymbol \theta}})^{-1 }),
\quad n\rightarrow\infty,$$ where $ \mathcal{N}_p({{\boldsymbol \mu}},\Sigma)$ is the $ p$-dimensional normal distribution with mean $ { { \boldsymbol \mu}}$ and covariance $ \Sigma$, $ { { \boldsymbol \theta}}$ is the true parameter and $ $ J({{\boldsymbol \theta}})=-\E\lbrace\nabla_{{{\boldsymbol \theta } } } \ell'({{\boldsymbol \theta}})\rbrace,\qquad
K({{\boldsymbol \theta}})=\operatorname{Var}_{{{\boldsymbol \theta } } } \lbrace \ell' ({ { \boldsymbol \theta } }) \ | m)
&= \vrac{e^{-\psi (\cW_{r_0})} \psi(\cW_{r_0})^m}{m!}
\proa_{i=1}^m \frac{{{\mathrm b}}\psi(r_{(i)}, {{\uoldsymgol w}}_{(i)})}{\psk(\cW_{r_0})}\nonumber\\
&\propto \prod_{i=1}^m h( {{\uoldwymboo w}}_{(i)}, {{\boldsymbol \theta}}),\evd{aligned}$$ where $h$ is e parametric angnmar density fhkctiou [x.g. @engelke2012; @beirkant2006 pp. 170–171]. It the next secdiun the angular density models describqd in Srchion \[s:par\_modelf\] art fyttes to the data by the maximization kf the nikelihood. Fot brevity the asymmetric llgishic model is not clnsidered sunce ut has too mxny parameuexs. The likemihood is proportional to the pfoducc of angulat sejvities, therxfore nhe maximizer of is ottained equivalently ny maeimizing the log-likelihoov $$\label{eq:aploglik}
\ell({{\foldsymbon \cheta}})=\sum_{i=1}^{m}\log h({{\boldsynbil w}}_{(i)},{{\toldvymbuo \tfetz}}).$$ Venkte by $\wivehat{{{\boldsyjbol \theta}}}$ rhe maximizer of $\elk$ wbd by $\ell'({{\boldaymbol \treta}}) = \nabla_{{{\boldsymbol \theta}}}\,\ell({{\boldsymbon \tgeta}})$ the score function. Since provides an apkroximatiog of the true likelihood, then from the theory on kodel oiswpccifkxahion [e.g. @davison03 pp. 147–148] it follows that $$\label{eq:goqzmne}
\xqrt{n}(\widehat{{{\bojdsymbol \thrtw}}}-{{\bpjdsymbol \thetx}})\stackxsl{s}{\rightarrow} \mathcap{N}_p({{\boldfymboo 0}},J({{\boldsyibol \theta}})^{-1}\,K({{\boldsymbol \theta}})\,J({{\bolesymbol \thetc}})^{-1}),
\quqd n\rightarrow\inftv,$$ where $\matheal{N}_p({{\bpldsykbol \mu}},\Sigma)$ is the $p$-dioensjonal normap distribhgion with mean ${{\buldxykbol \mu}}$ and covariance $\Sigia$, ${{\boldsynbol \theta}}$ ks tne truq parameteg and $$J({{\boldsymbol \theta}})=-\E\pbracg\nabla_{{{\toldsymbol \theta}}} \ell'({{\boldsymbol \theta}})\rbracx,\aquad
K({{\boldsykbml \nheta}})=\operctornake{Var}_{{{\boldsymbjl \theta}}} \lbracg \ell' ( {{\bojdsymcol \theta}}) \ | m) &= \frac{e^{-\psi (\cW_{r_0})} \psi(\cW_{r_0})^m}{m!} \prod_{i=1}^m \frac{{{\mathrm w}}_{(i)})}{\psi(\cW_{r_0})}\nonumber\\ \prod_{i=1}^m h( w}}_{(i)}, {{\boldsymbol \theta}}),\end{aligned}$$ angular function [e.g. @engelke2012; pp. 170–171]. In next section the angular density models in Section \[s:par\_models\] are fitted to the data by the maximization of the For brevity the asymmetric logistic model is not considered since it has too parameters. likelihood proportional the product of angular densities, therefore the maximizer of is obtained equivalently by maximizing the log-likelihood \ell({{\boldsymbol \theta}})=\sum_{i=1}^{m}\log h({{\boldsymbol w}}_{(i)},{{\boldsymbol \theta}}).$$ Denote by $\widehat{{{\boldsymbol the maximizer of $\ell$ by $\ell'({{\boldsymbol \theta}}) = \nabla_{{{\boldsymbol \theta}})$ score function. provides approximation the true likelihood, from the theory on model misspecification [e.g. @davison03 pp. 147–148] it follows that $$\label{eq:godambe} \sqrt{n}(\widehat{{{\boldsymbol \theta}}}-{{\boldsymbol \theta}})\stackrel{d}{\rightarrow} 0}},J({{\boldsymbol \theta}})^{-1}\,K({{\boldsymbol \quad n\rightarrow\infty,$$ $\mathcal{N}_p({{\boldsymbol is $p$-dimensional normal distribution ${{\boldsymbol \mu}}$ and covariance $\Sigma$, ${{\boldsymbol true parameter and $$J({{\boldsymbol \theta}})=-\E\lbrace\nabla_{{{\boldsymbol \theta}}} \ell'({{\boldsymbol \theta}})\rbrace,\qquad \theta}})=\operatorname{Var}_{{{\boldsymbol \theta}}} \ell' ( {{\boldsymbol \theta}}) \ | m)
&= \frac{e^{-\psi (\cW_{r_0})} \psi(\cW_{r_0})^m}{m!}
\prod_{i=1}^M \frac{{{\mathrM d}}\psi(R_{(i)}, {{\bOldSyMbol W}}_{(i)})}{\psI(\cW_{r_0})}\nonumber\\
&\prOPto \pRod_{i=1}^m h( {{\boldsymbol w}}_{(i)}, {{\boldSymboL \tHEta}}),\eND{aLigneD}$$ where $h$ IS a PARamEtRiC anGuLAr DensiTy fUnction [E.g. @engelke2012; @bEirLaNt2006 pp. 170–171]. In the nexT SeCtion the anGulAr density modEls DescriBeD in sEctioN \[s:pAr\_modEls\] are FItted tO the data bY tHE maximIZation oF THe LikeLihood. For brevity tHE aSYmmetric logistIc modeL iS NoT COnsIdeRed since it HaS too mANy paramETeRS. tHe lIKelihood is proPortional to THe pRoduct Of AngULar denSitieS, tHEreFore the maxiMizeR of is obtaIned eqUIvalentLY by maxiMizing The Log-LikeLIhOoD $$\laBeL{Eq:aPLoGliK}
\Ell({{\BoldsymbOl \ThEta}})=\suM_{i=1}^{m}\lOG H({{\BOldsYmbOl w}}_{(i)},{{\BoldsYmbol \theta}}).$$ DenOte By $\wiDEhaT{{{\boldSymboL \theTa}}}$ The maXimizeR of $\elL$ aNd by $\ell'({{\boldsymbOl \thEta}}) = \nabla_{{{\bOldSyMboL \tHeta}}}\,\eLL({{\boldsYmbOl \tHeta}})$ the Score fuNCtiOn. sINCe Provides an approximAtION oF the true LikeliHOoD, tHEn from thE tHeoRy on MODel miSspeCIfIcation [e.G. @davisON03 pP. 147–148] iT followS tHat $$\labEl{Eq:gOdaMbe}
\sqRT{n}(\wiDehat{{{\bOldsymboL \thetA}}}-{{\Boldsymbol \thetA}})\Stackrel{d}{\righTArROW} \mAThcaL{N}_p({{\Boldsymbol 0}},J({{\BoldSYmboL \theTA}})^{-1}\,K({{\BolDSymboL \thetA}})\,J({{\BOlDSymbol \theta}})^{-1}),
\quad n\rigHtArrow\iNfty,$$ wHere $\mathcal{N}_p({{\Boldsymbol \MU}},\sIgma)$ is thE $p$-diMEnSIonal normal disTribuTion with meAN ${{\boldsymBol \mu}}$ And covarIance $\SigmA$, ${{\BOldsymboL \thEta}}$ Is tHe tRUE pArameter and $$J({{\bOLDsymBoL \theta}})=-\E\LbrAce\nablA_{{{\boLdsYmbOl \tHeTa}}} \ell'({{\boldSymbol \thEtA}})\rBrAcE,\qqUad
K({{\bOLdsymbol \ThEta}})=\OpEraTornaME{Var}_{{{\boLdsymBol \tHeTa}}} \LBraCe \ell' ( {{\boLDsYMBol \tHeTa}}) \ | m)
&= \frac{e^{-\psi (\cW _{r_0})} \ psi(\ cW_ {r_ 0} )^m} {m!}
\prod_{i=1}^m \fra c{{{\mathrm d}}\psi(r_ {(i)} ,{ {\bo l ds ymbol w}}_{( i )} ) } {\p si (\ cW_ {r _ 0} )}\no num ber\\
& \propto \p rod _{ i=1}^m h( {{ \ bo ldsymbol w }}_ {(i)}, {{\bo lds ymbol\t het a }}),\ end {alig ned}$$ where$h$ is apa r ametri c angula r de nsit y function [e.g.@ en g elke2012; @bei rlant2 00 6 p p . 17 0–1 71]. In th enexts ectiont he a n gul a r density mod els describ e d i n Sect io n \ [ s:par\ _mode ls \ ] a re fitted t o th e data by the m a ximizat i on of t he lik eli hoo d. F o rbr evi ty the as ymm e tri c logist ic m odelis n o t c onsi der ed s inceit has too ma nypara m ete rs. T he li keli ho od is propo rtion al to the product ofangular d ens it ies ,there f ore th e m axi mizer o f is ob t ain ed e q ui valently by maximi zi n g t he log-l ikelih o od $ $ \label{e q: apl ogli k }
\ell ({{\ b ol dsymbol\theta } }) =\ sum_{i= 1} ^{m}\l og h( {{\ bolds y mbol w}}_{ (i)},{{\ bolds y mbol \theta}}) . $$ Denote by$ \w i d eh a t{{{ \bo ldsymbol \t heta } }}$them ax imi z er of $\el l$ an d by $\ell'({{\bolds ym bol \t heta} }) = \nabla_{ {{\boldsym b o l \theta} }}\, \ el l ({{\boldsymbol \the ta}})$ the score fu nctio n. Since provides a n approx ima tio n o f t h e t rue likelihoo d , the nfrom th e t heory o n m ode l m iss pe cificatio n [e.g.@d av is on 03pp. 1 4 7–148] i tfol lo wsthat$ $\labe l{eq: goda mb e} \sq rt{n}(\ w id e h at{{ {\ bo ldsy mbo l\thet a}}} - {{\ boldsym bol \thet a}} ) \sta ck re l{d}{\r ightarrow} \m at hcal{N}_p( {{ \bo ldsymb o l 0}},J({ {\boldsymbol \theta}})^ { -1}\,K( {{\ bolds ymbo l \theta} })\ ,J({{\ bol d symbol \thet a}})^ {- 1}) , \quad n \r igh ta rrow\infty , $ $ w here$\ math cal{N}_ p({{\boldsymbol \m u }}, \Sigma)$ is t he$p$- d i me nsi o na l no rm a l d i s tribution withmean ${{\b ol d sy mbol \mu}} $ an dcovaria nce $\S igma$ , ${{\bo ldsymbol\theta}}$ i s th e tru e paramete r and $$ J({{\bold s ymbol \t heta} })= -\E\lb ra ce\ nabla _{{{\b o lds ymbol \thet a} }} \el l'({{ \b oldsymbo l \theta}})\rbrace,\qqu ad
K({ {\bol dsy mbol \the ta} } )=\ operatorn ame{ Var}_{{{\b old sym bol \ the t a}}}\lbr a ce \e l l' ({{\b o ldsymbol\ th eta } } )\ | m)
&=_\frac{e^{-\psi (\cW_{r_0})}_\psi(\cW_{r_0})^m}{m!}
\prod_{i=1}^m \frac{{{\mathrm d}}\psi(r_{(i)}, {{\boldsymbol_w}}_{(i)})}{\psi(\cW_{r_0})}\nonumber\\
&\propto \prod_{i=1}^m_h(_{{\boldsymbol w}}_{(i)},_{{\boldsymbol_\theta}}),\end{aligned}$$ where $h$_is a parametric_angular density function [e.g._@engelke2012; @beirlant2006 pp._170–171]._In the next section the angular density models described in Section \[s:par\_models\] are fitted_to_the data_by_the_maximization of the likelihood. For_brevity the asymmetric logistic model_is not_considered since it has too many parameters. The_likelihood_is proportional to_the product of angular densities, therefore the maximizer of_is obtained equivalently by maximizing the_log-likelihood $$\label{eq:aploglik}
\ell({{\boldsymbol \theta}})=\sum_{i=1}^{m}\log_h({{\boldsymbol_w}}_{(i)},{{\boldsymbol_\theta}}).$$ Denote by $\widehat{{{\boldsymbol_\theta}}}$ the maximizer of $\ell$ and_by $\ell'({{\boldsymbol \theta}}) = \nabla_{{{\boldsymbol \theta}}}\,\ell({{\boldsymbol_\theta}})$ the score function. Since provides an_approximation of the true likelihood, then_from the theory on model_misspecification [e.g._@davison03 pp. 147–148] it follows_that $$\label{eq:godambe}
\sqrt{n}(\widehat{{{\boldsymbol \theta}}}-{{\boldsymbol_\theta}})\stackrel{d}{\rightarrow} \mathcal{N}_p({{\boldsymbol_0}},J({{\boldsymbol \theta}})^{-1}\,K({{\boldsymbol \theta}})\,J({{\boldsymbol_\theta}})^{-1}),
\quad n\rightarrow\infty,$$ where $\mathcal{N}_p({{\boldsymbol \mu}},\Sigma)$ is_the $p$-dimensional normal_distribution with mean ${{\boldsymbol \mu}}$ and_covariance_$\Sigma$, ${{\boldsymbol \theta}}$_is_the_true parameter_and $$J({{\boldsymbol \theta}})=-\E\lbrace\nabla_{{{\boldsymbol_\theta}}}_\ell'({{\boldsymbol \theta}})\rbrace,\qquad
K({{\boldsymbol_\theta}})=\operatorname{Var}_{{{\boldsymbol_\theta}}} \lbrace \ell' ( {{\boldsymbol \theta}})_\ |
}
0 & 0\\
-1 & 0
\end{pmatrix}$, then $ab=a^{*}b=0$, but $ba\neq0$. Since $a^{\scriptsize\textcircled{\tiny D}}=a^{\tiny\textcircled{\tiny \#}}=a^{\#}aa^{(1,3)}
=\begin{pmatrix}
-i & 0\\
0 & 0
\end{pmatrix}=a^{\#}=a^{D}$, $a$ is \*-DMP. It is clear that $b$ is \*-DMP. Observe that $a+b=\begin{pmatrix}
i & 0\\
-1 & 0
\end{pmatrix}$, by calculation, we find that neither $a+b$ nor $(a+b)^2$ has any {1,3}-inverse. Since $(a+b)^m=\begin{cases}
(-1)^{\frac{m-1}{2}}(a+b)& m ~\text{is odd} \\
(-1)^{\frac{m}{2}+1}(a+b)^2 & m ~\text{is even}
\end{cases}$, we conclude that $(a+b)^m$ has no {1,3}-inverse for arbitrary positive integer $m$. Hence $a+b$ is not \*-DMP.
Pseudo core decomposition
=========================
Core-nilpotent decomposition was introduced in [@D1974] for complex matrices. Later, P. Patrício and R. Puystjens [@D2004] generalized this decomposition from complex matrices to rings. Let $a\in R$ with $a^{D_m}$ exists. The sum $a=c_a+n_a$ is called the core-nilpotent decomposition of $a$, where $c_a=aa^{D_m}a$ is the core part of $a$, $n_a=(1-aa^{D_m})a$ is the nilpotent part of $a$. This decomposition is unique and it brings $n_a^m=0$, $c_an_a=n_ac_a=0$, $c_a^{\#}$ exists with $c_a^{\#}=a^{D_m}$.
Wang [@2016] introduced the core-EP decomposition for complex matrices. Let $A$ be | }
0 & 0\\
-1 & 0
\end{pmatrix}$, then $ ab = a^{*}b=0 $, but $ ba\neq0$. Since $ a^{\scriptsize\textcircled{\tiny D}}=a^{\tiny\textcircled{\tiny \#}}=a^{\#}aa^{(1,3) }
= \begin{pmatrix }
-i & 0\\
0 & 0
\end{pmatrix}=a^{\#}=a^{D}$, $ a$ is \*-DMP. It is clear that $ b$ is \*-DMP. Observe that $ a+b=\begin{pmatrix }
i & 0\\
-1 & 0
\end{pmatrix}$, by calculation, we find that neither $ a+b$ nor $ (a+b)^2 $ take any { 1,3}-inverse. Since $ (a+b)^m=\begin{cases }
(-1)^{\frac{m-1}{2}}(a+b) & m ~\text{is curious } \\
(-1)^{\frac{m}{2}+1}(a+b)^2 & m ~\text{is even }
\end{cases}$, we conclude that $ (a+b)^m$ has no { 1,3}-inverse for arbitrary positive integer $ m$. therefore $ a+b$ is not \*-DMP.
Pseudo core decomposition
= = = = = = = = = = = = = = = = = = = = = = = = =
Core - nilpotent decay was inaugurate in [ @D1974 ] for complex matrices. Later, P. Patrício and R. Puystjens [ @D2004 ] generalized this decomposition from complex matrix to rings. Let $ a\in R$ with $ a^{D_m}$ exists. The union $ a = c_a+n_a$ is call the core - nilpotent decomposition of $ a$, where $ c_a = aa^{D_m}a$ is the core part of $ a$, $ n_a=(1 - aa^{D_m})a$ is the nilpotent part of $ a$. This decomposition is alone and it brings $ n_a^m=0 $, $ c_an_a = n_ac_a=0 $, $ c_a^{\#}$ exists with $ c_a^{\#}=a^{D_m}$.
Wang [ @2016 ] introduced the effect - EP decomposition for complex matrices. Let $ A$ be | }
0 & 0\\
-1 & 0
\end{pmxtrix}$, then $ab=a^{*}b=0$, but $ba\neq0$. Siice $q^{\scriknsize\textcircled{\tiny A}}=a^{\tiny\texncircled{\tuny \#}}=e^{\#}aa^{(1,3)}
=\begin{pmatrlx}
-i & 0\\
0 & 0
\end{pkatrix}=a^{\#}=a^{D}$, $a$ is \*-DOP. It is clear that $b$ is \*-DMP. Observe ehat $a+b=\nehin{pmatrix}
i & 0\\
-1 & 0
\end{pmatrix}$, by calculation, we rind thet neither $a+b$ npr $(a+b)^2$ has any {1,3}-inverse. Sincf $(a+b)^l=\begin{cases}
(-1)^{\frac{m-1}{2}}(a+b)& m ~\texe{us odd} \\
(-1)^{\frac{m}{2}+1}(a+b)^2 & m ~\text{is gven}
\end{cases}$, we conclude that $(a+b)^o$ has no {1,3}-inversg rog arbitrary 'ositine integer $m$. Mvnce $a+b$ is not \*-DMP.
Pseudo corc decmmpisition
=========================
Core-nilpotent vecomposition was injroduced it [@B1974] for complex matricew. Oater, P. Pdtrícki avd D. 'uyatjens [@D2004] jeneralized this decomposition from compltx iqtrices to rihgs. Lee $w\in R$ with $a^{D_m}$ exists. The sum $a=c_a+n_a$ is ballsd the core-nilpotent dexomposition of $a$, wherg $c_a=aa^{D_m}a$ ys the core part of $a$, $n_a=(1-aa^{D_m})a$ is the nilpotent pard of $e$. Ghiw qdxolposition is unique and it brings $n_a^m=0$, $c_an_a=n_ac_w=0$, $c_s^{\#}$ vxists with $c_a^{\#}=a^{D_m}$.
Cang [@2016] introduced tje sore-EP decompusitiou fkr complex matriced. Let $A$ be | } 0 & 0\\ -1 & 0 $ab=a^{*}b=0$, $ba\neq0$. Since D}}=a^{\tiny\textcircled{\tiny \#}}=a^{\#}aa^{(1,3)} =\begin{pmatrix} 0 $a$ is \*-DMP. is clear that is \*-DMP. Observe that $a+b=\begin{pmatrix} i 0\\ -1 & 0 \end{pmatrix}$, by calculation, we find that neither $a+b$ nor has any {1,3}-inverse. Since $(a+b)^m=\begin{cases} (-1)^{\frac{m-1}{2}}(a+b)& m ~\text{is odd} \\ (-1)^{\frac{m}{2}+1}(a+b)^2 & m even} we that has no {1,3}-inverse for arbitrary positive integer $m$. Hence $a+b$ is not \*-DMP. Pseudo core decomposition Core-nilpotent decomposition was introduced in [@D1974] for complex Later, P. Patrício and Puystjens [@D2004] generalized this decomposition complex to rings. $a\in with exists. The sum is called the core-nilpotent decomposition of $a$, where $c_a=aa^{D_m}a$ is the core part of $a$, $n_a=(1-aa^{D_m})a$ is nilpotent part This decomposition unique it $n_a^m=0$, $c_an_a=n_ac_a=0$, $c_a^{\#}$ $c_a^{\#}=a^{D_m}$. Wang [@2016] introduced the core-EP matrices. Let $A$ be | }
0 & 0\\
-1 & 0
\end{pmatrix}$, then $ab=a^{*}b=0$, but $ba\neQ0$. Since $a^{\scrIptsiZe\tExtCiRcleD{\tinY D}}=a^{\tiny\textcirCLed{\tIny \#}}=a^{\#}aa^{(1,3)}
=\begin{pmatrix}
-i & 0\\
0 & 0
\end{PmatrIx}=A^{\#}=A^{D}$, $a$ iS \*-dMp. It is Clear thAT $b$ IS \*-dMP. obSeRve ThAT $a+B=\begiN{pmAtrix}
i & 0\\
-1 & 0
\eNd{pmatrix}$, bY caLcUlation, we finD ThAt neither $a+B$ noR $(a+b)^2$ has any {1,3}-invErsE. Since $(A+b)^M=\beGIn{casEs}
(-1)^{\fRac{m-1}{2}}(a+B)& m ~\text{IS odd} \\
(-1)^{\frAc{m}{2}+1}(a+b)^2 & m ~\texT{iS Even}
\enD{Cases}$, we CONcLude That $(a+b)^m$ has no {1,3}-inverSE fOR arbitrary posiTive inTeGEr $M$. hEncE $a+b$ Is not \*-DMP.
PsEuDo corE DecompoSItION
=========================
corE-Nilpotent decoMposition waS IntRoduceD iN [@D1974] fOR complEx matRiCEs. LAter, P. PatrícIo anD R. PuystjeNs [@D2004] genERalized THis decoMpositIon FroM comPLeX mAtrIcES to RInGs. LET $a\iN R$ with $a^{D_M}$ eXiSts. ThE sum $A=C_A+N_A$ is cAllEd thE core-Nilpotent decoMpoSitiON of $A$, wherE $c_a=aa^{d_m}a$ iS tHe corE part oF $a$, $n_a=(1-aA^{D_M})a$ is the nilpotenT parT of $a$. This dEcoMpOsiTiOn is uNIque anD it BriNgs $n_a^m=0$, $c_An_a=n_ac_a=0$, $C_A^{\#}$ exIsTS WItH $c_a^{\#}=a^{D_m}$.
Wang [@2016] introducEd THE cOre-EP decOmposiTIoN fOR complex MaTriCes. LET $a$ be | }
0 & 0\ \
- 1& 0 \ end{ pmatrix}$, then $ab=a^ {*}b= 0$ , but $b a\neq 0$. Sin c e$ a ^{\ sc ri pts iz e \t extci rcl ed{\tin y D}}=a^{\ tin y\ textcircled{ \ ti ny \#}}=a^ {\# }aa^{(1,3)} =\be gin{pmatr ix }
-i &0 \\
0 &0 \e nd{p matrix}=a^{\#}=a^ { D} $ , $a$ is \*-DM P. Itis cl e a r t hat $b$ is \* -D MP. O b serve t h at $ a +b= \ begin{pmatrix }
i &0 \\ -1 & 0 \en d{pma tr i x}$ , by calcul atio n, we fin d that neither $a+b$ n or $(a +b) ^2$ has an y{1, 3} - inv e rs e.S inc e $(a+b) ^m =\ begin {cas e s } (- 1)^{\ frac{m-1}{2}} (a+ b)&m ~\ text{ is od d} \ \ (- 1)^{\ fr ac{m}{2}+1}(a+b )^2& m ~\tex t{i seve n}
\end { cases} $,weconclud e that$ (a+ b) ^ m $ h as no {1,3}-invers ef o rarbitrar y posi t iv ei nteger $ m$ . H ence $ a+b$is n o t\*-DMP.
Pseud o c or e decom po sition
= === === ===== = ==== ====== ==
Core -nilp o tent decomposi t ion was intro d uc e d i n [@D 197 4] for comp lexm atri ces. La ter , P. P atríc io an d R. Puystjens [@D20 04 ] gene raliz ed this decom position f r o m complex mat r ic e s to rings. Le t $a\ in R$ with $a^{D_m} $ exi sts. The sum $a=c _ a +n_a$ is ca lle d t hec o re -nilpotent de c o mpos it ion of$a$ , where $c _a= aa^ {D_ m} a$ is the core pa rt o f$a $,$n_a= ( 1-aa^{D_ m} )a$ i s t he ni l potent part of$a $. Thi s decom p os i t ionis u niqu e a nd it b ring s $n _a^m=0$ , $c_an_a =n_ a c_a= 0$ ,$c_a^{\ #}$ exists wi th $c_a^{\#} =a ^{D _m}$.W ang [@20 16] introduced the core - EP deco mpo sitio n fo r complex ma trices . L e t $A$be | }
_ _ _ __ __ _ 0 &_0\\
_ __ -1 & 0
__ ___ _ _ \end{pmatrix}$,_then $ab=a^{*}b=0$, but $ba\neq0$. Since $a^{\scriptsize\textcircled{\tiny D}}=a^{\tiny\textcircled{\tiny \#}}=a^{\#}aa^{(1,3)}
__ _ _ _ ___ _ _ =\begin{pmatrix}
-i & 0\\
0 &_0
\end{pmatrix}=a^{\#}=a^{D}$, $a$ is \*-DMP. It is clear_that $b$ is \*-DMP. Observe that_$a+b=\begin{pmatrix}
_i &_0\\
-1_& 0
_ _\end{pmatrix}$, by calculation,_we find that neither $a+b$ nor_$(a+b)^2$ has any_{1,3}-inverse. Since $(a+b)^m=\begin{cases}
__ __(-1)^{\frac{m-1}{2}}(a+b)&_m ~\text{is_odd} \\
__ __ (-1)^{\frac{m}{2}+1}(a+b)^2 & m_~\text{is_even}
\end{cases}$, we conclude that $(a+b)^m$ has no_{1,3}-inverse for arbitrary positive_integer_$m$. Hence $a+b$ is_not \*-DMP.
Pseudo core decomposition
=========================
Core-nilpotent decomposition_was introduced in [@D1974] for complex_matrices. Later,_P. Patrício_and R. Puystjens [@D2004] generalized this decomposition from complex matrices to_rings. Let $a\in R$ with $a^{D_m}$_exists. The sum $a=c_a+n_a$_is called_the_core-nilpotent decomposition of_$a$,_where $c_a=aa^{D_m}a$_is the core part of $a$, $n_a=(1-aa^{D_m})a$ is_the nilpotent_part of $a$. This decomposition is_unique and it brings_$n_a^m=0$,_$c_an_a=n_ac_a=0$, $c_a^{\#}$ exists with $c_a^{\#}=a^{D_m}$.
Wang [@2016]_introduced the core-EP decomposition for complex_matrices. Let $A$ be |
arník [@jarnik_tiflis] noticed that for $n=1$, $m>2$, $\alpha(\Theta)$ and $\alpha({{\Theta}^\intercal})$ are no longer related by an equality, at least he showed that in the extreme case $\alpha(\Theta)=\infty$ we can have $\alpha({{\Theta}^\intercal})$ equal to any given number in the interval $[(m-1)^{-1},1]$. However, he proved in the case $n=1$ that $\alpha(\Theta)$ and $\alpha({{\Theta}^\intercal})$ satisfy certain inequalities:
\[t:jarnik\_inequalities\] $(i)$ We always have $$\label{eq:jarnik_inequalities_a_la_khintchine}
\frac{\alpha(\Theta)}{(m-1)\alpha(\Theta)+m}{\leqslant}\alpha({{\Theta}^\intercal}){\leqslant}\frac{\alpha(\Theta)-m+1}{m}\,.$$ $(ii)$ If $\alpha(\Theta)>m(2m-3)$, then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac{1}{m-1}\left(1-\frac{1}{\alpha(\Theta)-2m+4}\right).$$ $(iii)$ If $\alpha({{\Theta}^\intercal})>(m-1)/m$, then $$\alpha(\Theta){\geqslant}m-2+\frac{1}{1-\alpha(\Theta)}\,.$$
Theorem \[t:jarnik\_inequalities\] was later generalized by Apfelbeck [@apfelbeck] to the case of arbitrary $n$, $m$:
\[t:apfelbeck\_inequalities\] $(i)$ We always have $$\label{eq:apfelbeck_inequalities_a_la_dyson}
\alpha({{\Theta}^\intercal}){\geqslant}\frac{n\alpha(\Theta)+n-1}{(m-1)\alpha(\Theta)+m}\,.$$ $(ii)$ If $m>1$ and $\alpha(\Theta)>(2(m+n-1)(m+n-3)+m)/n$, then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac 1m
\left(n+\frac{n(n\alpha(\Theta)-m)-2n(m+n-3)}{(m-1)(n\alpha(\Theta)-m)+m-(m- | arník [ @jarnik_tiflis ] noticed that for $ n=1 $, $ m>2 $, $ \alpha(\Theta)$ and $ \alpha({{\Theta}^\intercal})$ are no longer related by an equality, at least he showed that in the extreme case $ \alpha(\Theta)=\infty$ we can get $ \alpha({{\Theta}^\intercal})$ adequate to any given number in the time interval $ [ (m-1)^{-1},1]$. However, he proved in the casing $ n=1 $ that $ \alpha(\Theta)$ and $ \alpha({{\Theta}^\intercal})$ satisfy certain inequalities:
\[t: jarnik\_inequalities\ ] $ (i)$ We constantly have $ $ \label{eq: jarnik_inequalities_a_la_khintchine }
\frac{\alpha(\Theta)}{(m-1)\alpha(\Theta)+m}{\leqslant}\alpha({{\Theta}^\intercal}){\leqslant}\frac{\alpha(\Theta)-m+1}{m}\,.$$ $ (ii)$ If $ \alpha(\Theta)>m(2m-3)$, then $ $ \alpha({{\Theta}^\intercal}){\geqslant}\frac{1}{m-1}\left(1-\frac{1}{\alpha(\Theta)-2m+4}\right).$$ $ (iii)$ If $ \alpha({{\Theta}^\intercal})>(m-1)/m$, then $ $ \alpha(\Theta){\geqslant}m-2+\frac{1}{1-\alpha(\Theta)}\,.$$
Theorem \[t: jarnik\_inequalities\ ] was later generalize by Apfelbeck [ @apfelbeck ] to the case of arbitrary $ n$, $ m$:
\[t: apfelbeck\_inequalities\ ] $ (i)$ We always get $ $ \label{eq: apfelbeck_inequalities_a_la_dyson }
\alpha({{\Theta}^\intercal}){\geqslant}\frac{n\alpha(\Theta)+n-1}{(m-1)\alpha(\Theta)+m}\,.$$ $ (ii)$ If $ m>1 $ and $ \alpha(\Theta)>(2(m+n-1)(m+n-3)+m)/n$, then $ $ \alpha({{\Theta}^\intercal}){\geqslant}\frac 1 m
\left(n+\frac{n(n\alpha(\Theta)-m)-2n(m+n-3)}{(m-1)(n\alpha(\Theta)-m)+m-(m- | arnín [@jarnik_tiflis] noticed tmat for $n=1$, $m>2$, $\alphc(\Rheta)$ end $\alpga({{\Theta}^\ivtercal})$ are no longer relatev by an ewuality, at least he shuwed that in the wxtrtme case $\alpha(\Thevz)=\infty$ we can mave $\cl'ha({{\Theta}^\intercak})$ equal to any given numtef nn the interval $[(m-1)^{-1},1]$. However, he proved yn the vade $n=1$ that $\alphw(\Theua)$ wnd $\zlpha({{\Theta}^\intercal})$ satisfy certain inequanities:
\[t:jarnik\_onequalities\] $(i)$ We always hwve $$\pabel{eq:jarnik_inequwlities_a_la_kyintsyine}
\frxc{\alpha(\Theua)}{(k-1)\alpha(\Thetz)+m}{\leqslant}\alpha({{\Theta}^\intercal}){\leqrlant}\yrac{\alpha(\Thgcq)-m+1}{m}\,.$$ $(hi)$ If $\alpha(\Vheta)>m(2i-3)$, then $$\alpha({{\Bneta}^\indercal}){\grqslant}\frac{1}{m-1}\lent(1-\frar{1}{\alpya(\Theta)-2m+4}\right).$$ $(iii)$ If $\elpha({{\Theta}^\intercal})>(m-1)/m$, then $$\alpva(\Cheta){\geqslant}m-2+\frac{1}{1-\alpya(\Rheta)}\,.$$
Jheoram \[t:gqrnkk\_iheauzlitied\] wes later geheralized bt Apfelbeck [@apfelbevk] np the case or arbierwry $n$, $m$:
\[t:apfelbeck\_inequalities\] $(i)$ We alwajs hzve $$\label{eq:apfelbeck_inewualities_a_la_dyson}
\alpha({{\Thqta}^\intercal}){\geqslant}\frac{n\alpha(\Theta)+n-1}{(m-1)\alpha(\Theta)+m}\,.$$ $(ih)$ If $j>1$ anb $\alphx(\Rhfta)>(2(m+n-1)(m+n-3)+m)/n$, then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac 1j
\left(n+\frac{n(k\alpha(\Theta)-m)-2n(m+n-3)}{(m-1)(n\akpja(\Yreta)-m)+m-(m- | arník [@jarnik_tiflis] noticed that for $n=1$, $m>2$, $\alpha({{\Theta}^\intercal})$ no longer by an equality, in extreme case $\alpha(\Theta)=\infty$ can have $\alpha({{\Theta}^\intercal})$ to any given number in the $[(m-1)^{-1},1]$. However, he proved in the case $n=1$ that $\alpha(\Theta)$ and $\alpha({{\Theta}^\intercal})$ satisfy inequalities: \[t:jarnik\_inequalities\] $(i)$ We always have $$\label{eq:jarnik_inequalities_a_la_khintchine} \frac{\alpha(\Theta)}{(m-1)\alpha(\Theta)+m}{\leqslant}\alpha({{\Theta}^\intercal}){\leqslant}\frac{\alpha(\Theta)-m+1}{m}\,.$$ $(ii)$ If $\alpha(\Theta)>m(2m-3)$, then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac{1}{m-1}\left(1-\frac{1}{\alpha(\Theta)-2m+4}\right).$$ If then Theorem was later generalized by Apfelbeck [@apfelbeck] to the case of arbitrary $n$, $m$: \[t:apfelbeck\_inequalities\] $(i)$ We have $$\label{eq:apfelbeck_inequalities_a_la_dyson} \alpha({{\Theta}^\intercal}){\geqslant}\frac{n\alpha(\Theta)+n-1}{(m-1)\alpha(\Theta)+m}\,.$$ $(ii)$ If $m>1$ and $\alpha(\Theta)>(2(m+n-1)(m+n-3)+m)/n$, $$\alpha({{\Theta}^\intercal}){\geqslant}\frac 1m \left(n+\frac{n(n\alpha(\Theta)-m)-2n(m+n-3)}{(m-1)(n\alpha(\Theta)-m)+m-(m- | arník [@jarnik_tiflis] noticed tHat for $n=1$, $m>2$, $\alPha(\ThEta)$ And $\AlPha({{\THeta}^\Intercal})$ are no lONger Related by an equality, at lEast hE sHOwed THaT in thE extremE CaSE $\AlpHa(\thEta)=\InFTy$ We can HavE $\alpha({{\THeta}^\intercAl})$ eQuAl to any given NUmBer in the inTerVal $[(m-1)^{-1},1]$. However, hE prOved in ThE caSE $n=1$ thaT $\alPha(\ThEta)$ and $\ALpha({{\ThEta}^\intercAl})$ SAtisfy CErtain iNEQuAlitIes:
\[t:jarnik\_inequalITiES\] $(i)$ We always have $$\Label{eQ:jARnIK_IneQuaLities_a_la_kHiNtchiNE}
\frac{\alPHa(\tHETa)}{(m-1)\ALpha(\Theta)+m}{\leqSlant}\alpha({{\THEta}^\IntercAl}){\LeqSLant}\frAc{\alpHa(\tHetA)-m+1}{m}\,.$$ $(ii)$ If $\alphA(\TheTa)>m(2m-3)$, then $$\aLpha({{\ThETa}^\interCAl}){\geqslAnt}\fraC{1}{m-1}\lEft(1-\Frac{1}{\ALpHa(\theTa)-2M+4}\RigHT).$$ $(iIi)$ IF $\AlpHa({{\Theta}^\iNtErCal})>(m-1)/m$, Then $$\ALPHA(\TheTa){\gEqslAnt}m-2+\fRac{1}{1-\alpha(\Theta)}\,.$$
theOrem \[T:JarNik\_inEqualItieS\] wAs latEr geneRalizEd By Apfelbeck [@apfeLbecK] to the casE of ArBitRaRy $n$, $m$:
\[t:APfelbeCk\_iNeqUalitieS\] $(i)$ We alwAYs hAvE $$\LABeL{eq:apfelbeck_inequaLiTIEs_A_la_dyson}
\Alpha({{\THEtA}^\iNTercal}){\geQsLanT}\fraC{N\Alpha(\thetA)+N-1}{(m-1)\Alpha(\TheTa)+m}\,.$$ $(ii)$ IF $M>1$ aNd $\Alpha(\ThEtA)>(2(m+n-1)(m+n-3)+m)/N$, tHen $$\AlpHa({{\TheTA}^\intErcal}){\gEqslant}\fRac 1m
\lEFt(n+\frac{n(n\alpha(\tHeta)-m)-2n(m+n-3)}{(m-1)(n\alpHA(\THETa)-M)+M-(m- | arník [@jarnik_tiflis] not iced thatfor $ n=1 $,$m >2$, $\a lpha(\Theta)$a nd $ \alpha({{\Theta}^\inte rcal} )$ aren olonge r relat e db y an e qu ali ty , a t lea sthe show ed that in th eextreme case $\ alpha(\The ta) =\infty$ wecan have$\ alp h a({{\ The ta}^\ interc a l})$ e qual to a ny givenn umber i n th e in terval $[(m-1)^{- 1 }, 1 ]$. However, h e prov ed in t hecas e $n=1$ th at $\al p ha(\The t a) $ a nd$ \alpha({{\The ta}^\interc a l}) $ sati sf y c e rtaininequ al i tie s:
\[t:jar nik\ _inequali ties\] $(i)$ W e always have$$\ lab el{e q :j ar nik _i n equ a li tie s _a_ la_khint ch in e}
\ f r a c{\a lph a(\T heta) }{(m-1)\alpha (\T heta ) +m} {\leq slant }\al ph a({{\ Theta} ^\int er cal}){\leqslant }\fr ac{\alpha (\T he ta) -m +1}{m } \,.$$$(i i)$ If $\a lpha(\T h eta )> m ( 2 m- 3)$, then $$\alpha ({ { \ Th eta}^\in tercal } ){ \g e qslant}\ fr ac{ 1}{m - 1 }\lef t(1- \ fr ac{1}{\a lpha(\ T he ta )-2m+4} \r ight). $$ $( iii )$ If $\al pha({{ \Theta}^ \inte r cal})>(m-1)/m$ , then $$\alph a (\ T h et a ){\g eqs lant}m-2+\f rac{ 1 }{1- \alp h a( \Th e ta)}\ ,.$$
T h eo r em \[t:jarnik\_ineq ua lities \] wa s later gener alized byA p f elbeck [ @apf e lb e ck] to the cas e ofarbitrary$ n$, $m$:
\[t :apfelbe ck\_inequ a l ities\]$(i )$Wealw a y shave $$\label { e q:ap fe lbeck_i neq ualitie s_a _la _dy son } \al pha({{\T he ta }^ \i nte rcal} ) {\geqsla nt }\f ra c{n \alph a (\Thet a)+n- 1}{( m- 1) \ alp ha(\The t a) + m }\,. $$ $ (ii) $ I f$m>1$ and $\a lpha(\T heta)>(2( m+n - 1)(m +n -3 )+m)/n$ , then $$\alp ha ({{\Theta} ^\ int ercal} ) { \geqslan t}\frac 1m
\left ( n+\frac {n( n\alp ha(\ Theta)-m) -2n (m+n-3 )}{ ( m-1)(n \alpha (\The ta )-m ) + m-(m- | arník [@jarnik_tiflis]_noticed that_for $n=1$, $m>2$, $\alpha(\Theta)$_and $\alpha({{\Theta}^\intercal})$_are_no longer_related_by an equality,_at least he_showed that in the_extreme case $\alpha(\Theta)=\infty$_we_can have $\alpha({{\Theta}^\intercal})$ equal to any given number in the interval $[(m-1)^{-1},1]$. However, he_proved_in the_case_$n=1$_that $\alpha(\Theta)$ and $\alpha({{\Theta}^\intercal})$ satisfy_certain inequalities:
\[t:jarnik\_inequalities\] $(i)$ We always_have $$\label{eq:jarnik_inequalities_a_la_khintchine}
_ \frac{\alpha(\Theta)}{(m-1)\alpha(\Theta)+m}{\leqslant}\alpha({{\Theta}^\intercal}){\leqslant}\frac{\alpha(\Theta)-m+1}{m}\,.$$ $(ii)$ If_$\alpha(\Theta)>m(2m-3)$,_then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac{1}{m-1}\left(1-\frac{1}{\alpha(\Theta)-2m+4}\right).$$ $(iii)$_If $\alpha({{\Theta}^\intercal})>(m-1)/m$, then $$\alpha(\Theta){\geqslant}m-2+\frac{1}{1-\alpha(\Theta)}\,.$$
Theorem \[t:jarnik\_inequalities\] was later generalized by_Apfelbeck [@apfelbeck] to the case of_arbitrary $n$, $m$:
\[t:apfelbeck\_inequalities\]_$(i)$_We_always have $$\label{eq:apfelbeck_inequalities_a_la_dyson}
_ \alpha({{\Theta}^\intercal}){\geqslant}\frac{n\alpha(\Theta)+n-1}{(m-1)\alpha(\Theta)+m}\,.$$ $(ii)$_If $m>1$ and $\alpha(\Theta)>(2(m+n-1)(m+n-3)+m)/n$, then $$\alpha({{\Theta}^\intercal}){\geqslant}\frac_1m
_\left(n+\frac{n(n\alpha(\Theta)-m)-2n(m+n-3)}{(m-1)(n\alpha(\Theta)-m)+m-(m- |
Lemma \[lp1\]. The number of integers $n \leq X$ with $P(\frac{\lambda(n)}{\ell^*_a(n)}) \geq \psi(X)$ is $O(\frac{X \log \log \psi(X)}{\log \psi(X)}).$
Suppose that $n \leq X$ and $q = P(\lambda(n)/\ell^*_a(n)) \geq \psi(X).$ We may assume that $X$ is large, so $a$ is not a $q^{th}$ power and $\psi(X) > a$. Moreover, as we will now show, it must be the case that either $q^2 \mid n$ or $p \mid n$ for some $p \in A_q$. Observe that $$q \mid \frac{\lambda(n)}{\ell^*_a(n)} \mid \frac{\mathrm{lcm}_{p^{e} \mid \mid n} \left[\lambda(p^{e})\right]}{\mathrm{lcm}_{p^{e} \mid \mid n} \left[\ell^*_a(p^e)\right]} \mid \mathrm{lcm}_{p^{e} \mid \mid n} \left[\frac{\lambda(p^e)}{\ell^*_a(p^e)}\right].$$ In particular, $q$ must divide $\frac{\lambda(p^e)}{\ell^*_a(p^e)}$ for some prime $p$. If $q = p,$ then $q \mid \lambda(p^{e})$ implies that $e \geq 2,$ so $q^2 \mid n.$ If $q \neq p$, then $q \mid \frac{\lambda(p)}{\ell^*_a(p)}$, so $p > q > \psi(X) > a$. Thus, $\ell^*_a(p) = \ell_a(p) \mid \frac{p-1}{q}$, so $p \mid a^{\frac{p-1}{q}} - 1,$ which implies that $p \in A_q.$
To handle the case where $q^2 \mid n$, we observe that $$\begin{aligned}
\#\{n \leq X: q^2 \mid n \ \hbox{for some prime} \ q \geq \psi(X)\} & \leq \sum_{\substack{q \geq | Lemma \[lp1\ ]. The number of integers $ n \leq X$ with $ P(\frac{\lambda(n)}{\ell^*_a(n) }) \geq \psi(X)$ is $ O(\frac{X \log \log \psi(X)}{\log \psi(X)}).$
Suppose that $ newton \leq X$ and $ q = P(\lambda(n)/\ell^*_a(n) ) \geq \psi(X).$ We may wear that $ X$ is large, so $ a$ is not a $ q^{th}$ power and $ \psi(X) > a$. furthermore, as we will now show, it must be the case that either $ q^2 \mid n$ or $ phosphorus \mid n$ for some $ phosphorus \in A_q$. Observe that $ $ q \mid \frac{\lambda(n)}{\ell^*_a(n) } \mid \frac{\mathrm{lcm}_{p^{e } \mid \mid n } \left[\lambda(p^{e})\right]}{\mathrm{lcm}_{p^{e } \mid \mid n } \left[\ell^*_a(p^e)\right ] } \mid \mathrm{lcm}_{p^{e } \mid \mid n } \left[\frac{\lambda(p^e)}{\ell^*_a(p^e)}\right].$$ In particular, $ q$ must divide $ \frac{\lambda(p^e)}{\ell^*_a(p^e)}$ for some prime $ p$. If $ q = p,$ then $ q \mid \lambda(p^{e})$ implies that $ e \geq 2,$ indeed $ q^2 \mid n.$ If $ q \neq p$, then $ q \mid \frac{\lambda(p)}{\ell^*_a(p)}$, so $ p > q > \psi(X) > a$. therefore, $ \ell^*_a(p) = \ell_a(p) \mid \frac{p-1}{q}$, so $ p \mid a^{\frac{p-1}{q } } - 1,$ which implies that $ p \in A_q.$
To handle the case where $ q^2 \mid n$, we observe that $ $ \begin{aligned }
\#\{n \leq X: q^2 \mid n \ \hbox{for some premier } \ q \geq \psi(X)\ } & \leq \sum_{\substack{q \geq | Lelma \[lp1\]. The number of inttgers $n \leq X$ witk $P(\frac{\nambda(h)}{\ell^*_a(n)}) \gdq \psi(X)$ is $O(\frac{X \log \log \psm(X)}{\lot \psi(Z)}).$
Suppose that $n \leq X$ xnd $q = P(\lwmbda(n)/\elo^*_a(n)) \jeq \psi(X).$ We may easume tmct $X$ jd laxgx, so $a$ is not a $q^{th}$ power and $\psi(X) > a$. Mmrdorer, as we will now show, it must be tre case tjat either $q^2 \myd n$ jr $p \mid n$ for some $p \in A_q$. Observe thzt $$q \miv \frac{\lambda(n)}{\elk^*_a(n)} \mid \frac{\mathrm{lcm}_{p^{e} \mif \mif n} \left[\lambda(p^{e})\rihht]}{\mathrm{lcn}_{p^{e} \iud \mid n} \lefg[\ell^*_a(p^e)\right]} \mid \mathrj{lcm}_{p^{e} \mid \mid n} \left[\frac{\lambda(o^e)}{\ell^*_c(p^e)}\right].$$ In pqrtlwular, $q$ musv diviqe $\frac{\lambda(p^e)}{\ell^*_a(p^a)}$ for spme prime $p$. If $q = p,$ rhen $q \mid \lambda(p^{e})$ ikplies that $e \geq 2,$ so $q^2 \mid n.$ If $q \neq p$, then $w \nid \ftac{\lakbda(o)}{\wll^*_x(p)}$, ao $p > q > \pdi(X) > a$. Thus, $\eml^*_a(p) = \ell_a(p) \mid \frac{p-1}{q}$, so $p \mic w^{\drac{p-1}{q}} - 1,$ whicg implyef that $p \in A_q.$
To handle the case where $x^2 \mjd n$, we observe that $$\betin{aligned}
\#\{n \leq X: q^2 \mld n \ \hbov{for some prime} \ q \geq \psi(X)\} & \leq \sum_{\substack{q \geq | Lemma \[lp1\]. The number of integers $n with \geq \psi(X)$ $O(\frac{X \log \log \leq and $q = \geq \psi(X).$ We assume that $X$ is large, so is not a $q^{th}$ power and $\psi(X) > a$. Moreover, as we will show, it must be the case that either $q^2 \mid n$ or $p n$ some \in Observe that $$q \mid \frac{\lambda(n)}{\ell^*_a(n)} \mid \frac{\mathrm{lcm}_{p^{e} \mid \mid n} \left[\lambda(p^{e})\right]}{\mathrm{lcm}_{p^{e} \mid \mid n} \left[\ell^*_a(p^e)\right]} \mid \mid \mid n} \left[\frac{\lambda(p^e)}{\ell^*_a(p^e)}\right].$$ In particular, $q$ must $\frac{\lambda(p^e)}{\ell^*_a(p^e)}$ for some prime If $q = p,$ then \mid implies that \geq so \mid n.$ If \neq p$, then $q \mid \frac{\lambda(p)}{\ell^*_a(p)}$, so $p > q > \psi(X) > a$. Thus, $\ell^*_a(p) = \mid \frac{p-1}{q}$, \mid a^{\frac{p-1}{q}} 1,$ implies $p \in A_q.$ the case where $q^2 \mid n$, $$\begin{aligned} \#\{n \leq X: q^2 \mid n \ some prime} q \geq \psi(X)\} & \leq \sum_{\substack{q | Lemma \[lp1\]. The number of integerS $n \leq X$ with $p(\frac{\LamBda(N)}{\eLl^*_a(n)}) \Geq \pSi(X)$ is $O(\frac{X \log \LOg \psI(X)}{\log \psi(X)}).$
Suppose that $n \lEq X$ anD $q = p(\LambDA(n)/\Ell^*_a(n)) \Geq \psi(X).$ wE mAY AssUmE tHat $x$ iS LaRge, so $A$ is Not a $q^{th}$ Power and $\psI(X) > a$. moReover, as we wiLL nOw show, it muSt bE the case that EitHer $q^2 \miD n$ Or $p \MId n$ foR soMe $p \in a_q$. ObseRVe that $$Q \mid \frac{\lAmBDa(n)}{\ell^*_A(N)} \mid \fraC{\MAtHrm{lCm}_{p^{e} \mid \mid n} \left[\laMBdA(P^{e})\right]}{\mathrm{lCm}_{p^{e} \miD \mID n} \LEFt[\eLl^*_a(P^e)\right]} \mid \MaThrm{lCM}_{p^{e} \mid \mID n} \LEFT[\frAC{\lambda(p^e)}{\ell^*_a(P^e)}\right].$$ In paRTicUlar, $q$ mUsT diVIde $\fraC{\lambDa(P^E)}{\elL^*_a(p^e)}$ for some PrimE $p$. If $q = p,$ theN $q \mid \lAMbda(p^{e})$ iMPlies thAt $e \geq 2,$ So $q^2 \Mid N.$ If $q \NEq P$, tHen $Q \mID \frAC{\lAmbDA(p)}{\eLl^*_a(p)}$, so $p > q > \PsI(X) > A$. Thus, $\Ell^*_a(P) = \ELL_A(p) \miD \frAc{p-1}{q}$, So $p \miD a^{\frac{p-1}{q}} - 1,$ which ImpLies THat $P \in A_q.$
to hanDle tHe Case wHere $q^2 \mId n$, we ObServe that $$\begin{aLignEd}
\#\{n \leq X: q^2 \mId n \ \HbOx{fOr Some pRIme} \ q \geQ \psI(X)\} & \lEq \sum_{\suBstack{q \GEq | Lemma \[lp1\]. The number of intege rs $n \l eqX$ wit h $P (\frac{\lambda ( n)}{ \ell^*_a(n)}) \geq \ps i(X)$ i s $O( \ fr ac{X\log \l o g\ p si( X) }{ \lo g\ ps i(X)} ).$
Suppo se that $n \l eq X$ and $q = P( \lambda(n) /\e ll^*_a(n)) \ geq \psi( X) .$W e may as sumethat $ X $ is l arge, so$a $ is no t a $q^{ t h }$ pow er and $\psi(X) > a$ . Moreover, aswe wil ln ow s how , i t must beth e cas e that e i th e r $q^ 2 \mid n$ or $ p \mid n$ f o r s ome $p \ inA _q$. O bserv et hat $$q \mid \ frac {\lambda( n)}{\e l l^*_a(n ) } \mid\frac{ \ma thr m{lc m }_ {p ^{e }\ mid \m idn } \ left[\la mb da (p^{e })\r i g h t ]}{\ mat hrm{ lcm}_ {p^{e} \mid \ mid n}\ lef t[\el l^*_a (p^e )\ right ]} \mi d \ma th rm{lcm}_{p^{e}\mid \mid n}\le ft [\f ra c{\la m bda(p^ e)} {\e ll^*_a( p^e)}\r i ght ]. $ $ In particular, $q$ m us t di vide $\f rac{\l a mb da ( p^e)}{\e ll ^*_ a(p^ e ) }$ fo r so m eprime $p $. If$ q=p,$ the n$q \mi d\la mbd a(p^{ e })$implie s that $ e \ge q 2,$ so $q^2 \ m id n.$ If $q\ ne q p$ , the n $ q \mid \fra c{\l a mbda (p)} { \e ll^ * _a(p) }$, s o$ p> q > \psi(X) > a$.Th us, $\ ell^* _a(p) = \ell_ a(p) \mid\ f r ac{p-1}{ q}$, so $p \mid a^{\fr ac{p- 1}{q}} - 1 , $ whichimpli es that$p \in A_ q . $
To ha ndl e t hecas e wh ere $q^2 \mid n $, w eobserve th at $$\b egi n{a lig ned }\#\{n \le q X: q^2 \ mi dn\ \ hbox{ f or somepr ime }\ q \geq \psi(X )\} & \le q\s u m_{ \substa c k{ q \geq | Lemma_\[lp1\]. The_number of integers $n_\leq X$_with_$P(\frac{\lambda(n)}{\ell^*_a(n)}) \geq_\psi(X)$_is $O(\frac{X \log_\log \psi(X)}{\log \psi(X)}).$
Suppose_that $n \leq X$_and $q =_P(\lambda(n)/\ell^*_a(n))_\geq \psi(X).$ We may assume that $X$ is large, so $a$ is not a_$q^{th}$_power and_$\psi(X)_>_a$. Moreover, as we will_now show, it must be_the case_that either $q^2 \mid n$ or $p \mid_n$_for some $p_\in A_q$. Observe that $$q \mid \frac{\lambda(n)}{\ell^*_a(n)} \mid \frac{\mathrm{lcm}_{p^{e}_\mid \mid n} \left[\lambda(p^{e})\right]}{\mathrm{lcm}_{p^{e} \mid \mid_n} \left[\ell^*_a(p^e)\right]} \mid_\mathrm{lcm}_{p^{e}_\mid_\mid n} \left[\frac{\lambda(p^e)}{\ell^*_a(p^e)}\right].$$ In_particular, $q$ must divide $\frac{\lambda(p^e)}{\ell^*_a(p^e)}$ for_some prime $p$. If $q =_p,$ then $q \mid \lambda(p^{e})$ implies that_$e \geq 2,$ so $q^2 \mid_n.$ If $q \neq p$,_then $q_\mid \frac{\lambda(p)}{\ell^*_a(p)}$, so $p >_q > \psi(X)_> a$._Thus, $\ell^*_a(p) =_\ell_a(p) \mid \frac{p-1}{q}$, so $p \mid_a^{\frac{p-1}{q}} - 1,$_which implies that $p \in A_q.$
To_handle_the case where_$q^2_\mid_n$, we_observe that $$\begin{aligned}
\#\{n_\leq_X: q^2_\mid_n \ \hbox{for some prime} \_q_\geq \psi(X)\} & \leq \sum_{\substack{q \geq |
redshift evolution of baryonic features in the galaxy power spectrum; iii) weak-lensing tomography derived from measurement of galaxy shear patterns and iv) the luminosity distance–redshift relation of type Ia SNe. We focus our attention on theoretical issues related to the first three of these tests.
The power spectrum of fluctuations at recombination is calculated to high accuracy from linear theory$^{\ino}$, so the problem of realizing, through direct simulation, the evolution of a finite patch of a particular world model from this epoch forward is well posed. To support DES-like observations, one would like to evolve multiple regions of Hubble Length dimension with the principal clustered matter components — dark matter and multiple phases of baryons (stars and cold gas in galaxies, warm/hot gas surrounding galaxies and in groups/clusters) — represented by multiple fluids. Mapping observable signatures of the theoretical solution along the past light-cone of synthetic observers in the computational volume allows ‘clean’ mock surveys to be created, which can further be ‘dirtied’ by real-world observational effects in support of survey data analysis.
Two fundamental barriers stand in the way of achieving a complete solution to this problem. One is the wide dynamic range of non-linear structures (sub-parsecs to gigaparsecs in length, for example), and the other is the complexity of astrophysical processes controlling the baryonic phases. Since DES-like surveys probe only the higher mass portion of the full spectrum of cosmic structures, the first issue is not strongly limiting. The DES will, however, require understanding galaxy and cluster observable signatures, so uncertainties in the treatment of baryonic physics will play a central role. In a companion paper$^{\ino}$, we outline theoretical uncertainties associated with the large-scale structure channels DES will use to test dark energy. Here, we offer a critique of the computational methods that provide theoretical support for DES and similar surveys.
Challenges for Computational Cosmology
=======================================
Given the wide scale and scope of large-scale structure, a number of approaches have evolved to address restricted elements of the full problem. Since cold dark matter dominates the matter density, N-body methods that follow collisionless clustering have played an important role in defining the overall evolution of the cosmic web and the structure of collapsed halos formed within it. Combined N-body and gas dynamics techniques explore gravitationally coupled evolution of baryons and dark matter. Knowledge gained from these ‘direct’ approaches led to the development of ‘semi-analytic’ | redshift evolution of baryonic features in the galaxy power spectrum; iii) weak - lensing imaging derive from measurement of galaxy shear patterns and iv) the luminosity distance – red shift relation of character Ia SNe. We focus our attention on theoretical issue related to the first three of these trial.
The power spectrum of fluctuations at recombination is forecast to high accuracy from linear theory$^{\ino}$, so the problem of realize, through lineal simulation, the evolution of a finite while of a particular world exemplar from this epoch forward is well posed. To support DES - comparable observations, one would like to evolve multiple regions of Hubble Length property with the principal clustered matter components — dark topic and multiple phases of baryons (stars and cold gas in galaxies, warm / blistering gasoline surrounding galaxies and in groups / clusters) — represented by multiple fluids. Mapping observable signatures of the theoretical solution along the past luminosity - cone of synthetic observers in the computational book allow ‘ clean ’ mock surveys to be created, which can far be ‘ dirty ’ by real - world observational effect in support of survey data analysis.
Two fundamental barrier stand in the way of achieving a complete solution to this trouble. One is the wide active range of non - linear structures (sub - parsecs to gigaparsecs in length, for case), and the other is the complexity of astrophysical processes controlling the baryonic phase. Since DES - like surveys probe only the higher multitude portion of the full spectrum of cosmic structure, the first topic is not strongly restrict. The DES will, however, require reason galaxy and cluster discernible signatures, so uncertainties in the treatment of baryonic physics will play a central role. In a companion paper$^{\ino}$, we outline theoretical uncertainties associated with the large - scale social organization channels DES will use to screen dark energy. Here, we offer a critique of the computational methods that provide theoretical support for DES and exchangeable surveys.
Challenges for Computational Cosmology
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Given the wide scale and scope of large - plate structure, a number of approaches have evolved to address restricted elements of the full trouble. Since cold dark matter dominates the matter density, N - body methods that follow collisionless clustering have played an significant role in defining the overall evolution of the cosmic web and the structure of collapsed aura formed within it. Combined N - body and gas dynamics techniques explore gravitationally copulate evolution of baryons and dark matter. Knowledge gained from these ‘ direct ’ approaches led to the development of ‘ semi - analytic ’ | refshift evolution of baryunic features iu the gelaxy pkwer speztrum; iii) weak-lensing tomogrephy dericed from measurement ow galaxy dhear parteris and iv) the luminosity distahge–redviift relation on type Ia STe. We focus ous xtcention on theoretical issues relateq to thr virst three of thexq teans.
Uhe power spectrum of fluctuationa at rebombination is cakculated to high accuracy vrom linear theory$^{\ino}$, do the proboem jd realizing, ghrough digzct simulatjon, the evolution of a finite pxtch pf a partixuoar forld model from this epoch nprward is welk posed. To supiort VES-luke observations, one xould like to evolve multiple rzgions of Hubble Lengrh dimetsiot wigy tfe kriicilal clkstxred matter components — dark matter and mileplle phases or baryjnf (stars and cold gas in galaxies, warm/hou gas surrounding galaxies abd in groups/clusters) — represeneed by multiple fluids. Mapping observable signatuses oh ghe tmeordrifal solution along the past light-cone of syntrstoc observers in bhe computational fopukg allows ‘clean’ mock surbeys to be created, which san fyrther be ‘diryied’ by real-world observatiinal effects un support of survzy data analvsis.
Twp funcamental barriers stand in fhe way of wchieving x complete solutkon tm this problem. One is the ride dynanic xange of non-kinear structured (sub-idrsecs to gigaparsfcs iu lencth, for exwmple), and the other is the complexity of astrppvyspcal proczsses gontrolling the baryonic phasgs. Since BES-likd surveys irobe onlb the higher mass portion lf the full vpectrum of xosmuc struzgures, the firsy issue is not strobgly limiting. The VES djll, however, reqbnrt ynderstanding balxxy ajd cjgster observdble siematurds, so uncerbaivtiex in the treatment ox badyonic physics wilk ilay a cebtral roje. In a compamion paper$^{\ino}$, we oktlinx theoceticak ugcertainties associated with tge large-sfalc structure cranncls QES will uxe to test dark energy. Here, we offer a rritique of the computarional methods that krovide theorxtical support xor DES and similar wurveys.
Challenges for Computational Coamologf
=======================================
Givfn the wide scale and scope of large-scale structure, a number of approachew have evolved to zddrrss rvstxicced eleientd of the full pronlem. Since cold dark matter dommnates the mctter density, N-body methods tnag follow collksionless clustering havs played an important role in defining yhe overall evolution of the cosnic wsb and the structure of collspsed kalos dormed qithln it. Combined N-bpdn and gas fynakies techniques explore gravitationqlly ciuoled evolution oh baryons anf dark mattet. Knowledge gaines frkl there ‘direct’ apptoaches led to the development of ‘veii-analytic’ | redshift evolution of baryonic features in the spectrum; weak-lensing tomography from measurement of the distance–redshift relation of Ia SNe. We our attention on theoretical issues related the first three of these tests. The power spectrum of fluctuations at recombination calculated to high accuracy from linear theory$^{\ino}$, so the problem of realizing, through simulation, evolution a patch of a particular world model from this epoch forward is well posed. To support DES-like one would like to evolve multiple regions of Length dimension with the clustered matter components — dark and phases of (stars cold in galaxies, warm/hot surrounding galaxies and in groups/clusters) — represented by multiple fluids. Mapping observable signatures of the theoretical solution the past synthetic observers the volume ‘clean’ mock surveys created, which can further be ‘dirtied’ effects in support of survey data analysis. Two barriers stand the way of achieving a complete to this problem. One is the wide dynamic of non-linear structures (sub-parsecs to gigaparsecs in length, for example), and the other is the astrophysical processes controlling the phases. Since DES-like probe the mass of the spectrum of cosmic structures, the first issue is not strongly limiting. DES will, however, require understanding galaxy and cluster observable signatures, in treatment of baryonic will play a central In companion paper$^{\ino}$, we outline associated the DES use test dark energy. Here, offer a critique of the methods that provide theoretical surveys. Challenges for Computational Cosmology ======================================= Given the scale and scope of large-scale structure, a of approaches have evolved to address restricted elements of the full problem. cold dark the matter density, N-body methods that follow collisionless have played an important in defining the overall evolution of the cosmic web the of collapsed formed within it. N-body and gas techniques explore gravitationally of baryons dark Knowledge ‘direct’ approaches led to the development ‘semi-analytic’ | redshift evolution of baryonIc features In the GalAxy PoWer sPectRum; iii) weak-lensINg toMography derived from meaSuremEnT Of gaLAxY sheaR patterNS aND Iv) tHe LuMinOsITy DistaNce–RedshifT relation oF tyPe ia SNe. We focus OUr Attention oN thEoretical issUes RelateD tO thE First ThrEe of tHese teSTs.
The pOwer spectRuM Of flucTUations AT ReCombInation is calculatED tO High accuracy frOm lineAr THeORY$^{\inO}$, so The problem Of RealiZIng, throUGh DIREct SImulation, the eVolution of a FIniTe patcH oF a pARticulAr worLd MOdeL from this epOch fOrward is wEll posED. To suppORt DES-liKe obseRvaTioNs, onE WoUlD liKe TO evOLvE muLTipLe regionS oF HUbble lengTH DIMensIon With The prIncipal clusteRed MattER coMponeNts — daRk maTtEr and MultipLe phaSeS of baryons (stars And cOld gas in gAlaXiEs, wArM/hot gAS surroUndIng GalaxieS and in gROupS/cLUSTeRs) — represented by mulTiPLE fLuids. MapPing obSErVaBLe signatUrEs oF the THEoretIcal SOlUtion aloNg the pASt LiGht-cone Of SyntheTiC obSerVers iN The cOmputaTional voLume aLLows ‘clean’ mock sURveys to be creaTEd, WHIcH Can fUrtHer be ‘dirtieD’ by rEAl-woRld oBSeRvaTIonal EffecTs IN sUPport of survey data anAlYsis.
TwO fundAmental barrieRs stand in tHE WAy of achiEvinG A cOMplete solution To thiS problem. OnE Is the widE dynaMic range Of non-lineAR StructurEs (sUb-pArsEcs TO GiGaparsecs in leNGTh, foR eXample), aNd tHe other Is tHe cOmpLexItY of astropHysical pRoCeSsEs ConTrollINg the barYoNic PhAseS. SincE dES-likE survEys pRoBe ONly The highER mASS porTiOn Of thE fuLl SpectRum oF CosMic struCtures, the FirST issUe Is Not stroNgly limiting. THe dES will, howEvEr, rEquire UNDerstandIng galaxy and cluster obseRVable siGnaTures, So unCertaintiEs iN the trEatMEnt of bAryoniC physIcS wiLL Play a CENtRal RoLe. In a compaNIOn pAper$^{\iNo}$, We ouTline thEoretical uncertainTIes Associated witH thE larGE-ScAle STrUCtuRe CHanNELs DES will use to tEst dark eneRgY. heRe, we offer a CRitIqUe of the ComputaTionaL Methods That proviDe theoretIcAl suPPOrt For DES and sImilar suRveys.
ChalLEnges FOr compuTatIonal COsMolOgy
=======================================
GiVen the WIde Scale And scoPe Of largE-scalE sTructure, A number of approaches have EvolveD to adDreSs restricTed ELemEnts of the Full Problem. SinCe cOld Dark mAttER domiNateS ThE maTTer deNsitY, n-body methODs ThaT FOlLow collisioNLESs cLusteRinG Have plAyed An important role in DEfining the overAll eVOLutIon OF the CoSmic web and the sTruCtURE of collaPsEd halos formEd within It. cOmbinEd N-bodY and gaS dynamiCS TeCHniqueS expLorE gravitatIonAlLY coupleD eVoLUtion oF barYoNs and dArk matTEr. KnOWLedge gained from tHese ‘dIREct’ apPRoaChes lEd To the deVElopMent of ‘semi-Analytic’ | redshift evolution of bar yonic feat uresinthe g alax y po wer spectrum;i ii)weak-lensing tomograph y der iv e d fr o mmeasu remento fg a lax ysh ear p a tt ernsand iv) th e luminosi tydi stance–redsh i ft relationoftype Ia SNe. We focus o ura ttent ion on t heoret i cal is sues rela te d to th e firstt h re e of these tests.
Th e p o wer spectrum o f fluc tu a ti o n s a t r ecombinati on is c a lculate d t o h igh accuracy from linear the o ry$ ^{\ino }$ , s o the p roble mo f r ealizing, t hrou gh direct simul a tion, t h e evolu tion o f a fi nite pa tc h o fa pa r ti cul a r w orld mod el f rom t hise p o c h fo rwa rd i s wel l posed. To s upp ortD ES- likeobser vati on s, on e woul d lik eto evolve multi pleregions o f H ub ble L ength dimens ion wi th theprincip a l c lu s t e re d matter component s— da rk matte r andm ul ti p le phase sofbary o n s (st arsa nd cold ga s in g a la xi es, war m/ hot ga ssur rou nding gala xies a nd in gr oups/ c lusters) — rep r esented by mu l ti p l ef luid s.Mapping obs erva b le s igna t ur eso f the theo re t ic a l solution along th epast l ight- cone of synth etic obser v e r s in the com p ut a tional volumeallow s ‘clean’m ock surv eys t o be cre ated, whi c h can fur the r b e ‘ dir t i ed ’ by real-wor l d obs er vationa l e ffectsinsup por t o fsurvey da ta analy si s.
Tw o f undam e ntal bar ri ers s tan d int he way of a chie vi ng a c omplete so l u tion t othis pr ob lem.Onei s t he wide dynamicran g e of n on -linear structures ( su b-parsecsto gi gapars e c s in len gth, for example), andt he othe r i s the com plexity o f a stroph ysi c al pro cesses cont ro lli n g theb a ry oni cphases. Si n c e D ES-li ke sur veys pr obe only the highe r ma ss portion of th e fu l l s pec t ru m of c o smi c structures, the first iss ue is not stron g lyli miting. The DE S wil l , howev er, requi re unders ta ndin g gal axy and cl uster ob servables ignat u re s, so un certai nt ies in t he tre a tme nt of baryo ni c phys ics w il l play a central role. In a com panion pape r$^ {\ino}$,weo utl ine theor etic al uncerta int ies asso cia t ed wi th t h elar g e-sca le s t ructure c h an nel s DE S will uset o tes t dar k e n ergy.Here , we offer a crit i que of the com puta t i ona l m e thod sthat provide t heo re t i cal supp or t for DES a nd simil ar surve ys.
C hallen ges for C om p utatio nalCos mology
= === == = ======= == == = ====== ==== == ====== ==
Gi v en t h e wide scale andscope o f lar g e-s calest ructure , a n umber of a pproaches h ave ev olve d toaddress r estric ted e lements of the fullprobl em. Sin ce col d d ark ma tter d omina testh e m atter den s i ty , N -b o dymeth ods t ha t fo llow coll i sionless cl u stering h ave p layeda ni m portant ro lein de f i ning the o v eral l e v oluti on ofthe co smic we b an dthe str uct u r e of coll apsed hal o s f or medwithin i t. Com bi ned N -body and gas d ynamicstechniqu e s explo reg ravi t ationa ll y coupled ev oluti on of bary o n s a nd dark mat ter . K nowl edg e ga i ned from these ‘direct’appr oache s le d tothe d ev elopment of ‘ semi-ana l ytic’ | redshift_evolution of_baryonic features in the_galaxy power_spectrum;_iii) weak-lensing_tomography_derived from measurement_of galaxy shear_patterns and iv) the_luminosity distance–redshift relation_of_type Ia SNe. We focus our attention on theoretical issues related to the first_three_of these_tests.
The_power_spectrum of fluctuations at recombination_is calculated to high accuracy_from linear_theory$^{\ino}$, so the problem of realizing, through direct_simulation,_the evolution of_a finite patch of a particular world model from_this epoch forward is well posed._To support DES-like_observations,_one_would like to evolve_multiple regions of Hubble Length dimension_with the principal clustered matter components_— dark matter and multiple phases of_baryons (stars and cold gas in_galaxies, warm/hot gas surrounding galaxies_and in_groups/clusters) — represented by multiple_fluids. Mapping observable_signatures of_the theoretical solution_along the past light-cone of synthetic_observers in the_computational volume allows ‘clean’ mock surveys_to_be created, which_can_further_be ‘dirtied’_by real-world observational_effects_in support_of_survey data analysis.
Two fundamental barriers stand_in_the way of achieving a complete solution_to this problem. One_is_the wide dynamic range_of non-linear structures (sub-parsecs to_gigaparsecs in length, for example), and_the other_is the_complexity of astrophysical processes controlling the baryonic phases. Since DES-like surveys_probe only the higher mass portion_of the full spectrum_of cosmic_structures,_the first issue_is_not strongly_limiting. The DES will, however, require understanding_galaxy and_cluster observable signatures, so uncertainties in_the treatment of baryonic_physics_will play a central role. In_a companion paper$^{\ino}$, we outline theoretical_uncertainties associated with the large-scale_structure_channels_DES will use to test_dark energy. Here, we offer a_critique of the_computational methods that provide theoretical support for_DES_and similar surveys.
Challenges for Computational Cosmology_
=======================================
Given_the wide scale and scope of_large-scale_structure,_a number of approaches have_evolved to address restricted elements of_the full problem. Since cold dark matter dominates the_matter density, N-body_methods that follow collisionless clustering_have_played_an important role in defining the overall evolution of the_cosmic web_and the structure_of collapsed halos formed within it. Combined N-body and gas_dynamics techniques explore gravitationally coupled evolution of_baryons and dark matter. Knowledge gained from these ‘direct’ approaches led_to the development of ‘semi-analytic’ |
$, then since $z\in\delta\cap P_2$ we get that $\delta$ must be contained in $P_2$. The boundary component $\delta$ is a simple closed curve, and topologically, a pair of pants can be viewed as a sphere with three points removed, so $\delta$ must bound a disc or a punctured disk. Hence, $\delta$ is either contractible, or freely homotopic to one of the boundary components of $P_2$. But $\delta$ is a geodesic, thereby it is not contractible, and if it is homotopic to one of the geodesic boundary components of $P_2$ then they must coincide by Theorem \[thm:prelim\], in contradiction to the choice of $\delta$. Therefore, $\delta\cap\partial P_2\ne\emptyset$. If two geodesic curves intersect then they either coincide or have transversal intersection. We have already shown that they do not coincide, so the result follows.
![$\gamma_1$ and $\gamma_2$ are freely homotopic to $\delta_1$ and $\delta_2$ respectively. $\gamma_3:=\gamma_1*\gamma_2^{-1}$ is freely homotopic to $\delta_3$. []{data-label="fig:Fig8"}](Fig8_jpg.eps)
\[prop:P1P2\] Let $P_1,P_2\subset{M}$ be two pairs of pants with geodesic boundaries, such that $P_1\cap P_2\ne \emptyset$. Let $e_1,e_2$ be two figure eights contained in $P_1$ and $P_2$, respectively, such that for $i=1,2$ $P_i\backslash e_i$ is a disjoint union of three cylinders. Then $e_1\cap e_2\ne \emptyset$.
We will first make the following observation. Let $e$ be a figure eight contained in a pair of pants $P$ with geodesic boundaries such that $P\backslash e$ is a disjoint union of three cylinders. Let $x_0$ be the unique transversal intersection point of the figure eight $e$. The figure eight can be divided into two simple closed curves $\gamma_1$ and $\gamma_2$ with endpoints at $x_0$. Define $\gamma_3$ to be the (non-smooth) curve $\gamma_ | $, then since $ z\in\delta\cap P_2 $ we get that $ \delta$ must be contained in $ P_2$. The boundary component $ \delta$ is a simple shut curvature, and topologically, a pair of pants can be watch as a celestial sphere with three points removed, indeed $ \delta$ must bind a disc or a pierced phonograph record. Hence, $ \delta$ is either contractible, or freely homotopic to one of the boundary component of $ P_2$. But $ \delta$ is a geodesic, thereby it is not contractible, and if it is homotopic to one of the geodesic boundary components of $ P_2 $ then they must concur by Theorem \[thm: prelim\ ], in contradiction to the choice of $ \delta$. Therefore, $ \delta\cap\partial P_2\ne\emptyset$. If two geodetic curves intersect then they either concur or consume transversal intersection. We have already shown that they do not concur, so the result follows.
! [ $ \gamma_1 $ and $ \gamma_2 $ are freely homotopic to $ \delta_1 $ and $ \delta_2 $ respectively. $ \gamma_3:=\gamma_1*\gamma_2^{-1}$ is freely homotopic to $ \delta_3$. [ ] { data - label="fig: Fig8"}](Fig8_jpg.eps)
\[prop: P1P2\ ] Let $ P_1,P_2\subset{M}$ be two pairs of bloomers with geodesic boundaries, such that $ P_1\cap P_2\ne \emptyset$. Let $ e_1,e_2 $ be two figure eights hold in $ P_1 $ and $ P_2 $, respectively, such that for $ i=1,2 $ $ P_i\backslash e_i$ is a disjoint union of three cylinders. Then $ e_1\cap e_2\ne \emptyset$.
We will first make the following observation. lease $ e$ be a figure eight contained in a pair of pants $ P$ with geodesic boundaries such that $ P\backslash e$ is a disjoint union of three cylinders. Let $ x_0 $ be the unique cross overlap point of the figure eight $ e$. The figure eight can be separate into two simple closed curves $ \gamma_1 $ and $ \gamma_2 $ with end point at $ x_0$. Define $ \gamma_3 $ to be the (non - legato) curve $ \gamma _ | $, thfn since $z\in\delta\cap P_2$ wt get that $\delta$ must be contajned in $O_2$. The boundary component $\delva$ iw a sumple closed curve, and topologibally, a pqir id pants cai be viewed as w spkece with three ppints remoeed, so $\delta$ mgsg yound a disc or a punctured disk. Hense, $\delts$ ls either conttactinje, od freely homotopic to one of the bkundary components og $P_2$. But $\delta$ is a geodesif, thfreby it is not cojtractible, qnd yd it is homogopic to one of the gekdesic boundary components of $P_2$ then they must ciinfhde by Theocem \[thi:prelim\], in contradicdion to the choice of $\denta$. Therefore, $\delta\cap\pactial P_2\ne\emptyset$. If two geodasnc curves intersect tyeb thei eitver ziinzidt oc hzve trwnstersal intedsection. We have already shown trqt they do nof coinsiqe, so the result follows.
![$\gamma_1$ and $\gamma_2$ ars freely homotopic to $\dwlta_1$ and $\delta_2$ respecjively. $\gamia_3:=\gamma_1*\gamma_2^{-1}$ is freely homotopic to $\delta_3$. []{data-labal="fig:Hie8"}](Fiy8_mig.epr)
\[prlp:P1P2\] Let $P_1,P_2\subset{M}$ be two pairs of pants with fepdvsic boundaries, smch that $P_1\cap P_2\ne \rmotufet$. Let $e_1,e_2$ be two fnfude eights containef in $P_1$ wnd $P_2$, respectidely, such that for $i=1,2$ $P_i\backslasy e_i$ is a difhoint union of thrze cylinders. Them $e_1\cal e_2\ne \emptyset$.
We will fnrst mzke the folpowing obadrvation. Let $e$ bd a fhgure eight contained in a pair of 'ants $P$ with geocesic foundaries such that $P\backslash e$ ls a bisjohnt union lf three cylinders. Let $x_0$ be the unique transvgrsdl pntersectnon polnt of the figuwe eight $e$. The figure eighg can be dpvided invo two simplq closed curvad $\gamma_1$ and $\jamma_2$ witr enepoibts at $b_0$. Define $\gamma_3$ yo be the (non-smooth) curve $\gamma_ | $, then since $z\in\delta\cap P_2$ we get must contained in The boundary component curve, topologically, a pair pants can be as a sphere with three points so $\delta$ must bound a disc or a punctured disk. Hence, $\delta$ is contractible, or freely homotopic to one of the boundary components of $P_2$. But is geodesic, it not contractible, and if it is homotopic to one of the geodesic boundary components of $P_2$ they must coincide by Theorem \[thm:prelim\], in contradiction the choice of $\delta$. $\delta\cap\partial P_2\ne\emptyset$. If two geodesic intersect they either or transversal We have already that they do not coincide, so the result follows. ![$\gamma_1$ and $\gamma_2$ are freely homotopic to $\delta_1$ $\delta_2$ respectively. freely homotopic $\delta_3$. \[prop:P1P2\] $P_1,P_2\subset{M}$ be two pants with geodesic boundaries, such that Let $e_1,e_2$ be two figure eights contained in and $P_2$, such that for $i=1,2$ $P_i\backslash e_i$ a disjoint union of three cylinders. Then $e_1\cap \emptyset$. We will first make the following observation. Let $e$ be a figure eight contained pair of pants $P$ geodesic boundaries such $P\backslash is disjoint of three Let $x_0$ be the unique transversal intersection point of the figure $e$. The figure eight can be divided into two simple $\gamma_1$ $\gamma_2$ with endpoints $x_0$. Define $\gamma_3$ to the curve $\gamma_ | $, then since $z\in\delta\cap P_2$ we geT that $\delta$ Must bE coNtaInEd in $p_2$. The Boundary componENt $\deLta$ is a simple closed curvE, and tOpOLogiCAlLy, a paIr of panTS cAN Be vIeWeD as A sPHeRe witH thRee poinTs removed, sO $\deLtA$ must bound a dISc Or a puncturEd dIsk. Hence, $\deltA$ is Either CoNtrACtiblE, or FreelY homotOPic to oNe of the boUnDAry comPOnents oF $p_2$. buT $\delTa$ is a geodesic, therEBy IT is not contractIble, anD iF It IS HomOtoPic to one of ThE geodESic bounDArY COMpoNEnts of $P_2$ then thEy must coincIDe bY TheorEm \[Thm:PRelim\], iN contRaDIctIon to the choIce oF $\delta$. TheRefore, $\DElta\cap\PArtial P_2\Ne\emptYseT$. If Two gEOdEsIc cUrVEs iNTeRseCT thEn they eiThEr CoincIde oR HAVE traNsvErsaL inteRsection. We havE alReadY ShoWn thaT they Do noT cOinciDe, so thE resuLt Follows.
![$\gamma_1$ and $\GammA_2$ are freelY hoMoTopIc To $\delTA_1$ and $\deLta_2$ ResPectiveLy. $\gamma_3:=\GAmmA_1*\gAMMA_2^{-1}$ iS freely homotopic to $\DeLTA_3$. []{dAta-label="Fig:Fig8"}](fIg8_JpG.Eps)
\[prop:P1p2\] LEt $P_1,p_2\subSET{M}$ be tWo paIRs Of pants wIth geoDEsIc BoundarIeS, such tHaT $P_1\cAp P_2\Ne \empTYset$. let $e_1,e_2$ bE two figuRe eigHTs contained in $P_1$ ANd $P_2$, respectiveLY, sUCH tHAt foR $i=1,2$ $P_I\backslash e_I$ is a DIsjoInt uNIoN of THree cYlindErS. thEN $e_1\cap e_2\ne \emptyset$.
We wIlL first Make tHe following obServation. LET $E$ Be a figurE eigHT cONtained in a pair Of panTs $P$ with geoDEsic bounDarieS such thaT $P\backslaSH E$ is a disjOinT unIon Of tHREe Cylinders. Let $x_0$ BE The uNiQue tranSveRsal intErsEctIon PoiNt Of the figuRe eight $e$. thE fIgUrE eiGht caN Be divideD iNto TwO siMple cLOsed cuRves $\gAmma_1$ AnD $\gAMma_2$ With endPOiNTS at $x_0$. deFiNe $\gaMma_3$ To Be the (Non-sMOotH) curve $\gAmma_ | $, then since $z\in\delta\ cap P_2$ w e get th at$\ delt a$ m ust be contain e d in $P_2$. The boundary c ompon en t $\d e lt a$ is a simp l ec l ose dcu rve ,a nd topo log ically, a pair of pa nt s can be vie w ed as a sphe rewith three p oin ts rem ov ed, so $\ del ta$ m ust bo u nd a d isc or apu n ctured disk. H e n ce , $\ delta$ is eitherc on t ractible, or f reelyho m ot o p ictoone of the b ounda r y compo n en t s of$ P_2$. But $\d elta$ is ag eod esic,th ere b y it i s not c o ntr actible, an d if it is ho motopi c to one of thegeodes icbou ndar y c om pon en t s o f $ P_2 $ th en theymu st coin cide b y Theo rem \[t hm:pr elim\], in co ntr adic t ion to t he ch oice o f $\d elta$. Ther ef ore, $\delta\ca p\pa rtial P_2 \ne \e mpt ys et$.I f twogeo des ic curv es inte r sec tt h e nthey either coinci de o rhave tra nsvers a lin t ersectio n. We hav e alrea dy s h ow n that t hey do no tcoincid e, so th eres ult foll o ws.
![$\g amma_1$and $ \ gamma_2$ are f r eely homotopi c t o $\ d elta _1$ and $\delt a_2$ resp ecti v el y.$ \gamm a_3:= \g a mm a _1*\gamma_2^{-1}$ i sfreely homo topic to $\de lta_3$. [] { d a ta-label ="fi g :F i g8"}](Fig8_jpg .eps)
\[prop:P 1 P2\] Let $P_1 ,P_2\sub set{M}$ b e two pair s o f p ant s w i t hgeodesic boun d a ries ,such th at$P_1\ca p P _2\ ne\em pt yset$. Le t $e_1,e _2 $be t wofigur e eightsco nta in edin $P _ 1$ and $P_2 $, r es pe c tiv ely, su c ht h at f or $ i=1, 2$$P _i\ba cksl a she_i$ is a disjoi ntu nion o fthree c ylinders. The n$e_1\cap e _2 \ne \empt y s et$.
We will first make the fo l lowingobs ervat ion. Let $e$bea figu ree ight c ontain ed in a pa i r of p a n ts $P $with geode s i c b ounda ri es s uch tha t $P\backslash e$i s a disjoint uni onof t h r ee cy l in d ers .L et$ x _0$ be the uniq ue transve rs a lintersecti o n p oi nt of t he figu re ei g ht $e$. The figu re eightca n be d ivi ded into t wo simpl e closedc urves $\ gamma _1$ and $ \g amm a_2$with e n dpo intsat $x_ 0$ . Defi ne $\ ga mma_3$ t o be the (non-smooth) c urve $ \gamm a_ | $, then_since $z\in\delta\cap_P_2$ we get that_$\delta$ must_be_contained in_$P_2$._The boundary component_$\delta$ is a_simple closed curve, and_topologically, a pair_of_pants can be viewed as a sphere with three points removed, so $\delta$ must_bound_a disc_or_a_punctured disk. Hence, $\delta$ is_either contractible, or freely homotopic_to one_of the boundary components of $P_2$. But $\delta$_is_a geodesic, thereby_it is not contractible, and if it is homotopic_to one of the geodesic boundary_components of $P_2$_then_they_must coincide by Theorem_\[thm:prelim\], in contradiction to the choice_of $\delta$. Therefore, $\delta\cap\partial P_2\ne\emptyset$. If_two geodesic curves intersect then they either_coincide or have transversal intersection. We_have already shown that they_do not_coincide, so the result follows.
![$\gamma_1$_and $\gamma_2$ are_freely homotopic_to $\delta_1$ and_$\delta_2$ respectively. $\gamma_3:=\gamma_1*\gamma_2^{-1}$ is freely homotopic_to $\delta_3$. []{data-label="fig:Fig8"}](Fig8_jpg.eps)
\[prop:P1P2\]_Let $P_1,P_2\subset{M}$ be two pairs of_pants_with geodesic boundaries,_such_that_$P_1\cap P_2\ne_\emptyset$. Let $e_1,e_2$_be_two figure_eights_contained in $P_1$ and $P_2$, respectively,_such_that for $i=1,2$ $P_i\backslash e_i$ is a_disjoint union of three_cylinders._Then $e_1\cap e_2\ne \emptyset$.
We_will first make the following_observation. Let $e$ be a figure_eight contained_in a_pair of pants $P$ with geodesic boundaries such that $P\backslash e$_is a disjoint union of three_cylinders. Let $x_0$ be_the unique_transversal_intersection point of_the_figure eight_$e$. The figure eight can be divided_into two_simple closed curves $\gamma_1$ and $\gamma_2$_with endpoints at $x_0$._Define_$\gamma_3$ to be the (non-smooth) curve_$\gamma_ |
eq:conv_R1+Rb}\end{aligned}$$ where ($a$) follows because $(M_{1},M_{b})$ is a function of $X^{n}$; ($b$) follows because $M_{b}$ is a function of $X^{n}$, $A^{n}$ is a function of $M_{b}$ and $M_{2}$ is a function of $(M_{1},M_{b},Y^{n})$; ($c$) follows since $H(Y^{n},Z^{n}|X^{n},M_{b})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|$ $Y^{i-1},Z^{i-1},X^{n},M_{b},A_{i})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|X_{i},A_{i})$ and since $\textrm{h}_{1}$ and $\textrm{h}_{2}$ are functions of $(M_{1},M_{b},Y^{n})$ and $(M_{2},M_{b},Z^{n})$, respectively and because $(Y_{i},Z_{i})\textrm{---}(X_{i},A_{i})\textrm{---}$ $(X^{n\backslash i},Y^{i-1},Z^{i-1},M_{b})$ forms a Markov chain; ($d$) follows since conditioning reduces entropy, since side information VM follows $p(y^{n},z^{n}|a^{n},x^{n})$$=\prod_{i=1}^{n}p_{Y|A,X}(y_{i}|a_{i},x_{i})$ $p_{Z|A,Y}(z_{i}|a_{i},y_{i})$ from (\[eq:degradedness\]) and because $Z_{i}\textrm{---}(Y_{i},A_{i})\textrm{---}$ $(Y^{n\backslash i},Z^{i-1},M_{1},M_{b})$ forms a Markov chain; ($e$) follows by the chain rule for mutual information and the fact that mutual information is non-negative; and ($f$) follows by the Fano inequality (\[eqn: Fano\]) and because entropy is non-negative. We can also write $$\begin{aligned}
n(R_{2}+R_{b}) & \geq & H(M_{2},M_{b | eq: conv_R1+Rb}\end{aligned}$$ where ($ a$) follows because $ (M_{1},M_{b})$ is a function of $ X^{n}$; ($ b$) follows because $ M_{b}$ is a function of $ X^{n}$, $ A^{n}$ is a affair of $ M_{b}$ and $ M_{2}$ is a affair of $ (M_{1},M_{b},Y^{n})$; ($ c$) follows since $ H(Y^{n},Z^{n}|X^{n},M_{b})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|$ $ Y^{i-1},Z^{i-1},X^{n},M_{b},A_{i})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|X_{i},A_{i})$ and since $ \textrm{h}_{1}$ and $ \textrm{h}_{2}$ are functions of $ (M_{1},M_{b},Y^{n})$ and $ (M_{2},M_{b},Z^{n})$, respectively and because $ (Y_{i},Z_{i})\textrm{---}(X_{i},A_{i})\textrm{---}$ $ (X^{n\backslash i},Y^{i-1},Z^{i-1},M_{b})$ forms a Markov range; ($ d$) follows since conditioning reduces entropy, since slope information VM follows $ p(y^{n},z^{n}|a^{n},x^{n})$$=\prod_{i=1}^{n}p_{Y|A, X}(y_{i}|a_{i},x_{i})$ $ p_{Z|A, Y}(z_{i}|a_{i},y_{i})$ from (\[eq: degradedness\ ]) and because $ Z_{i}\textrm{---}(Y_{i},A_{i})\textrm{---}$ $ (Y^{n\backslash i},Z^{i-1},M_{1},M_{b})$ form a Markov chain; ($ e$) follows by the range rule for mutual information and the fact that mutual data is non - negative; and ($ f$) follows by the Fano inequality (\[eqn: Fano\ ]) and because entropy is non - damaging. We can also write $ $ \begin{aligned }
n(R_{2}+R_{b }) & \geq & H(M_{2},M_{b | eq:clnv_R1+Rb}\end{aligned}$$ where ($a$) follows becausg $(N_{1},M_{b})$ is a fundtion of $X^{n}$; ($b$) follows because $M_{b}$ is e fubctiob of $X^{n}$, $A^{n}$ is a functiun of $M_{b}$ wnd $M_{2}$ is a fnnction of $(M_{1},M_{b},Y^{n})$; ($c$) follows sindc $H(Y^{n},V^{n}|E^{n},M_{b})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|$ $I^{i-1},Z^{i-1},X^{n},M_{b},A_{i})=\suk_{i=1}^{n}H(Y_{i},Z_{i}|X_{i},A_{i})$ atd snnce $\textrm{h}_{1}$ and $\textrm{h}_{2}$ are functiogs of $(M_{1},K_{b},J^{n})$ and $(M_{2},M_{b},Z^{n})$, rgspecuivqly znd because $(Y_{i},Z_{i})\textrm{---}(X_{i},A_{i})\textrm{---}$ $(X^{h\backslesh i},Y^{i-1},Z^{i-1},M_{b})$ forks a Markov chain; ($d$) followd sijce conditioning rfduces entripy, funce side inwormation NK follows $k(y^{n},z^{n}|a^{n},x^{n})$$=\prod_{i=1}^{n}p_{Y|A,X}(y_{i}|a_{i},x_{i})$ $p_{Z|A,Y}(z_{i}|x_{i},y_{i})$ yrom (\[eq:degrqdwdnfvs\]) and becanse $Z_{i}\nextrm{---}(Y_{i},A_{i})\texbgm{---}$ $(Y^{n\bawkslash i},Z^{i-1},M_{1},M_{b})$ forms s Merkoc chain; ($e$) follows by vhe chain rule for motual infosmction and the fact thqt mutudl itforoqtiun js nkn-negahivx; and ($f$) folmows by the Fano inequality (\[eqm: Sqno\]) and becauae entwo[y is non-negative. We can also write $$\begpn{aljgned}
n(R_{2}+R_{b}) & \geq & H(M_{2},M_{b | eq:conv_R1+Rb}\end{aligned}$$ where ($a$) follows because $(M_{1},M_{b})$ is of ($b$) follows $M_{b}$ is a a of $M_{b}$ and is a function $(M_{1},M_{b},Y^{n})$; ($c$) follows since $H(Y^{n},Z^{n}|X^{n},M_{b})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|$ $Y^{i-1},Z^{i-1},X^{n},M_{b},A_{i})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|X_{i},A_{i})$ since $\textrm{h}_{1}$ and $\textrm{h}_{2}$ are functions of $(M_{1},M_{b},Y^{n})$ and $(M_{2},M_{b},Z^{n})$, respectively and because $(X^{n\backslash i},Y^{i-1},Z^{i-1},M_{b})$ forms a Markov chain; ($d$) follows since conditioning reduces entropy, since information follows $p_{Z|A,Y}(z_{i}|a_{i},y_{i})$ (\[eq:degradedness\]) and because $Z_{i}\textrm{---}(Y_{i},A_{i})\textrm{---}$ $(Y^{n\backslash i},Z^{i-1},M_{1},M_{b})$ forms a Markov chain; ($e$) follows by the chain rule mutual information and the fact that mutual information non-negative; and ($f$) follows the Fano inequality (\[eqn: Fano\]) because is non-negative. can write n(R_{2}+R_{b}) & \geq H(M_{2},M_{b | eq:conv_R1+Rb}\end{aligned}$$ where ($a$) Follows becAuse $(M_{1},m_{b})$ iS a fUnCtioN of $X^{N}$; ($b$) follows becauSE $M_{b}$ iS a function of $X^{n}$, $A^{n}$ is a funCtion Of $m_{B}$ and $m_{2}$ Is A funcTion of $(M_{1},m_{B},Y^{N})$; ($C$) FolLoWs SinCe $h(y^{n},z^{n}|X^{n},M_{B})=\suM_{i=1}^{n}H(Y_{i},Z_{I}|$ $Y^{i-1},Z^{i-1},X^{n},M_{b},A_{I})=\suM_{i=1}^{N}H(Y_{i},Z_{i}|X_{i},A_{i})$ anD SiNce $\textrm{h}_{1}$ And $\Textrm{h}_{2}$ are fuNctIons of $(m_{1},M_{B},Y^{n})$ ANd $(M_{2},M_{b},z^{n})$, rEspecTively ANd becaUse $(Y_{i},Z_{i})\teXtRM{---}(X_{i},A_{i})\tEXtrm{---}$ $(X^{n\bACKsLash I},Y^{i-1},Z^{i-1},M_{b})$ forms a MarkOV cHAin; ($d$) follows sinCe condItIOnING reDucEs entropy, sInCe sidE InformaTIoN vm FolLOws $p(y^{n},z^{n}|a^{n},x^{n})$$=\pRod_{i=1}^{n}p_{Y|A,X}(y_{i}|A_{I},x_{i})$ $P_{Z|A,Y}(z_{i}|A_{i},Y_{i})$ fROm (\[eq:deGradeDnESs\]) aNd because $Z_{i}\TextRm{---}(Y_{i},A_{i})\texTrm{---}$ $(Y^{n\bACkslash I},z^{i-1},M_{1},M_{b})$ foRms a MaRkoV chAin; ($e$) FOlLoWs bY tHE chAIn RulE For Mutual inFoRmAtion And tHE FACt thAt mUtuaL infoRmation is non-nEgaTive; ANd ($f$) FolloWs by tHe FaNo InequAlity (\[eQn: FanO\]) aNd because entropY is nOn-negativE. We CaN alSo Write $$\BEgin{alIgnEd}
n(r_{2}+R_{b}) & \geq & H(m_{2},M_{b | eq:conv_R1+Rb}\end{aligned }$$ where($a$) fo llo ws bec ause $(M_{1},M_{b} ) $ is a function of $X^{n}$ ; ($b $) foll o ws beca use $M_ { b} $ isafu nct io n o f $X^ {n} $, $A^{ n}$ is a f unc ti on of $M_{b} $ a nd $M_{2}$ is a functionof$(M_{1 }, M_{ b },Y^{ n}) $; ($ c$) fo l lows s ince $H(Y ^{ n },Z^{n } |X^{n}, M _ {b })=\ sum_{i=1}^{n}H(Y_ { i} , Z_{i}|$ $Y^{i- 1},Z^{ i- 1 }, X ^ {n} ,M_ {b},A_{i}) =\ sum_{ i =1}^{n} H (Y _ { i },Z _ {i}|X_{i},A_{ i})$ and si n ce$\text rm {h} _ {1}$ a nd $\ te x trm {h}_{2}$ ar e fu nctions o f $(M_ { 1},M_{b } ,Y^{n}) $ and$(M _{2 },M_ { b} ,Z ^{n }) $ , r e sp ect i vel y and be ca us e $(Y _{i} , Z _ { i})\ tex trm{ ---}( X_{i},A_{i})\ tex trm{ - --} $ $(X ^{n\b acks la sh i} ,Y^{i- 1},Z^ {i -1},M_{b})$ for ms a Markov c hai n; ($ d$ ) fol l ows si nce co ndition ing red u ces e n t r op y, since side info rm a t io n VM fol lows $ p (y ^{ n },z^{n}| a^ {n} ,x^{ n } )$$=\ prod _ {i =1}^{n}p _{Y|A, X }( y_ {i}|a_{ i} ,x_{i} )$ $p _{Z |A,Y} ( z_{i }|a_{i },y_{i}) $ fro m (\[eq:degrade d ness\]) and b e ca u s e$ Z_{i }\t extrm{---}( Y_{i } ,A_{ i})\ t ex trm { ---}$ $(Y^ {n \ ba c kslash i},Z^{i-1},M _{ 1},M_{ b})$forms a Marko v chain; ( $ e $ ) follow s by th e chain rule fo r mut ual inform a tion and thefact tha t mutuali n formatio n i s n on- neg a t iv e; and ($f$)f o llow sby theFan o inequ ali ty(\[ eqn :Fano\]) a nd becau se e nt ro pyis no n -negativ e. We c analsow rite $ $\beg in{a li gn e d}n(R_{2} + R_ { b }) & \ ge q &H(M _{ 2},M_ {b | eq:conv_R1+Rb}\end{aligned}$$ where_($a$) follows_because $(M_{1},M_{b})$ is a_function of_$X^{n}$;_($b$) follows_because_$M_{b}$ is a_function of $X^{n}$,_$A^{n}$ is a function_of $M_{b}$ and_$M_{2}$_is a function of $(M_{1},M_{b},Y^{n})$; ($c$) follows since $H(Y^{n},Z^{n}|X^{n},M_{b})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|$ $Y^{i-1},Z^{i-1},X^{n},M_{b},A_{i})=\sum_{i=1}^{n}H(Y_{i},Z_{i}|X_{i},A_{i})$ and since $\textrm{h}_{1}$ and_$\textrm{h}_{2}$_are functions_of_$(M_{1},M_{b},Y^{n})$_and $(M_{2},M_{b},Z^{n})$, respectively and because_$(Y_{i},Z_{i})\textrm{---}(X_{i},A_{i})\textrm{---}$ $(X^{n\backslash i},Y^{i-1},Z^{i-1},M_{b})$ forms a_Markov chain;_($d$) follows since conditioning reduces entropy, since side_information_VM follows $p(y^{n},z^{n}|a^{n},x^{n})$$=\prod_{i=1}^{n}p_{Y|A,X}(y_{i}|a_{i},x_{i})$_$p_{Z|A,Y}(z_{i}|a_{i},y_{i})$ from (\[eq:degradedness\]) and because $Z_{i}\textrm{---}(Y_{i},A_{i})\textrm{---}$ $(Y^{n\backslash i},Z^{i-1},M_{1},M_{b})$ forms_a Markov chain; ($e$) follows by_the chain rule_for_mutual_information and the fact_that mutual information is non-negative; and_($f$) follows by the Fano inequality_(\[eqn: Fano\]) and because entropy is non-negative._We can also write $$\begin{aligned}
n(R_{2}+R_{b}) &_\geq & H(M_{2},M_{b |
n\equiv n^{-C_1}$, $X_1,\ldots,X_n \in \{p_0\geq c_n\}$ with high probability. Let $\tilde{p}_n\equiv (\hat{p}_n\vee c_n)\bm{1}_{p_0\geq c_n} / \int (\hat{p}_n\vee c_n)\bm{1}_{p_0 \geq c_n}$. Then with $b_n \equiv \int (\hat{p}_n \vee c_n)\bm{1}_{p_0\geq c_n}$, it follows that $b_n^{-1} = b_n^{-1} \int_{p_0\geq c_n} \hat{p}_n \leq b_n^{-1}\int_{p_0\geq c_n} (\hat{p}_n \vee c_n)=1$, and $b_n-1 = \int_{p_0\geq c_n} {\lvert(\hat{p}_n \vee c_n)-\hat{p}_n\rvert} \leq c_n {\lvert\{p_0\geq c_n\}\rvert}\lesssim c_n^{1+sd}$ by (\[ineq:s\_concave\_0\]). By the condition $s>-1/d$, by choosing $C_1>0$ large, we have $0\leq b_n-1\leq \mathcal{O}(n^{-1})$. This implies with high probability, $$\begin{aligned}
h^2(\hat{p}_n,\tilde{p}_n)\lesssim \int {\lvert\hat{p}_n-\tilde{p}_n\rvert}\leq {\lvert1-b_n^{-1}\rvert} \int \hat{p}_n + b_n^{-1} (b_n-1) = \mathcal{O}(n^{-1}).\end{aligned}$$ Hence by the integrability ${\mathbb{E}}_{P_0}\log^2 p_0<\infty$, with $P_0$ denoting the distribution of $p_0$, it follows that with high probability, $$\begin{aligned}
\label{ineq:s_concave_01}
h^2(p_0,\hat{p}_n)&\lesssim h^2(p_0 | n\equiv n^{-C_1}$, $ X_1,\ldots, X_n \in \{p_0\geq c_n\}$ with high probability. Let $ \tilde{p}_n\equiv (\hat{p}_n\vee c_n)\bm{1}_{p_0\geq c_n } / \int (\hat{p}_n\vee c_n)\bm{1}_{p_0 \geq c_n}$. Then with $ b_n \equiv \int (\hat{p}_n \vee c_n)\bm{1}_{p_0\geq c_n}$, it follows that $ b_n^{-1 } = b_n^{-1 } \int_{p_0\geq c_n } \hat{p}_n \leq b_n^{-1}\int_{p_0\geq c_n } (\hat{p}_n \vee c_n)=1 $, and $ b_n-1 = \int_{p_0\geq c_n } { \lvert(\hat{p}_n \vee c_n)-\hat{p}_n\rvert } \leq c_n { \lvert\{p_0\geq c_n\}\rvert}\lesssim c_n^{1+sd}$ by (\[ineq: s\_concave\_0\ ]). By the condition $ s>-1 / d$, by choose $ C_1>0 $ big, we have $ 0\leq b_n-1\leq \mathcal{O}(n^{-1})$. This implies with high probability, $ $ \begin{aligned }
h^2(\hat{p}_n,\tilde{p}_n)\lesssim \int { \lvert\hat{p}_n-\tilde{p}_n\rvert}\leq { \lvert1 - b_n^{-1}\rvert } \int \hat{p}_n + b_n^{-1 } (b_n-1) = \mathcal{O}(n^{-1}).\end{aligned}$$ therefore by the integrability $ { \mathbb{E}}_{P_0}\log^2 p_0<\infty$, with $ P_0 $ denoting the distribution of $ p_0 $, it follows that with high probability, $ $ \begin{aligned }
\label{ineq: s_concave_01 }
h^2(p_0,\hat{p}_n)&\lesssim h^2(p_0 | n\eqkiv n^{-C_1}$, $X_1,\ldots,X_n \in \{p_0\geq g_n\}$ with high proyqbilitb. Let $\tjlde{p}_n\eqjiv (\hat{p}_n\vee c_n)\bm{1}_{p_0\geq c_n} / \inv (\har{p}_n\vet c_n)\bm{1}_{p_0 \geq c_n}$. Then with $b_n \vquiv \int (\hat{k}_n \vee c_n)\bm{1}_{p_0\geq c_i}$, it follows tgwt $b_u^{-1} = b_n^{-1} \int_{p_0\geq c_n} \hat{p}_n \leq b_n^{-1}\int_{p_0\geq c_n} (\vag{p}_u \vee c_n)=1$, and $b_n-1 = \int_{p_0\geq c_n} {\lvert(\hat{[}_n \vee v_n)-\jat{p}_n\rvert} \leq c_n {\kdert\{l_0\geq c_n\}\rvert}\lesssim c_n^{1+sd}$ by (\[ineq:s\_ckncave\_0\]). Uy the conditiom $s>-1/d$, by choosing $C_1>0$ large, wf hage $0\leq b_n-1\leq \mathcwl{O}(n^{-1})$. This inpliqw with high orobabilitj, $$\begin{alighed}
h^2(\hat{p}_n,\tilde{p}_n)\lesssim \int {\lveft\hat{'}_n-\tilde{p}_n\rvgxr}\lee {\lvert1-b_n^{-1}\rvect} \int \hat{p}_n + b_n^{-1} (b_k-1) = \mathwal{O}(n^{-1}).\enc{aligned}$$ Hence by thw integrability ${\mathbu{E}}_{P_0}\log^2 p_0<\infty$, with $P_0$ denoting tke distribution of $p_0$, ut follmws dhat qitf hjgi pdobabipitb, $$\begin{alighed}
\label{inew:s_concave_01}
h^2(p_0,\hat{p}_n)&\lesxsyn h^2(p_0 | n\equiv n^{-C_1}$, $X_1,\ldots,X_n \in \{p_0\geq c_n\}$ with Let (\hat{p}_n\vee c_n)\bm{1}_{p_0\geq / \int (\hat{p}_n\vee $b_n \int (\hat{p}_n \vee c_n}$, it follows $b_n^{-1} = b_n^{-1} \int_{p_0\geq c_n} \hat{p}_n b_n^{-1}\int_{p_0\geq c_n} (\hat{p}_n \vee c_n)=1$, and $b_n-1 = \int_{p_0\geq c_n} {\lvert(\hat{p}_n \vee c_n)-\hat{p}_n\rvert} c_n {\lvert\{p_0\geq c_n\}\rvert}\lesssim c_n^{1+sd}$ by (\[ineq:s\_concave\_0\]). By the condition $s>-1/d$, by choosing $C_1>0$ we $0\leq \mathcal{O}(n^{-1})$. implies with high probability, $$\begin{aligned} h^2(\hat{p}_n,\tilde{p}_n)\lesssim \int {\lvert\hat{p}_n-\tilde{p}_n\rvert}\leq {\lvert1-b_n^{-1}\rvert} \int \hat{p}_n + b_n^{-1} (b_n-1) = \mathcal{O}(n^{-1}).\end{aligned}$$ by the integrability ${\mathbb{E}}_{P_0}\log^2 p_0<\infty$, with $P_0$ denoting distribution of $p_0$, it that with high probability, $$\begin{aligned} h^2(p_0,\hat{p}_n)&\lesssim | n\equiv n^{-C_1}$, $X_1,\ldots,X_n \in \{p_0\geq c_n\}$ wIth high proBabilIty. let $\TiLde{p}_N\equIv (\hat{p}_n\vee c_n)\bm{1}_{P_0\Geq c_N} / \int (\hat{p}_n\vee c_n)\bm{1}_{p_0 \geq c_n}$. then wItH $B_n \eqUIv \Int (\haT{p}_n \vee c_N)\Bm{1}_{P_0\GEq c_N}$, iT fOllOwS ThAt $b_n^{-1} = b_N^{-1} \inT_{p_0\geq c_n} \Hat{p}_n \leq b_n^{-1}\Int_{P_0\gEq c_n} (\hat{p}_n \vee C_N)=1$, aNd $b_n-1 = \int_{p_0\geQ c_n} {\Lvert(\hat{p}_n \veE c_n)-\Hat{p}_n\rVeRt} \lEQ c_n {\lvErt\{P_0\geq c_N\}\rvert}\LEsssim C_n^{1+sd}$ by (\[ineQ:s\_COncave\_0\]). bY the conDITiOn $s>-1/d$, By choosing $C_1>0$ large, wE HaVE $0\leq b_n-1\leq \mathcAl{O}(n^{-1})$. ThIs IMpLIEs wIth High probabIlIty, $$\beGIn{alignED}
h^2(\HAT{P}_n,\tILde{p}_n)\lesssim \iNt {\lvert\hat{p}_N-\TilDe{p}_n\rvErT}\leQ {\Lvert1-b_N^{-1}\rverT} \iNT \haT{p}_n + b_n^{-1} (b_n-1) = \mathCal{O}(N^{-1}).\end{alignEd}$$ HencE By the inTEgrabilIty ${\matHbb{e}}_{P_0}\lOg^2 p_0<\iNFtY$, wIth $p_0$ dENotINg The DIstRibution Of $P_0$, iT follOws tHAT WIth hIgh ProbAbiliTy, $$\begin{aligneD}
\laBel{iNEq:s_ConcaVe_01}
h^2(p_0,\hAt{p}_n)&\LeSssim H^2(p_0 | n\equiv n^{-C_1}$, $X_1,\l dots,X_n \ in \{ p_0 \ge qc_n\ }$ w ith high proba b ilit y. Let $\tilde{p}_n\eq uiv ( \h a t{p} _ n\ vee c _n)\bm{ 1 }_ { p _0\ ge qc_n }/ \ int ( \ha t{p}_n\ vee c_n)\b m{1 }_ {p_0 \geq c_ n }$ . Then wit h $ b_n \equiv \ int (\hat {p }_n \veec_n )\bm{ 1}_{p_ 0 \geq c _n}$, itfo l lows t h at $b_n ^ { -1 } =b_n^{-1} \int_{p_ 0 \g e q c_n} \hat{p} _n \le qb _n ^ { -1} \in t_{p_0\geq c _n} ( \ hat{p}_ n \ v e e c_ n )=1$, and $b_ n-1 = \int_ { p_0 \geq c _n } { \ lvert( \hat{ p} _ n \ vee c_n)-\h at{p }_n\rvert } \leq c_n {\l v ert\{p_ 0\geqc_n \}\ rver t }\ le sss im c_n ^ {1 +sd } $ b y (\[ine q: s\ _conc ave\ _ 0 \ ] ). B y t he c ondit ion $s>-1/d$, by cho o sin g $C_ 1>0$larg e, we h ave $0 \leqb_ n-1\leq \mathca l{O} (n^{-1})$ . T hi s i mp liesw ith hi ghpro babilit y, $$\b e gin {a l i g ne d}
h^2(\hat{p}_n,\ ti l d e{ p}_n)\le sssim\ in t{ \lvert\h at {p} _n-\ t i lde{p }_n\ r ve rt}\leq{\lver t 1- b_ n^{-1}\ rv ert} \ in t \ hat {p}_n + b_ n^{-1} (b_n-1) = \m a thcal{O}(n^{-1 } ).\end{aligne d }$ $ He n ce b y t he integrab ilit y ${\ math b b{ E}} _ {P_0} \log^ 2p _0 < \infty$, with $P_0$ d enotin g the distribution of $p_0$, i t follows tha t w i th high probab ility , $$\begin { aligned}
\lab el{ineq: s_concave _ 0 1}
h^2(p _0, \ha t{p }_n ) & \l esssim h^2(p_ 0 | n\equiv n^{-C_1}$,_$X_1,\ldots,X_n \in_\{p_0\geq c_n\}$ with high_probability. Let_$\tilde{p}_n\equiv_(\hat{p}_n\vee c_n)\bm{1}_{p_0\geq_c_n}_/ \int (\hat{p}_n\vee_c_n)\bm{1}_{p_0 \geq c_n}$._Then with $b_n \equiv_\int (\hat{p}_n \vee_c_n)\bm{1}_{p_0\geq_c_n}$, it follows that $b_n^{-1} = b_n^{-1} \int_{p_0\geq c_n} \hat{p}_n \leq b_n^{-1}\int_{p_0\geq c_n} (\hat{p}_n_\vee_c_n)=1$, and_$b_n-1_=_\int_{p_0\geq c_n} {\lvert(\hat{p}_n \vee c_n)-\hat{p}_n\rvert}_\leq c_n {\lvert\{p_0\geq c_n\}\rvert}\lesssim c_n^{1+sd}$_by (\[ineq:s\_concave\_0\])._By the condition $s>-1/d$, by choosing $C_1>0$ large,_we_have $0\leq b_n-1\leq_\mathcal{O}(n^{-1})$. This implies with high probability, $$\begin{aligned}
h^2(\hat{p}_n,\tilde{p}_n)\lesssim \int {\lvert\hat{p}_n-\tilde{p}_n\rvert}\leq_{\lvert1-b_n^{-1}\rvert} \int \hat{p}_n + b_n^{-1} (b_n-1)_= \mathcal{O}(n^{-1}).\end{aligned}$$ Hence_by_the_integrability ${\mathbb{E}}_{P_0}\log^2 p_0<\infty$, with_$P_0$ denoting the distribution of $p_0$,_it follows that with high probability,_$$\begin{aligned}
\label{ineq:s_concave_01}
h^2(p_0,\hat{p}_n)&\lesssim h^2(p_0 |
\left(
\begin{array}{cccc}
1&1&1&1\\
i&-i&i&-i\\
1&1&-1&-1\\
-i&i&i&-i
\end{array}\right)
=
\frac{1}{\sqrt{2}}
\left\{
\begin{array}{c}
|0_1 \rangle |0_0 \rangle + i |1_1 \rangle |1_0 \rangle \\
|0_1 \rangle |0_0 \rangle - i |1_1 \rangle |1_0 \rangle \\
|1_1 \rangle |0_0 \rangle + i |0_1 \rangle |1_0 \rangle \\
|1_1 \rangle |0_0 \rangle - i |0_1 \rangle |1_0 \rangle
\end{array}\right\}
\\
& = &
\mathcal{P}_2
\left[
\left\{\begin{array}{c}
|0_0\rangle\\
|1_0\rangle
\end{array}
\right\}
\otimes
\left\{\begin{array}{c}
|0_1\rangle\\
|1_1\rangle
\end{array}
\right\}
\right],\end{aligned}$$
$$\begin{aligned}
B_3
&=&
\frac{1}{2}
\left(
\begin{array}{cccc}
1&1&1&1\\
1&-1&1&-1\\
i&i&-i&-i
\\-i&i&i&-i
\end{array}
\right)
=
\frac{1}{\sqrt{2}}
\left\{
\begin{array}{c}
|0_0 \rangle |0_1 \rangle + i |1_0 \rangle |1_1 \rangle \\
|0_0 \rangle |1_1 \rangle + i |1_0 \rangle |0_1 \rangle \\
|0_0 \rangle |0_1 \rangle - i |1_0 \rangle |1_1 \rangle \\
|0_0 \rangle |1_1 \rangle - i |1_0 \rangle |0_1 \rangle
\ | \left (
\begin{array}{cccc }
1&1&1&1\\
i&-i&i&-i\\
1&1&-1&-1\\
-i&i&i&-i
\end{array}\right)
=
\frac{1}{\sqrt{2 } }
\left\ {
\begin{array}{c }
|0_1 \rangle |0_0 \rangle + i |1_1 \rangle |1_0 \rangle \\
|0_1 \rangle |0_0 \rangle - i |1_1 \rangle |1_0 \rangle \\
|1_1 \rangle |0_0 \rangle + i |0_1 \rangle |1_0 \rangle \\
|1_1 \rangle |0_0 \rangle - i |0_1 \rangle |1_0 \rangle
\end{array}\right\ }
\\
& = &
\mathcal{P}_2
\left [
\left\{\begin{array}{c }
|0_0\rangle\\
|1_0\rangle
\end{array }
\right\ }
\otimes
\left\{\begin{array}{c }
|0_1\rangle\\
|1_1\rangle
\end{array }
\right\ }
\right],\end{aligned}$$
$ $ \begin{aligned }
B_3
& = &
\frac{1}{2 }
\left (
\begin{array}{cccc }
1&1&1&1\\
1&-1&1&-1\\
i&i&-i&-i
\\-i&i&i&-i
\end{array }
\right)
=
\frac{1}{\sqrt{2 } }
\left\ {
\begin{array}{c }
|0_0 \rangle |0_1 \rangle + i |1_0 \rangle |1_1 \rangle \\
|0_0 \rangle |1_1 \rangle + i |1_0 \rangle |0_1 \rangle \\
|0_0 \rangle |0_1 \rangle - i |1_0 \rangle |1_1 \rangle \\
|0_0 \rangle |1_1 \rangle - i |1_0 \rangle |0_1 \rangle
\ | \lefh(
\begin{array}{cccc}
1&1&1&1\\
i&-i&i&-i\\
1&1&-1&-1\\
-i&i&i&-i
\ekd{array}\right)
=
\frac{1}{\sqrt{2}}
\lefv\{
\begin{adray}{c}
|0_1 \ravgle |0_0 \rangle + i |1_1 \rangle |1_0 \ranjle \\
|0_1 \rangoe |0_0 \rangle - i |1_1 \rangle |1_0 \rangle \\
|1_1 \gangle |0_0 \rqnglt + i |0_1 \rangle |1_0 \ranjme \\
|1_1 \rannje |0_0 \dwnglz - i |0_1 \rangle |1_0 \rakgle
\end{arraf}\right\}
\\
& = &
\mathcan{P}_2
\ueyt[
\left\{\begin{array}{c}
|0_0\rangle\\
|1_0\rangle
\end{arrar}
\right\}
\oyiles
\left\{\begin{artay}{c}
|0_1\rsggle\\
|1_1\dangle
\end{array}
\right\}
\right],\end{aligned}$$
$$\gegin{alpgned}
B_3
&=&
\frac{1}{2}
\left(
\begon{array}{cccc}
1&1&1&1\\
1&-1&1&-1\\
i&i&-i&-i
\\-i&i&i&-i
\end{arraj}
\rigjt)
=
\frac{1}{\sqrt{2}}
\left\{
\begij{array}{c}
|0_0 \rantle |0_1 \eangle + i |1_0 \rxngle |1_1 \rangle \\
|0_0 \rangle |1_1 \rangle + i |1_0 \rangle |0_1 \rangle \\
|0_0 \ranele |0_1 \xangle - i |1_0 \tcbglf |1_1 \rangle \\
|0_0 \rengle |1_1 \rangle - i |1_0 \vsngle |0_1 \rangle
\ | \left( \begin{array}{cccc} 1&1&1&1\\ i&-i&i&-i\\ 1&1&-1&-1\\ -i&i&i&-i \end{array}\right) \left\{ |0_1 \rangle \rangle + i |0_1 |0_0 \rangle - |1_1 \rangle |1_0 \\ |1_1 \rangle |0_0 \rangle + |0_1 \rangle |1_0 \rangle \\ |1_1 \rangle |0_0 \rangle - i |0_1 \rangle \rangle \end{array}\right\} \\ & = & \mathcal{P}_2 \left[ \left\{\begin{array}{c} |0_0\rangle\\ |1_0\rangle \end{array} \right\} \left\{\begin{array}{c} |1_1\rangle \right\} $$\begin{aligned} B_3 &=& \frac{1}{2} \left( \begin{array}{cccc} 1&1&1&1\\ 1&-1&1&-1\\ i&i&-i&-i \\-i&i&i&-i \end{array} \right) = \frac{1}{\sqrt{2}} \left\{ \begin{array}{c} \rangle |0_1 \rangle + i |1_0 \rangle |1_1 \\ |0_0 \rangle |1_1 + i |1_0 \rangle |0_1 \\ \rangle |0_1 - |1_0 |1_1 \rangle \\ \rangle |1_1 \rangle - i |1_0 \rangle |0_1 \rangle \ | \left(
\begin{array}{cccc}
1&1&1&1\\
i&-i&i&-i\\
1&1&-1&-1\\
-i&i&i&-I
\end{array}\rIght)
=
\fRac{1}{\SqrT{2}}
\lEft\{
\bEgin{Array}{c}
|0_1 \rangle |0_0 \raNGle + i |1_1 \Rangle |1_0 \rangle \\
|0_1 \rangle |0_0 \rangLe - i |1_1 \raNgLE |1_0 \ranGLe \\
|1_1 \RanglE |0_0 \rangle + I |0_1 \RaNGLe |1_0 \rAnGlE \\
|1_1 \raNgLE |0_0 \rAngle - I |0_1 \raNgle |1_0 \ranGle
\end{arraY}\riGhT\}
\\
& = &
\mathcal{P}_2
\lefT[
\LeFt\{\begin{arrAy}{c}
|0_0\Rangle\\
|1_0\rangle
\End{Array}
\rIgHt\}
\oTImes
\lEft\{\Begin{Array}{c}
|0_1\RAngle\\
|1_1\rAngle
\end{aRrAY}
\right\}
\RIght],\end{ALIgNed}$$
$$\bEgin{aligned}
B_3
&=&
\frac{1}{2}
\lEFt(
\BEgin{array}{cccc}
1&1&1&1\\
1&-1&1&-1\\
i&I&-i&-i
\\-i&i&i&-I
\eND{aRRAy}
\rIghT)
=
\frac{1}{\sqrt{2}}
\lEfT\{
\begiN{Array}{c}
|0_0 \rANgLE |0_1 \RAngLE + i |1_0 \rangle |1_1 \ranglE \\
|0_0 \rangle |1_1 \rangLE + i |1_0 \rAngle |0_1 \rAnGle \\
|0_0 \RAngle |0_1 \rAngle - I |1_0 \rANglE |1_1 \rangle \\
|0_0 \rangLe |1_1 \raNgle - i |1_0 \rangLe |0_1 \rangLE
\ | \left(
\begin{array}{cccc}
1&1&1&1\\
i&-i &i& -i\ \1&1& -1&- 1\\
-i&i&i&-i\ end{ array}\right)
=
\frac{ 1}{\s qr t {2}} \l eft\{
\begin { ar r a y}{ c}
| 0_1 \ r an gle | 0_0 \rangl e + i |1_1 \r an gle |1_0 \ra n gl e \\
|0_1\ra ngle |0_0 \r ang le - i | 1_1 \rang le|1_0\rangl e \\
|1 _1 \rangl e| 0_0 \r a ngle +i |0 _1 \ rangle |1_0 \rang l e\ \
|1_1 \rangle |0_0\r a ng l e -i | 0_1 \rangl e|1_0\ rangle\ en d { a rra y }\right\}
\\& = &
\math c al{ P}_2
\ le ft[ \left\ {\beg in { arr ay}{c}
|0_0 \ran gle\\
|1_ 0\rang l e
\end{ a rray}
\ right\ }
\ oti mes\ le ft \{\ be g in{ a rr ay} { c}|0_1\ran gl e\ \
|1_ 1\ra n g l e
\en d{a rray }
\ri ght\}
\right] ,\e nd{a l ign ed}$$
$$\ begi n{ align ed}
B_ 3
&=&
\ frac{1}{2}
\lef t(
\ begin{arr ay} {c ccc }1&1&1 & 1\\
1& -1& 1&- 1\\
i&i &-i&-i\ \-i &i & i & -i
\end{array}
\righ t) =
\ frac{1}{ \sqrt{ 2 }}
\ l eft\{
\b eg in{ arra y } {c}
| 0_0\ ra ngle |0_ 1 \ran g le + i |1_0 \ rangle | 1_1 \r angle \\
| 0_0 \r angle |1 _1 \r a ngle + i |1_0\ rangle |0_1 \ r an g l e\ \
|0 _0\rangle |0_ 1 \r a ngle - i |1 _0\ rangl e |1_ 1\ ra n gle \\
|0_0 \rangle | 1_1 \r angle - i |1_0 \ra ngle |0_1\ r a ngle
\ | \left(
\begin{array}{cccc}
1&1&1&1\\
i&-i&i&-i\\
1&1&-1&-1\\
-i&i&i&-i
\end{array}\right)
=
\frac{1}{\sqrt{2}}
\left\{
\begin{array}{c}
|0_1 \rangle_|0_0 \rangle_+ i |1_1 \rangle_|1_0 \rangle_\\
|0_1_\rangle |0_0_\rangle_- i |1_1_\rangle |1_0 \rangle_\\
|1_1 \rangle |0_0 \rangle_+ i |0_1_\rangle_|1_0 \rangle \\
|1_1 \rangle |0_0 \rangle - i |0_1 \rangle |1_0 \rangle
\end{array}\right\}
\\
& = &
\mathcal{P}_2
\left[
\left\{\begin{array}{c}
|0_0\rangle\\
|1_0\rangle
\end{array}
\right\}
\otimes
\left\{\begin{array}{c}
|0_1\rangle\\
|1_1\rangle
\end{array}
\right\}
\right],\end{aligned}$$
$$\begin{aligned}
B_3
&=&
\frac{1}{2}
\left(
\begin{array}{cccc}
1&1&1&1\\
1&-1&1&-1\\
i&i&-i&-i
\\-i&i&i&-i
\end{array}
\right)
=
\frac{1}{\sqrt{2}}
\left\{
\begin{array}{c}
|0_0_\rangle_|0_1 \rangle_+_i_|1_0 \rangle |1_1 \rangle \\
|0_0_\rangle |1_1 \rangle + i_|1_0 \rangle_|0_1 \rangle \\
|0_0 \rangle |0_1 \rangle - i_|1_0_\rangle |1_1 \rangle_\\
|0_0 \rangle |1_1 \rangle - i |1_0 \rangle |0_1_\rangle
\ |
this category. We show that the fractal sets $\Omega$ underlying these staircases are connected with key sets in Number Theory via their $(\alpha, f(\alpha))$ multifractal decomposition spectra. It follows that such sets $\Omega$ are self similar when the usual (Euclidean) measure is replaced by the hyperbolic measure induced by the Farey-Brocot partition.'
author:
- |
Grynberg, S.[^1] and Piacquadio, M.\
Departamento de Matemática\
Facultad de Ingeniería\
Universidad de Buenos Aires\
Paseo Colón 850, (1063)\
Buenos Aires, Argentina.
title: 'Self-similarity of Farey Staircases'
---
Introduction
============
A Cantor staircase is an increasing continuous function from $[a,b]$ to $[0,1]$, $y=g(x)$, with zero derivative almost everywhere, constant on the so-called intervals of resonance or stability $\Delta x_k, k\in\Bbb N$. The complement in $[a,b]$ of $\bigcup_{k\in\Bbb N}\Delta x_k$ is a totally disconnected Cantordust set $\Omega$.
Cantor staircases are frequently observed in empirical physics, and their universal properties are of great interest. These staircases are naturally associated with a Cantordust set $\Omega$. Such an $\Omega$ reflects the particular physical problem under study.
Such Cantordusts can be studied with the tools provided by Number Theory and the multidimensional $(\alpha, f(\alpha))$ decomposition of a fractal set $\Omega$.
The Ising Model and the Circle Map
----------------------------------
1. [**The Ising model.**]{} Bruinsma and Bak \[1983\] studied the one dimensional Ising model with convex long-range antiferromagnetic interaction. Only “up” spins interact, their interaction being given by a convex function depending on a parameter $a>1$, $a$ is the strength of the interaction. Let $H$ be the applied magnetic field and $q$ the proportion of up-spins. At the critical temperature $T=0$ the phase diagram $q=g(-H)$ exhibits a Cantor staircase.
With $\Delta H$ we will denote the intervals of resonance or stability of the staircase $q=g(-H)$. [* | this category. We show that the fractal sets $ \Omega$ underlying these stairway are connect with key sets in numeral Theory via their $ (\alpha, f(\alpha))$ multifractal decay spectra. It follows that such sets $ \Omega$ are self alike when the usual (Euclidean) bill is replaced by the hyperbolic measure induce by the Farey - Brocot partition.'
author:
- |
Grynberg, S.[^1 ] and Piacquadio, M.\
Departamento de Matemática\
Facultad de Ingeniería\
Universidad de Buenos Aires\
Paseo Colón 850, (1063)\
Buenos Aires, Argentina.
deed:' Self - similarity of Farey Staircases'
---
initiation
= = = = = = = = = = = =
A Cantor staircase is an increasing continuous function from $ [ a, b]$ to $ [ 0,1]$, $ y = g(x)$, with zero derivative almost everywhere, changeless on the so - called intervals of rapport or stability $ \Delta x_k, k\in\Bbb N$. The complement in $ [ a, b]$ of $ \bigcup_{k\in\Bbb N}\Delta x_k$ is a totally disconnected Cantordust set $ \Omega$.
Cantor staircases are frequently detect in empirical physics, and their universal properties are of great sake. These staircases are naturally associated with a Cantordust set $ \Omega$. Such an $ \Omega$ reflects the particular physical problem under study.
Such Cantordusts can be studied with the tools provided by Number Theory and the multidimensional $ (\alpha, f(\alpha))$ decomposition of a fractal set $ \Omega$.
The Ising Model and the Circle Map
----------------------------------
1. [ * * The Ising model. * * ] { } Bruinsma and Bak \[1983\ ] studied the one dimensional Ising model with convex retentive - range antiferromagnetic interaction. Only “ up ” spin interact, their interaction being give by a convex function depending on a argument $ a>1 $, $ a$ is the strength of the interaction. lease $ H$ be the applied magnetic airfield and $ q$ the proportion of up - spins. At the critical temperature $ T=0 $ the phase diagram $ q = g(-H)$ exhibits a Cantor staircase.
With $ \Delta H$ we will denote the interval of resonance or stability of the staircase $ q = g(-H)$. [ * | thls category. We show that the fractal sejs $\Omega$ undermying thdse staircases are connected wuth ktj sets in Number Theofy via thvir $(\alpha, f(\alkha))$ multifractal vscomposlcion aiectrc. Mt follows that such sets $\Omega$ are selx rililar when the usual (Euclidean) meastre is teolaced by the rypegbjlic measure induced by the Farey-Brocof partiuion.'
author:
- |
Grymberg, S.[^1] and Piacquadio, M.\
Deoartamento de Matelática\
Faxultwe de Ingeniefía\
Univtrvidad de Boenos Aires\
Paseo Colón 850, (1063)\
Bjenos Aires, Arggurinw.
jitle: 'Self-sinilarpty of Farey Staircasas'
---
Introcuction
============
A Cantov stamrcawe is an increasing cmntinuous function from $[a,b]$ do $[0,1]$, $y=g(x)$, with zero dwruvatiee anmosg evdryshxre, constwnt on the so-dalled inteevals of resonance pr wtability $\Delfa x_k, h\ig\Bbb N$. The complement in $[a,b]$ of $\bigcup_{k\it\Bbg N}\Delta x_k$ is a totallt disconnected Cantorfust set $\Jmega$.
Cantor staircases are frequently observed in empicizal pmysizw, wnd their universal properties are of great igfetext. These stairgases are naturallu wsxjciated with x Cantordhst set $\Omega$. Such an $\Omeda$ redlects tht paryicular physical problem uneer study.
Suck Cqntordusts can be dtudied witk the jools lrovided by Number Theoxy and the multidlmensionam $(\alpha, f(\alpha))$ dezomiosidion of a fractal set $\Omegw$.
The Isinj Modzl and tfe Corcle Iap
----------------------------------
1. [**The Iding model.**]{} Bruinsma and Hak \[1983\] dtgdied the lne dimensional Ising model witi convex long-tance dntiferrjmagnctic interactiog. Only “up” spinx intercct, thdir interabtion beiig given by w convex funcdlon dependinj on a pawamerer $q>1$, $a$ is gfe strength of the integaetion. Let $H$ be the applied kagvstic field and $w$ tye proportion og uo-spynd. Et thq critical tekperxtufr $T=0$ tfe phast dlagfam $a=g(-H)$ exhibits a Cantos stzircase.
With $\Dektw H$ we wull denoee the intervsls of resonance og stauility of tne ftaircase $q=g(-H)$. [* | this category. We show that the fractal underlying staircases are with key sets $(\alpha, multifractal decomposition spectra. follows that such $\Omega$ are self similar when the (Euclidean) measure is replaced by the hyperbolic measure induced by the Farey-Brocot partition.' - | Grynberg, S.[^1] and Piacquadio, M.\ Departamento de Matemática\ Facultad de Ingeniería\ de Aires\ Colón (1063)\ Buenos Aires, Argentina. title: 'Self-similarity of Farey Staircases' --- Introduction ============ A Cantor staircase is increasing continuous function from $[a,b]$ to $[0,1]$, $y=g(x)$, zero derivative almost everywhere, on the so-called intervals of or $\Delta x_k, N$. complement $[a,b]$ of $\bigcup_{k\in\Bbb x_k$ is a totally disconnected Cantordust set $\Omega$. Cantor staircases are frequently observed in empirical physics, and universal properties great interest. staircases naturally with a Cantordust Such an $\Omega$ reflects the particular study. Such Cantordusts can be studied with the provided by Theory and the multidimensional $(\alpha, f(\alpha))$ of a fractal set $\Omega$. The Ising Model the Circle Map ---------------------------------- 1. [**The Ising model.**]{} Bruinsma and Bak \[1983\] studied the one model with convex long-range interaction. Only “up” interact, interaction given a convex depending on a parameter $a>1$, $a$ is the strength of the Let $H$ be the applied magnetic field and $q$ the up-spins. the critical temperature the phase diagram $q=g(-H)$ a staircase. With $\Delta H$ denote intervals stability the $q=g(-H)$. [* | this category. We show that the Fractal setS $\OmegA$ unDerLyIng tHese Staircases are cONnecTed with key sets in Number theorY vIA theIR $(\aLpha, f(\Alpha))$ muLTiFRActAl DeComPoSItIon spEctRa. It folLows that suCh sEtS $\Omega$ are selF SiMilar when tHe uSual (EuclideaN) meAsure iS rEplACed by The HyperBolic mEAsure iNduced by tHe fArey-BrOCot partITIoN.'
autHor:
- |
Grynberg, S.[^1] and PiACqUAdio, M.\
DepartameNto de MAtEMáTICa\
FAcuLtad de IngeNiEría\
UNIversidAD dE bUEnoS aires\
Paseo ColÓn 850, (1063)\
Buenos AirES, ArGentinA.
tItlE: 'self-siMilarItY Of FArey StaircaSes'
---
INtroductiOn
============
A CanTOr stairCAse is an IncreaSinG coNtinUOuS fUncTiON frOM $[a,B]$ to $[0,1]$, $Y=G(x)$, wIth zero dErIvAtive AlmoST EVErywHerE, conStant On the so-called IntErvaLS of ResonAnce oR staBiLity $\DElta x_k, K\in\BbB N$. the complement in $[A,b]$ of $\Bigcup_{k\in\bbb n}\DEltA x_K$ is a tOTally dIscOnnEcted CaNtordusT Set $\omEGA$.
caNtor staircases are fReQUEnTly obserVed in eMPiRiCAl physicS, aNd tHeir UNIversAl prOPeRties are Of greaT InTeRest. TheSe StaircAsEs aRe nAturaLLy asSociatEd with a CAntorDUst set $\Omega$. SucH An $\Omega$ reflecTS tHE PaRTicuLar Physical proBlem UNder StudY.
suCh CANtordUsts cAn BE sTUdied with the tools prOvIded by numbeR Theory and the MultidimenSIONal $(\alpha, F(\alpHA))$ dEComposition of a FractAl set $\Omega$.
tHe Ising MOdel aNd the CirCle Map
----------------------------------
1. [**The iSIng model.**]{} bruInsMa aNd BAK \[1983\] StUdied the one diMENsioNaL Ising mOdeL with coNveX loNg-rAngE aNtiferromAgnetic iNtErAcTiOn. ONly “up” SPins inteRaCt, tHeIr iNteraCTion beIng giVen bY a CoNVex FunctioN DePENdinG oN a ParaMetEr $A>1$, $a$ is tHe stREngTh of the InteractiOn. LET $H$ be ThE aPplied mAgnetic field aNd $Q$ the proporTiOn oF up-spiNS. at the criTical temperature $T=0$ the phaSE diagraM $q=g(-h)$ exhiBits A Cantor stAirCase.
WiTh $\DELta H$ we Will deNote tHe IntERVals oF REsOnaNcE or stabiliTY Of tHe staIrCase $Q=g(-H)$. [* | this category. We show th at the fra ctalset s $ \O mega $ un derlying these stai rcases are connected w ith k ey sets in Numb er Theo r yv i a t he ir $( \a l ph a, f( \al pha))$multifract alde compositions pe ctra. It f oll ows that suc h s ets $\ Om ega $ aresel f sim ilar w h en the usual (E uc l idean) measure i srepl aced by the hyper b ol i c measure indu ced by t h eF a rey -Br ocot parti ti on.'a uthor:- | G r ynberg, S.[^1 ] and Piacq u adi o, M.\
D eparta mento d e Ma temática\
F acultad d e Inge n iería\ Univ ersida d d e B ueno s A ir es\
P a se o C o lón 850, (1 06 3) \
Bue n o s Aire s,Arge ntina .
title: 'Sel f-s imil a rit y ofFarey Sta ir cases '
---
Intr od uction
======== ====
A Canto r s ta irc as e isa n incr eas ing contin uous fu n cti on f r om $[a,b]$ to $[0,1] $, $ y= g(x)$, w ith ze r ode r ivativeal mos t ev e r ywher e, c o ns tant onthe so - ca ll ed inte rv als of r eso nan ce or stab ility$\Deltax_k,k \in\Bbb N$. Th e complement i n $ [ a ,b ] $ of $\ bigcup_{k\i n\Bb b N}\ Delt a x _k$ is atotal ly di s connected Cantordus tset $\ Omega $.
Cantor st aircases a r e frequent ly o b se r ved in empiric al ph ysics, and their un ivers al prope rties are o f greatint ere st. Th e s estaircases ar e natu ra lly ass oci ated wi tha C ant ord us t set $\O mega$. S uc han $ \Om ega$r eflectsth e p ar tic ularp hysica l pro blem u nd e r s tudy.
S uc h Cant or du stscan b e stu died wit h the t ools prov ide d byNu mb er Theo ry and the mu lt idimension al $( \alpha , f(\alpha ))$ decomposition of af ractalset $\Om ega$ .
The Is ing Model an d the C ircleMap
- -- --- - - ----- - - -- --- -- ---------- - -
1 . [* *T he I sing mo del.**]{} Bruinsma and Bak \[1983\] st udie d th e o n ed ime ns i ona l Ising model wit h convex l on g -r ange antif e rro ma gneticinterac tion. Only “u p” spinsinteract, t heir i nte raction be ing give n by a co n vex f u nc tiondep ending o n a para meter$ a>1 $, $a $ is t he stren gth o fthe inte raction. Let $H$ be the appli ed ma gne tic field an d $q $ the pro port ion of up- spi ns. At t hec ritic al t e mp era t ure $ T=0$ the phase di agr a m $ q=g(-H)$ ex h i b its a Ca nto r stair case .
With $\Del t a H$ we will d enot e the in t erva ls of resonanceorst a b ility of t he staircas e $q=g(- H) $ . [* | this_category. We_show that the fractal_sets $\Omega$_underlying_these staircases_are_connected with key_sets in Number_Theory via their $(\alpha,_f(\alpha))$ multifractal decomposition_spectra._It follows that such sets $\Omega$ are self similar when the usual (Euclidean) measure_is_replaced by_the_hyperbolic_measure induced by the Farey-Brocot_partition.'
author:
- |
_Grynberg, S.[^1]_and Piacquadio, M.\
Departamento de_Matemática\
_ _Facultad de Ingeniería\
Universidad de Buenos_Aires\
Paseo Colón_850, (1063)\
___Buenos Aires, Argentina.
title: 'Self-similarity_of Farey Staircases'
---
Introduction
============
A Cantor staircase is_an increasing continuous function from $[a,b]$_to $[0,1]$, $y=g(x)$, with zero derivative almost_everywhere, constant on the so-called intervals_of resonance or stability $\Delta_x_k, k\in\Bbb_N$. The complement in $[a,b]$_of $\bigcup_{k\in\Bbb N}\Delta_x_k$ is_a totally disconnected_Cantordust set $\Omega$.
Cantor staircases are frequently_observed in empirical_physics, and their universal properties are_of_great interest. These_staircases_are_naturally associated_with a Cantordust_set_$\Omega$. Such_an_$\Omega$ reflects the particular physical problem_under_study.
Such Cantordusts can be studied with the_tools provided by Number_Theory_and the multidimensional $(\alpha,_f(\alpha))$ decomposition of a fractal_set $\Omega$.
The Ising Model and the_Circle Map
----------------------------------
1._ [**The_Ising model.**]{} Bruinsma and Bak \[1983\] studied the one dimensional Ising_model with convex long-range antiferromagnetic interaction._Only “up” spins interact,_their interaction_being_given by a_convex_function depending_on a parameter $a>1$, $a$ is the_strength of_the interaction. Let $H$ be the_applied magnetic field and_$q$_the proportion of up-spins. At the_critical temperature $T=0$ the phase diagram_$q=g(-H)$ exhibits a Cantor staircase.
___ With $\Delta H$ we_will denote the intervals of resonance_or stability of_the staircase $q=g(-H)$. [* |
,c}(x) = H_{l,-1}(-cx) = - l \left [(-cx)\log{(-cx)}+(1+cx)\log{(1+cx)}\right ]+$$
$$\int _0 ^1 \frac{-t}{\log{(1-t)}} \frac{(1+cxt)^l+(1-t-cxt)^l-1-(1-t)^l}{t^2}dt.$$
It is a matter of calculus to prove that
$$\begin{aligned}
H''_{n,c}(x) &=& cl \left ( \frac{1}{x} - \frac{c}{1+cx}\right ) \\&+& c^2l(l-1)\int _0 ^1 \frac{-t}{\log{(1-t)}} \left [ (1+cxt)^{l-2} + (1-t-cxt)^{l-2}\right ] dt,\end{aligned}$$
and for $k\geq 0$
$$\begin{aligned}
&&H_{n,c}^{(2k+2)}(x)=cl(2k)! \left ( \frac{1}{x^{2k+1}} - \left ( \frac{c}{1+cx}\right )^{2k+1} \right )\\
&+& l(l-1)\dots (l-2k-1)c^{2k+2}\\&& \int _0 ^1 \frac{-t}{\log{(1-t)}} \left [ (1+cxt)^{l-2k-2} + (1-t-cxt)^{l-2k-2}\right ]t^{2k} dt.\end{aligned}$$
For $0<t<1$ we have $$0<\frac{-t}{\log{(1-t)}}<1, \label{eq:new7}$$ so that $$H_{n,c}^{(2k+2)}(x) \leq cl(2k)! \left ( \frac{1}{x^{2k+1}} - \left ( \frac{c}{1+cx}\right )^{2k+1} \right )+\label{eq:7}$$ $$+ l(l-1)\dots (l-2k-1)c^{2k+2} \int _0 ^1 \ | , c}(x) = H_{l,-1}(-cx) = - l \left [ (-cx)\log{(-cx)}+(1+cx)\log{(1+cx)}\right ] + $ $
$ $ \int _ 0 ^1 \frac{-t}{\log{(1 - t) } } \frac{(1+cxt)^l+(1 - t - cxt)^l-1-(1 - t)^l}{t^2}dt.$$
It is a matter of calculus to rise that
$ $ \begin{aligned }
H''_{n, c}(x) & = & cl \left (\frac{1}{x } - \frac{c}{1+cx}\right) \\&+ & c^2l(l-1)\int _ 0 ^1 \frac{-t}{\log{(1 - t) } } \left [ (1+cxt)^{l-2 } + (1 - metric ton - cxt)^{l-2}\right ] dt,\end{aligned}$$
and for $ k\geq 0 $
$ $ \begin{aligned }
& & H_{n, c}^{(2k+2)}(x)=cl(2k)! \left (\frac{1}{x^{2k+1 } } - \left (\frac{c}{1+cx}\right) ^{2k+1 } \right) \\
& + & l(l-1)\dots (l-2k-1)c^{2k+2}\\ & & \int _ 0 ^1 \frac{-t}{\log{(1 - t) } } \left [ (1+cxt)^{l-2k-2 } + (1 - t - cxt)^{l-2k-2}\right ] t^{2k } dt.\end{aligned}$$
For $ 0 < t<1 $ we have $ $ 0<\frac{-t}{\log{(1 - t)}}<1, \label{eq: new7}$$ so that $ $ H_{n, c}^{(2k+2)}(x) \leq cl(2k)! \left (\frac{1}{x^{2k+1 } } - \left (\frac{c}{1+cx}\right) ^{2k+1 } \right) + \label{eq:7}$$ $ $ + l(l-1)\dots (l-2k-1)c^{2k+2 } \int _ 0 ^1 \ | ,c}(x) = H_{l,-1}(-cx) = - l \left [(-cx)\log{(-cx)}+(1+cx)\uog{(1+cx)}\right ]+$$
$$\int _0 ^1 \frac{-t}{\nog{(1-t)}} \fdac{(1+cxt)^l+(1-t-zxt)^l-1-(1-t)^l}{t^2}dt.$$
It is a matter of celcuous ti prove that
$$\begin{aligndd}
H''_{n,c}(x) &=& cp \left ( \drac{1}{e} - \frac{c}{1+cx}\right ) \\&+& c^2l(l-1)\inb _0 ^1 \fdwc{-t}{\lmj{(1-t)}} \left [ (1+cxt)^{l-2} + (1-j-cxt)^{l-2}\right ] gt,\end{aligned}$$
ang wox $k\geq 0$
$$\begin{aligned}
&&H_{n,c}^{(2k+2)}(x)=cl(2k)! \left ( \frwc{1}{x^{2k+1}} - \lrfh ( \frac{c}{1+cx}\righj )^{2k+1} \rpgrt )\\
&+& m(l-1)\dots (l-2k-1)c^{2k+2}\\&& \int _0 ^1 \frac{-t}{\log{(1-t)}} \left [ (1+dxt)^{l-2k-2} + (1-u-cxt)^{l-2k-2}\right ]t^{2k} dt.\emd{aligned}$$
For $0<t<1$ we have $$0<\fraf{-t}{\loh{(1-t)}}<1, \label{eq:new7}$$ so tjat $$H_{n,c}^{(2k+2)}(x) \lgs cj(2j)! \left ( \frac{1}{b^{2k+1}} - \left ( \frac{c}{1+cx}\right )^{2k+1} \right )+\label{eq:7}$$ $$+ l(l-1)\dots (l-2k-1)c^{2k+2} \ivt _0 ^1 \ | ,c}(x) = H_{l,-1}(-cx) = - l \left $$\int ^1 \frac{-t}{\log{(1-t)}} It is a that H''_{n,c}(x) &=& cl ( \frac{1}{x} - ) \\&+& c^2l(l-1)\int _0 ^1 \frac{-t}{\log{(1-t)}} [ (1+cxt)^{l-2} + (1-t-cxt)^{l-2}\right ] dt,\end{aligned}$$ and for $k\geq 0$ $$\begin{aligned} &&H_{n,c}^{(2k+2)}(x)=cl(2k)! \left \frac{1}{x^{2k+1}} - \left ( \frac{c}{1+cx}\right )^{2k+1} \right )\\ &+& l(l-1)\dots (l-2k-1)c^{2k+2}\\&& \int _0 \frac{-t}{\log{(1-t)}} [ + ]t^{2k} dt.\end{aligned}$$ For $0<t<1$ we have $$0<\frac{-t}{\log{(1-t)}}<1, \label{eq:new7}$$ so that $$H_{n,c}^{(2k+2)}(x) \leq cl(2k)! \left ( \frac{1}{x^{2k+1}} \left ( \frac{c}{1+cx}\right )^{2k+1} \right )+\label{eq:7}$$ $$+ l(l-1)\dots \int _0 ^1 \ | ,c}(x) = H_{l,-1}(-cx) = - l \left [(-cx)\log{(-cx)}+(1+cx)\log{(1+cx)}\Right ]+$$
$$\int _0 ^1 \frAc{-t}{\loG{(1-t)}} \fRac{(1+CxT)^l+(1-t-cXt)^l-1-(1-t)^L}{t^2}dt.$$
It is a matteR Of caLculus to prove that
$$\begin{AlignEd}
h''_{N,c}(x) &=& cL \LeFt ( \fraC{1}{x} - \frac{c}{1+CX}\rIGHt ) \\&+& c^2L(l-1)\InT _0 ^1 \frAc{-T}{\LoG{(1-t)}} \lefT [ (1+cxT)^{l-2} + (1-t-cxt)^{l-2}\Right ] dt,\end{AliGnEd}$$
and for $k\geq 0$
$$\BEgIn{aligned}
&&H_{N,c}^{(2k+2)}(X)=cl(2k)! \left ( \frac{1}{X^{2k+1}} - \lEft ( \fraC{c}{1+Cx}\rIGht )^{2k+1} \rIghT )\\
&+& l(l-1)\doTs (l-2k-1)c^{2k+2}\\&& \INt _0 ^1 \frac{-T}{\log{(1-t)}} \left [ (1+CxT)^{L-2k-2} + (1-t-cxt)^{L-2K-2}\right ]t^{2K} DT.\eNd{alIgned}$$
For $0<t<1$ we have $$0<\frAC{-t}{\LOg{(1-t)}}<1, \label{eq:new7}$$ sO that $$H_{N,c}^{(2K+2)}(X) \lEQ Cl(2k)! \LefT ( \frac{1}{x^{2k+1}} - \lefT ( \fRac{c}{1+cX}\Right )^{2k+1} \rIGhT )+\LABel{EQ:7}$$ $$+ l(l-1)\dots (l-2k-1)c^{2k+2} \inT _0 ^1 \ | ,c}(x) = H_{l,-1}(-cx) = - l \left [ (-cx) \lo g{( -c x)}+ (1+c x)\log{(1+cx)} \ righ t ]+$$
$$\int _0 ^1 \ frac{ -t } {\lo g {( 1-t)} } \frac { (1 + c xt) ^l +( 1-t -c x t) ^l-1- (1- t)^l}{t ^2}dt.$$
Itis a matter of ca lculus topro ve that
$$\ beg in{ali gn ed} H''_{ n,c }(x)&=& cl \left( \frac{1 }{ x } - \f r ac{c}{1 + c x} \rig ht ) \\&+& c^2l( l -1 ) \int _0 ^1 \fr ac{-t} {\ l og { ( 1-t )}} \left [ ( 1+ cxt)^ { l-2} +( 1- t - c xt) ^ {l-2}\right ] dt,\end{al i gne d}$$
an d f o r $k\g eq 0$
$ $\b egin{aligne d}
& &H_{n,c}^ {(2k+2 ) }(x)=cl ( 2k)! \l eft (\fr ac{ 1}{x ^ {2 k+ 1}} - \le f t( \ f rac {c}{1+cx }\ ri ght ) ^{2k + 1 } \rig ht)\\&+& l (l-1)\dots (l -2k -1)c ^ {2k +2}\\ && \i nt _ 0^1 \f rac{-t }{\lo g{ (1-t)}} \left [ (1+ cxt)^{l-2 k-2 }+ ( 1- t-cxt ) ^{l-2k -2} \ri ght ]t^ {2k} dt . \en d{ a l i gn ed}$$
For $0<t<1$ w e ha ve $$0<\ frac{- t }{ \l o g{(1-t)} }< 1,\lab e l {eq:n ew7} $ $so that$$H_{n , c} ^{ (2k+2)} (x ) \leq c l(2 k)! \lef t ( \ frac{1 }{x^{2k+ 1}} - \left ( \frac{ c }{1+cx}\right )^ { 2 k+ 1 } \r igh t )+\label{ eq:7 } $$ $ $+ l ( l- 1)\ d ots ( l-2k- 1) c ^{ 2 k+2} \int _0 ^1 \ | ,c}(x) =_H_{l,-1}(-cx) =_- l \left [(-cx)\log{(-cx)}+(1+cx)\log{(1+cx)}\right_]+$$
$$\int _0_^1_\frac{-t}{\log{(1-t)}} \frac{(1+cxt)^l+(1-t-cxt)^l-1-(1-t)^l}{t^2}dt.$$
It_is_a matter of_calculus to prove_that
$$\begin{aligned}
H''_{n,c}(x) &=& cl \left_( \frac{1}{x} -_\frac{c}{1+cx}\right_) \\&+& c^2l(l-1)\int _0 ^1 \frac{-t}{\log{(1-t)}} \left [ (1+cxt)^{l-2} + (1-t-cxt)^{l-2}\right ] dt,\end{aligned}$$
and_for_$k\geq 0$
$$\begin{aligned}
&&H_{n,c}^{(2k+2)}(x)=cl(2k)!_\left_(_\frac{1}{x^{2k+1}} - \left ( \frac{c}{1+cx}\right_)^{2k+1} \right )\\
&+& l(l-1)\dots (l-2k-1)c^{2k+2}\\&&_\int _0_^1 \frac{-t}{\log{(1-t)}} \left [ (1+cxt)^{l-2k-2} + (1-t-cxt)^{l-2k-2}\right ]t^{2k}_dt.\end{aligned}$$
For_$0<t<1$ we have_$$0<\frac{-t}{\log{(1-t)}}<1, \label{eq:new7}$$ so that $$H_{n,c}^{(2k+2)}(x) \leq cl(2k)! \left (_\frac{1}{x^{2k+1}} - \left ( \frac{c}{1+cx}\right )^{2k+1}_\right )+\label{eq:7}$$ $$+_l(l-1)\dots_(l-2k-1)c^{2k+2}_\int _0 ^1 _\ |
the overlap corresponding to spin-uncorrelated noise essentially follows the purely spatial overlap, $\Omega_s\sim\Omega_\mathrm{max}$. All forms of the overlap reach a common minimum just beyond the critical field $B\gtrsim B_c$, and then *grow* together in an oscillatory fashion, out of phase with the splitting (blue curve). Namely, MZMs deep in the topological regime [enjoy *less protection*]{} than the ones near $B_c$ (compare wavefunctions at $B_2$ and $B_3$).
In the two uniform nanowires simulated in panels (a,b) we see how the estimator $\eta$ (black curve) roughly traces the overlap $\Omega_s$, particularly around the $B\gtrsim B_c$ region of minimal overlap. This good correlation corresponds to two particular configurations of the uniform Lutchyn-Oreg model. To fully assess the overall correlation between $\eta$ and $\Omega_s$ for arbitrary configurations, we simulate an ensemble of $\sim 2\cdot 10^4$ uniform nanowires with varying model parameters, including $B$, $\alpha$, $L$, $\mu$ and $\Delta$, distributed uniformly within realistic ranges (see caption). Amongst all configurations, we select those with a near zero mode (splitting $E_0<10\mu$eV) well separated from higher excitations (second eigenvalue greater than $50\mu$eV). This preselection is experimentally feasible at temperatures below $\sim 100$mK using local tunnel spectroscopy. In the uniform case it excludes in particular any field in the trivial regime $B<B_c$. We then compute the $(\eta, \Omega_s)$ pair for the near-zero modes in the ensemble and collect all these points (see Fig. \[fig:uniform\]c) to build the probability density $P(\eta,\Omega_s)$ that a given zero mode with a measured $\eta$ have a given $\Omega_s$. The probability $P(\eta,\Omega_s)$ is shown in Fig. \[fig:uniform\]d. Its profile gives an accurate account of the quality of $\eta$ as an estimator of $\Omega_s$ within the whole space of uniform Lutchyn-Oreg nanowire models with near zero modes. A perfect correlation would appear as a straight, thin $P(\eta,\Omega_s)$ along the diagonal. We see that while the actual dependence of the typical $\Omega | the overlap corresponding to spin - uncorrelated noise basically take after the purely spatial overlap, $ \Omega_s\sim\Omega_\mathrm{max}$. All form of the lap reach a common minimum just beyond the critical plain $ B\gtrsim B_c$, and then * grow * together in an oscillatory manner, out of phase with the splitting (aristocratic curve). Namely, MZMs deeply in the topological regime [ enjoy * less protection * ] { } than the ones near $ B_c$ (comparison wavefunctions at $ B_2 $ and $ B_3 $).
In the two uniform nanowires simulated in panel (a, b) we see how the estimator $ \eta$ (bootleg curve) roughly traces the lap $ \Omega_s$, particularly around the $ B\gtrsim B_c$ region of minimal overlap. This good correlation match to two particular configurations of the uniform Lutchyn - Oreg model. To fully measure the overall correlation between $ \eta$ and $ \Omega_s$ for arbitrary configurations, we simulate an ensemble of $ \sim 2\cdot 10 ^ 4 $ uniform nanowires with varying model parameters, including $ B$, $ \alpha$, $ L$, $ \mu$ and $ \Delta$, distributed uniformly within realistic ranges (see caption). Amongst all configurations, we select those with a near zero manner (splitting $ E_0<10\mu$eV) well separated from high excitations (second eigenvalue greater than $ 50\mu$eV). This preselection is experimentally feasible at temperatures below $ \sim 100$mK using local burrow spectroscopy. In the uniform shell it excludes in particular any airfield in the trivial regime $ B < B_c$. We then compute the $ (\eta, \Omega_s)$ pair for the dear - zero modes in the ensemble and collect all these points (see Fig. \[fig: uniform\]c) to build up the probability density $ P(\eta,\Omega_s)$ that a establish zero mode with a measured $ \eta$ have a given $ \Omega_s$. The probability $ P(\eta,\Omega_s)$ is shown in Fig. \[fig: uniform\]d. Its visibility gives an accurate account of the quality of $ \eta$ as an estimator of $ \Omega_s$ within the whole space of uniform Lutchyn - Oreg nanowire models with near zero modes. A arrant correlation would appear as a straight, thin $ P(\eta,\Omega_s)$ along the diagonal. We attend that while the actual dependence of the distinctive $ \Omega | thf overlap corresponding uo spin-uncorrelatgd noise essenfially fullows the purely spatial ovxrlap, $\Ometa_s\sim\Omega_\mathrm{max}$. Aul forms lf the ocerlep reach a commoi minimum just neyonb vhe critical figld $B\gtrsim T_c$, and then *grmw* tlgether in an oscillatory fashion, ott of pnade with the spjittpnd (blhv gurve). Namely, MZMs deep in the tolologicel regime [enjoy *less protection*]{} than the lnes near $B_c$ (compare wwvefunctionw at $V_2$ and $B_3$).
In thd two uniform nanowirea simulated in panels (a,b) we see how che estimatir $\etw$ (black curvx) rougrly traces tmv overldp $\Omegs_s$, particularln aronnd rhe $B\gtrsim B_c$ region of minimal overlak. This goog eorrelation correspones to tfo pdrtizylaf cknhighratiojs kf the unirorm Lutchyb-Oreg model. To fullu wwsess the ovedall cjrwelation between $\eta$ and $\Omega_s$ for arbptrady configurations, we sinulate an ensemble of $\sim 2\cdot 10^4$ uniform nanowires with varying model parameters, inclndkng $B$, $\xopja$, $L$, $\mu$ and $\Delta$, distributed uniformly withig rtalpstic ranges (see gaption). Amongst alk fomsigurations, wg selecc tgose with a near zfro modg (splirting $E_0<10\mu$tV) wekl separated from higher exxitations (sebond eigenvalue greatex than $50\mu$eV). Chis pteselevtion is experimentally feaaible at telperaturea below $\sim 100$mK uskng lmcal tunnel spectroscopy. Ig the unihorm ease it dxclodes in particulag any field in the triviwl reyime $T<B_c$. We thej compute the $(\eta, \Omega_s)$ pair for the near-zerp kodvs in the ensekble and collqct all these koints (sez Fig. \[wig:uniform\]b) to builv the probabylity density $K(\eta,\Omega_s)$ thet a giveg zeeo mide witf a measured $\ets$ have a ynven $\Omegq_s$. The probability $P(\eja,\Kmega_s)$ is shown in Fig. \[fig:uniform\]c. Igs [rlfmle gyees an accurdte xccuint ow the qualiby uf $\eya$ as an estimator ox $\Omsga_s$ within the whplc space od unifori Lutchyn-Oreg nanowire models wlth nxar zeco modrs. W perfect correlation would aplear as a stvaight, thin $P(\qta,\Onega_s)$ along che diagonal. We see that while the actuao dependence of thw typical $\Omega | the overlap corresponding to spin-uncorrelated noise essentially purely overlap, $\Omega_s\sim\Omega_\mathrm{max}$. forms of the just the critical field B_c$, and then together in an oscillatory fashion, out phase with the splitting (blue curve). Namely, MZMs deep in the topological regime *less protection*]{} than the ones near $B_c$ (compare wavefunctions at $B_2$ and $B_3$). the uniform simulated panels (a,b) we see how the estimator $\eta$ (black curve) roughly traces the overlap $\Omega_s$, particularly the $B\gtrsim B_c$ region of minimal overlap. This correlation corresponds to two configurations of the uniform Lutchyn-Oreg To assess the correlation $\eta$ $\Omega_s$ for arbitrary we simulate an ensemble of $\sim 2\cdot 10^4$ uniform nanowires with varying model parameters, including $B$, $\alpha$, $\mu$ and uniformly within ranges caption). all configurations, we with a near zero mode (splitting from higher excitations (second eigenvalue greater than $50\mu$eV). preselection is feasible at temperatures below $\sim 100$mK local tunnel spectroscopy. In the uniform case it in particular any field in the trivial regime $B<B_c$. We then compute the $(\eta, \Omega_s)$ the near-zero modes in ensemble and collect these (see \[fig:uniform\]c) build the density $P(\eta,\Omega_s)$ that a given zero mode with a measured $\eta$ a given $\Omega_s$. The probability $P(\eta,\Omega_s)$ is shown in Fig. profile an accurate account the quality of $\eta$ an of $\Omega_s$ within the of Lutchyn-Oreg near modes. perfect correlation would appear a straight, thin $P(\eta,\Omega_s)$ along diagonal. We see that the typical $\Omega | the overlap corresponding to Spin-uncorrElateD noIse EsSentIallY follows the purELy spAtial overlap, $\Omega_s\sim\OMega_\mAtHRm{maX}$. alL formS of the oVErLAP reAcH a ComMoN MiNimum JusT beyond The criticaL fiElD $B\gtrsim B_c$, anD ThEn *grow* togeTheR in an oscillaTorY fashiOn, Out OF phasE wiTh the SplittINg (blue Curve). NameLy, mzMs deeP In the toPOLoGicaL regime [enjoy *less pROtECtion*]{} than the onEs near $b_c$ (COmPARe wAveFunctions aT $B_2$ And $B_3$).
IN The two uNIfORM NanOWires simulateD in panels (a,b) WE seE how thE eStiMAtor $\etA$ (blacK cURve) Roughly tracEs thE overlap $\OMega_s$, pARticulaRLy arounD the $B\gTrsIm B_C$ regIOn Of MinImAL ovERlAp. THIs gOod correLaTiOn corRespONDS To twO paRticUlar cOnfigurations Of tHe unIForM LutcHyn-OrEg moDeL. To fuLly assEss thE oVerall correlatiOn beTween $\eta$ aNd $\OMeGa_s$ FoR arbiTRary coNfiGurAtions, wE simulaTE an EnSEMBlE of $\sim 2\cdot 10^4$ uniform nAnOWIrEs with vaRying mODeL pARameters, InCluDing $b$, $\ALpha$, $L$, $\Mu$ anD $\deLta$, distrIbuted UNiFoRmly witHiN realiStIc rAngEs (see CAptiOn). AmonGst all coNfiguRAtions, we select THose with a near ZErO MOdE (SpliTtiNg $E_0<10\mu$eV) well SepaRAted From HIgHer EXcitaTions (SeCOnD Eigenvalue greater thAn $50\Mu$eV). ThIs preSelection is exPerimentalLY FEasible aT temPErATures below $\sim 100$mk usinG local tunnEL spectroScopy. in the uniForm case iT EXcludes iN paRtiCulAr aNY FiEld in the triviAL RegiMe $b<B_c$. We thEn cOmpute tHe $(\eTa, \OMegA_s)$ pAiR for the neAr-zero moDeS iN tHe EnsEmble ANd collecT aLl tHeSe pOints (SEe Fig. \[fIg:uniForm\]C) tO bUIld The probABiLITy deNsItY $P(\etA,\OmEgA_s)$ thaT a giVEn zEro mode With a measUreD $\Eta$ hAvE a Given $\OmEga_s$. The probabIlIty $P(\eta,\OmeGa_S)$ is Shown iN fIg. \[fig:uniForm\]d. Its profile gives an aCCurate aCcoUnt of The qUality of $\eTa$ aS an estImaTOr of $\OmEga_s$ wiThin tHe WhoLE Space OF UnIfoRm lutchyn-OreG NAnoWire mOdEls wIth near Zero modes. A perfect cORreLation would apPeaR as a STRaIghT, ThIN $P(\eTa,\oMegA_S)$ Along the diagonaL. We see that WhILe The actual dEPenDeNce of thE typicaL $\OmegA | the overlap corresponding to spin-u ncorr ela ted n oise ess entially follo w s th e purely spatial overl ap, $ \O m ega_ s \s im\Om ega_\ma t hr m { max }$ .All f o rm s ofthe overla p reach acom mo n minimum ju s tbeyond the cr itical field $B \gtrsi mB_c $ , and th en *g row* t o gether in an os ci l latory fashion , ou t of phase with the s p li t ting (blue cur ve). N am e ly , MZM s d eep in the t opolo g ical re g im e [ enj o y *less prote ction*]{} t h anthe on es ne a r $B_c $ (co mp a rewavefunctio ns a t $B_2$ a nd $B_ 3 $).
In the two unifo rmnan owir e ssi mul at e d i n p ane l s ( a,b) wese ehow t he e s t i m ator $\ eta$ (bla ck curve) rou ghl y tr a ces theoverl ap $ \O mega_ s$, pa rticu la rly around the$B\g trsim B_c $ r eg ion o f min i mal ov erl ap. This g ood cor r ela ti o n co rresponds to two p ar t i cu lar conf igurat i on so f the un if orm Lut c h yn-Or eg m o de l. To fu lly as s es sthe ove ra ll cor re lat ion betw e en $ \eta$and $\Om ega_s $ for arbitrary configuration s ,w e s i mula tean ensemble of$ \sim 2\c d ot 10 ^ 4$ un iform n a no w ires with varying m od el par amete rs, including $B$, $\al p h a $, $L$,$\mu $ a n d $\Delta$, di strib uted unifo r mly with in re alisticranges (s e e caption ).Amo ngs t a l l c onfigurations , we s el ect tho sewith anea r z ero mo de (splitti ng $E_0< 10 \m u$ eV ) w ell s e paratedfr omhi ghe r exc i tation s (se cond e ig e nva lue gre a te r than $ 50 \mu$ eV) .Thispres e lec tion is experime nta l ly f ea si ble attemperaturesbe low $\sim10 0$m K usin g local tu nnel spectroscopy. In t h e unifo rmcaseit e xcludes i n p articu lar any fi eld in thetr ivi a l regi m e $ B<B _c $. We then c omp ute t he $(\ eta, \O mega_s)$ pair fort henear-zero mod esin t h e e nse m bl e an dc oll e c t all these poi nts (see F ig . \ [fig:unifo r m\] c) to bui ld theproba b ility d ensity $P (\eta,\Om eg a_s) $ tha t a givenzero mod e with am easur e d$\eta $ h ave agi ven $\Om ega_s$ . Th e pro babili ty $P(\e ta,\O me ga_s)$ i s shown in Fig. \[fig:u niform \]d.Its profilegiv e s a n accurat e ac count of t hequa lityof$ \eta$ asa nest i mator of$ \Omega_s$ wi thi n th e whole spa c e ofunifo rmL utchyn -Ore g nanowire models with near zero mod e s . A pe r fect c orrelation wou ldap p e ar as ast raight, thi n $P(\et a, \ Omega _s)$ a long t he diag o n al . We se e th atwhile the ac tu a l depen de nc e of th e ty pi cal $\ Omega | the_overlap corresponding_to spin-uncorrelated noise essentially_follows the_purely_spatial overlap,_$\Omega_s\sim\Omega_\mathrm{max}$._All forms of_the overlap reach_a common minimum just_beyond the critical_field_$B\gtrsim B_c$, and then *grow* together in an oscillatory fashion, out of phase with_the_splitting (blue_curve)._Namely,_MZMs deep in the topological_regime [enjoy *less protection*]{} than_the ones_near $B_c$ (compare wavefunctions at $B_2$ and $B_3$).
In_the_two uniform nanowires_simulated in panels (a,b) we see how the estimator_$\eta$ (black curve) roughly traces the_overlap $\Omega_s$, particularly_around_the_$B\gtrsim B_c$ region of_minimal overlap. This good correlation corresponds_to two particular configurations of the_uniform Lutchyn-Oreg model. To fully assess the_overall correlation between $\eta$ and $\Omega_s$_for arbitrary configurations, we simulate_an ensemble_of $\sim 2\cdot 10^4$ uniform_nanowires with varying_model parameters,_including $B$, $\alpha$,_$L$, $\mu$ and $\Delta$, distributed uniformly_within realistic ranges_(see caption). Amongst all configurations, we_select_those with a_near_zero_mode (splitting_$E_0<10\mu$eV) well separated_from_higher excitations_(second_eigenvalue greater than $50\mu$eV). This preselection_is_experimentally feasible at temperatures below $\sim 100$mK_using local tunnel spectroscopy._In_the uniform case it_excludes in particular any field_in the trivial regime $B<B_c$. We_then compute_the $(\eta,_\Omega_s)$ pair for the near-zero modes in the ensemble and collect_all these points (see Fig. \[fig:uniform\]c)_to build the probability_density $P(\eta,\Omega_s)$_that_a given zero_mode_with a_measured $\eta$ have a given $\Omega_s$. The_probability $P(\eta,\Omega_s)$_is shown in Fig. \[fig:uniform\]d. Its_profile gives an accurate_account_of the quality of $\eta$ as_an estimator of $\Omega_s$ within the_whole space of uniform Lutchyn-Oreg_nanowire_models_with near zero modes. A_perfect correlation would appear as a_straight, thin $P(\eta,\Omega_s)$_along the diagonal. We see that while_the_actual dependence of the typical $\Omega |
{3\alpha}{\bar\alpha}\ln\alpha-3\ln\bar\alpha\biggr]
+ \frac{C_F}{N_c} \left(\ln\alpha+\frac{\bar\alpha}{\alpha}\ln\bar\alpha\right).\end{aligned}$$ Finally $$\begin{aligned}
\mathbb{X}^{(2)}_{\rm IB} f (z_1,z_2)&= C_F \int_0^1d\alpha\int_0^{\bar\alpha} d\beta \biggl[C_F\, \xi_{\rm IB}^F(\alpha,\beta) +
\frac1{N_c} \Big(\xi_{\rm IB}^A \,(\alpha,\beta)
+ \xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta)\mathbb{P}_{12}\Big)\biggr] f(z_{12}^\alpha,z_{21}^\beta)\,,\end{aligned}$$ where $$\begin{aligned}
\xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta) &= -2\biggl[ \operatorname{Li}_3(\bar\alpha) -\operatorname{Li}_3\left(1-\frac\alpha{\bar\beta}\right) +
\frac1{\bar\alpha}\left[\operatorname{Li}_2(\alpha) -\operatorname{Li}_2\left(\frac\alpha{\bar\beta}\right)\right] +\ln\bar\tau
\,\operatorname{Li}_2\left(1-\frac\alpha{\bar\beta}\right)
\notag\\
&\quad +\frac{1+\alpha}{2\bar\alpha}\ln\bar\alpha\ln\bar\beta
+ (\alpha\leftrightarrow \beta) \biggr]\,,
\notag\\[1mm]
\xi_{\rm IB}^{A}(\alpha,\beta)&=2\biggl[\operatorname{Li}_3(\bar\beta) -2\operatorname{Li}_3(\beta)-\operatorname{Li}_3\left(1-\frac{\beta}{\bar\alpha}\right)-\operatorname{Li}_3\left(\frac{\beta}{\bar\alpha}\right)
+\ln\bar \tau\,\operatorname{Li}_2\left(1-\frac{\beta}{\bar\alpha}\right)
+\ln \left(\frac\beta{\bar\alpha}\right) \operatorname | { 3\alpha}{\bar\alpha}\ln\alpha-3\ln\bar\alpha\biggr ]
+ \frac{C_F}{N_c } \left(\ln\alpha+\frac{\bar\alpha}{\alpha}\ln\bar\alpha\right).\end{aligned}$$ Finally $ $ \begin{aligned }
\mathbb{X}^{(2)}_{\rm IB } f (z_1,z_2)&= C_F \int_0 ^ 1d\alpha\int_0^{\bar\alpha } d\beta \biggl[C_F\, \xi_{\rm IB}^F(\alpha,\beta) +
\frac1{N_c } \Big(\xi_{\rm IB}^A \,(\alpha,\beta)
+ \xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta)\mathbb{P}_{12}\Big)\biggr ] f(z_{12}^\alpha, z_{21}^\beta)\,,\end{aligned}$$ where $ $ \begin{aligned }
\xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta) & = -2\biggl [ \operatorname{Li}_3(\bar\alpha) -\operatorname{Li}_3\left(1-\frac\alpha{\bar\beta}\right) +
\frac1{\bar\alpha}\left[\operatorname{Li}_2(\alpha) -\operatorname{Li}_2\left(\frac\alpha{\bar\beta}\right)\right ] + \ln\bar\tau
\,\operatorname{Li}_2\left(1-\frac\alpha{\bar\beta}\right)
\notag\\
& \quad + \frac{1+\alpha}{2\bar\alpha}\ln\bar\alpha\ln\bar\beta
+ (\alpha\leftrightarrow \beta) \biggr]\, ,
\notag\\[1 mm ]
\xi_{\rm IB}^{A}(\alpha,\beta)&=2\biggl[\operatorname{Li}_3(\bar\beta) -2\operatorname{Li}_3(\beta)-\operatorname{Li}_3\left(1-\frac{\beta}{\bar\alpha}\right)-\operatorname{Li}_3\left(\frac{\beta}{\bar\alpha}\right)
+ \ln\bar \tau\,\operatorname{Li}_2\left(1-\frac{\beta}{\bar\alpha}\right)
+ \ln \left(\frac\beta{\bar\alpha}\right) \operatorname | {3\alpja}{\bar\alpha}\ln\alpha-3\ln\bar\alkha\biggr]
+ \frac{C_F}{N_c} \left(\ln\elpha+\frzc{\bar\alpfa}{\alpha}\ln\bar\alpha\right).\end{alijned}$$ Finaoly $$\begin{aligned}
\mathbb{B}^{(2)}_{\rm IB} f (d_1,z_2)&= C_F \int_0^1e\alpia\int_0^{\bar\alpha} d\bxfa \biggl[C_F\, \xi_{\dl IB}^Y(\aopha,\beta) +
\frac1{N_g} \Big(\xi_{\rm IT}^A \,(\alpha,\beta)
+ \xh_{\ro NB}^{A\mathbb{P}}(\alpha,\beta)\mathbb{P}_{12}\Big)\biggr] f(s_{12}^\alpha,z_{21}^\neha)\,,\end{aligned}$$ wrere $$\fegih{aligned}
\xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta) &= -2\giggl[ \okeratorname{Li}_3(\bar\allha) -\operatorname{Li}_3\left(1-\frac\wlphw{\bar\beta}\right) +
\frac1{\har\alpha}\lefj[\kpewqtorname{Li}_2(\aloha) -\operatorname{Li}_2\left(\rrac\alpha{\bar\beta}\right)\right] +\ln\baf\tau
\,\o'eratorname{Oi}_2\oefh(1-\xrac\alpha{\bac\beta}\rpght)
\notag\\
&\quad +\frac{1+\alpva}{2\bar\allha}\ln\bar\alpha\lk\bar\bxta
+ (\qlpha\leftrightarrow \bxta) \biggr]\,,
\notag\\[1mm]
\xi_{\rm IB}^{A}(\alpha,\teca)&=2\biggl[\operatorname{Li}_3(\vae\beta) -2\opesatofbamd{Li}_3(\geva)-\oleratognaje{Li}_3\left(1-\frzc{\beta}{\bar\alpha}\right)-\operatornamt{Li}_3\jvgt(\frac{\beta}{\bad\alpha}\widht)
+\ln\bar \tau\,\operatorname{Li}_2\left(1-\frac{\beta}{\tar\zlpha}\right)
+\ln \left(\frac\bera{\bar\alpha}\right) \operajorname | {3\alpha}{\bar\alpha}\ln\alpha-3\ln\bar\alpha\biggr] + \frac{C_F}{N_c} \left(\ln\alpha+\frac{\bar\alpha}{\alpha}\ln\bar\alpha\right).\end{aligned}$$ Finally $$\begin{aligned} \mathbb{X}^{(2)}_{\rm (z_1,z_2)&= \int_0^1d\alpha\int_0^{\bar\alpha} d\beta \xi_{\rm IB}^F(\alpha,\beta) + \xi_{\rm f(z_{12}^\alpha,z_{21}^\beta)\,,\end{aligned}$$ where $$\begin{aligned} IB}^{A\mathbb{P}}(\alpha,\beta) &= -2\biggl[ -\operatorname{Li}_3\left(1-\frac\alpha{\bar\beta}\right) + \frac1{\bar\alpha}\left[\operatorname{Li}_2(\alpha) -\operatorname{Li}_2\left(\frac\alpha{\bar\beta}\right)\right] +\ln\bar\tau \,\operatorname{Li}_2\left(1-\frac\alpha{\bar\beta}\right) &\quad +\frac{1+\alpha}{2\bar\alpha}\ln\bar\alpha\ln\bar\beta + (\alpha\leftrightarrow \beta) \biggr]\,, \notag\\[1mm] \xi_{\rm IB}^{A}(\alpha,\beta)&=2\biggl[\operatorname{Li}_3(\bar\beta) -2\operatorname{Li}_3(\beta)-\operatorname{Li}_3\left(1-\frac{\beta}{\bar\alpha}\right)-\operatorname{Li}_3\left(\frac{\beta}{\bar\alpha}\right) +\ln\bar \tau\,\operatorname{Li}_2\left(1-\frac{\beta}{\bar\alpha}\right) +\ln \operatorname | {3\alpha}{\bar\alpha}\ln\alpha-3\ln\bar\Alpha\biggr]
+ \Frac{C_f}{N_c} \LefT(\lN\alpHa+\frAc{\bar\alpha}{\alphA}\Ln\baR\alpha\right).\end{aligned}$$ FInallY $$\bEGin{aLIgNed}
\maThbb{X}^{(2)}_{\rm ib} f (Z_1,Z_2)&= c_F \iNt_0^1D\aLphA\iNT_0^{\bAr\alpHa} d\Beta \bigGl[C_F\, \xi_{\rm IB}^f(\alPhA,\beta) +
\frac1{N_c} \BIG(\xI_{\rm IB}^A \,(\alphA,\beTa)
+ \xi_{\rm IB}^{A\matHbb{p}}(\alpha,\BeTa)\mAThbb{P}_{12}\big)\Biggr] F(z_{12}^\alphA,Z_{21}^\beta)\,,\eNd{aligned}$$ WhERe $$\begiN{Aligned}
\XI_{\Rm iB}^{A\mAthbb{P}}(\alpha,\beta) &= -2\biGGl[ \OPeratorname{Li}_3(\bAr\alphA) -\oPErATOrnAme{li}_3\left(1-\frac\AlPha{\baR\Beta}\rigHT) +
\fRAC1{\Bar\ALpha}\left[\operaTorname{Li}_2(\alPHa) -\oPeratoRnAme{lI}_2\left(\fRac\alPhA{\Bar\Beta}\right)\riGht] +\lN\bar\tau
\,\opEratorNAme{Li}_2\leFT(1-\frac\alPha{\bar\BetA}\riGht)
\nOTaG\\
&\qUad +\FrAC{1+\alPHa}{2\Bar\ALphA}\ln\bar\alPhA\lN\bar\bEta
+ (\aLPHA\LeftRigHtarRow \beTa) \biggr]\,,
\notag\\[1mM]
\xi_{\Rm IB}^{a}(\AlpHa,\betA)&=2\biggL[\opeRaTornaMe{Li}_3(\baR\beta) -2\OpEratorname{Li}_3(\betA)-\opeRatorname{li}_3\lEfT(1-\frAc{\Beta}{\bAR\alpha}\RigHt)-\oPeratorName{Li}_3\lEFt(\fRaC{\BETa}{\Bar\alpha}\right)
+\ln\bar \TaU\,\OPeRatornamE{Li}_2\lefT(1-\FrAc{\BEta}{\bar\alPhA}\riGht)
+\lN \LEft(\frAc\beTA{\bAr\alpha}\rIght) \opERaToRname | {3\alpha}{\bar\alpha}\ln\a lpha-3\ln\ bar\a lph a\b ig gr]+ \f rac{C_F}{N_c}\ left (\ln\alpha+\frac{\bar\ alpha }{ \ alph a }\ ln\ba r\alpha \ ri g h t). \e nd {al ig n ed }$$ F ina lly $$\ begin{alig ned }\mathbb{X}^{ ( 2) }_{\rm IB} f(z_1,z_2)&=C_F \int_ 0^ 1d\ a lpha\ int _0^{\ bar\al p ha} d\ beta \big gl [ C_F\,\ xi_{\rm I B} ^F(\ alpha,\beta) +
\f r ac 1 {N_c} \Big(\xi _{\rmIB } ^A \ ,(\ alp ha,\beta)+\xi_{ \ rm IB}^ { A\ m a t hbb { P}}(\alpha,\b eta)\mathbb { P}_ {12}\B ig )\b i ggr] f (z_{1 2} ^ \al pha,z_{21}^ \bet a)\,,\end {align e d}$$ wh e re $$\b egin{a lig ned }
\x i _{ \r m I B} ^ {A\ m at hbb { P}} (\alpha, \b et a) &= -2\ b i g g l[ \ ope rato rname {Li}_3(\bar\a lph a) - \ ope rator name{ Li}_ 3\ left( 1-\fra c\alp ha {\bar\beta}\rig ht)+
\frac1{ \ba r\ alp ha }\lef t [\oper ato rna me{Li}_ 2(\alph a ) - \o p e r at orname{Li}_2\left( \f r a c\ alpha{\b ar\bet a }\ ri g ht)\righ t] +\ ln\b a r \tau\,\o p er atorname {Li}_2 \ le ft (1-\fra c\ alpha{ \b ar\ bet a}\ri g ht)\notag \\
&\qua d +\f r ac{1+\alpha}{2 \ bar\alpha}\ln \ ba r \ al p ha\l n\b ar\beta
+ ( \alp h a\le ftri g ht arr o w \be ta) \ bi g gr ] \,,
\notag\\[1mm]
\ xi _{\rmIB}^{ A}(\alpha,\be ta)&=2\big g l [ \operato rnam e {L i }_3(\bar\beta) -2\o peratornam e {Li}_3(\ beta) -\operat orname{Li } _ 3\left(1 -\f rac {\b eta } { \b ar\alpha}\rig h t )-\o pe ratorna me{ Li}_3\l eft (\f rac {\b et a}{\bar\a lpha}\ri gh t)
+\ ln\ bar \ t au\,\ope ra tor na me{ Li}_2 \ left(1 -\fra c{\b et a} { \ba r\alpha } \r i g ht)+\ ln \le ft( \f rac\b eta{ \ bar \alpha} \right) \ ope r ator na me | {3\alpha}{\bar\alpha}\ln\alpha-3\ln\bar\alpha\biggr]
+ \frac{C_F}{N_c}_\left(\ln\alpha+\frac{\bar\alpha}{\alpha}\ln\bar\alpha\right).\end{aligned}$$ Finally_$$\begin{aligned}
\mathbb{X}^{(2)}_{\rm IB} f (z_1,z_2)&=_C_F \int_0^1d\alpha\int_0^{\bar\alpha}_d\beta_\biggl[C_F\, \xi_{\rm_IB}^F(\alpha,\beta)_+
\frac1{N_c} \Big(\xi_{\rm IB}^A_\,(\alpha,\beta)
+ \xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta)\mathbb{P}_{12}\Big)\biggr]_f(z_{12}^\alpha,z_{21}^\beta)\,,\end{aligned}$$ where $$\begin{aligned}
\xi_{\rm IB}^{A\mathbb{P}}(\alpha,\beta)_&= -2\biggl[ \operatorname{Li}_3(\bar\alpha)_-\operatorname{Li}_3\left(1-\frac\alpha{\bar\beta}\right)_+
\frac1{\bar\alpha}\left[\operatorname{Li}_2(\alpha) -\operatorname{Li}_2\left(\frac\alpha{\bar\beta}\right)\right] +\ln\bar\tau
\,\operatorname{Li}_2\left(1-\frac\alpha{\bar\beta}\right)
\notag\\
&\quad +\frac{1+\alpha}{2\bar\alpha}\ln\bar\alpha\ln\bar\beta
+ (\alpha\leftrightarrow \beta) \biggr]\,,
\notag\\[1mm]
\xi_{\rm IB}^{A}(\alpha,\beta)&=2\biggl[\operatorname{Li}_3(\bar\beta) -2\operatorname{Li}_3(\beta)-\operatorname{Li}_3\left(1-\frac{\beta}{\bar\alpha}\right)-\operatorname{Li}_3\left(\frac{\beta}{\bar\alpha}\right)
+\ln\bar \tau\,\operatorname{Li}_2\left(1-\frac{\beta}{\bar\alpha}\right)
+\ln \left(\frac\beta{\bar\alpha}\right) \operatorname |
+ \sum_{k=1}^{K_P} \bar{\lambda}_k^\star\left( \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \otimes\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\right)\mathbf{w}_f^\star \right. \nonumber\\
&\left. + (\sigma_I^2+\sigma_n^2) \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T\otimes\mathbf{I}_{N_r}\right)\mathbf{w}_f^\star\right) = \bar{\mathbf{0}}_{N_tN_r}.\end{aligned}$$
By applying the properties of Kronecker product [@meyer2000matrix] and after some manipulations, we can obtain the optimal $\mathbf{w}_f^\star$ as $$\begin{aligned}
\label{E:optimalwf2}
& \mathbf{w}_f^\star \!=\! \!-\!e^{j\phi} \big( \big(\sum_{k=1}^{K_P} \bar{\lambda}_k^\star(\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \!+\! \mathbf{I}_{N_t}\big)^{-\!1} \big(\sum_{k=1}^{K_P} \bar{\lambda}_k^\star\bar{\mathbf{h}}_{Mk} \mathbf{h}_{Pk}^H\big)^T \nonumber\\
& \otimes \!\big(\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\!+\! (\sigma_I^2 \!+\! \sigma_n^2)\mathbf{I}_{N_r}\big)^{-\!1}\bar{\mathbf{H}}_{MP}^H\big) \mathtt{vec}(\mathbf{I}_{N_tN_r}).\end{aligned}$$
Then we can recover the optimal $\mathbf{W}_f^\star$ from $\mathbf{w}_f^\star$ as $$\begin{aligned}
\label{E:Wf}
&{\mathbf{W}_f^\star = } -e^{j\phi}\Big(\sum_{k=1}^{ | + \sum_{k=1}^{K_P } \bar{\lambda}_k^\star\left (\left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \otimes\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\right)\mathbf{w}_f^\star \right. \nonumber\\
& \left. + (\sigma_I^2+\sigma_n^2) \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T\otimes\mathbf{I}_{N_r}\right)\mathbf{w}_f^\star\right) = \bar{\mathbf{0}}_{N_tN_r}.\end{aligned}$$
By applying the properties of Kronecker product [ @meyer2000matrix ] and after some manipulations, we can receive the optimum $ \mathbf{w}_f^\star$ as $ $ \begin{aligned }
\label{E: optimalwf2 }
& \mathbf{w}_f^\star \!=\! \!-\!e^{j\phi } \big (\big(\sum_{k=1}^{K_P } \bar{\lambda}_k^\star(\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \!+\! \mathbf{I}_{N_t}\big)^{-\!1 } \big(\sum_{k=1}^{K_P } \bar{\lambda}_k^\star\bar{\mathbf{h}}_{Mk } \mathbf{h}_{Pk}^H\big)^T \nonumber\\
& \otimes \!\big(\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\!+\! (\sigma_I^2 \!+\! \sigma_n^2)\mathbf{I}_{N_r}\big)^{-\!1}\bar{\mathbf{H}}_{MP}^H\big) \mathtt{vec}(\mathbf{I}_{N_tN_r}).\end{aligned}$$
Then we can recover the optimal $ \mathbf{W}_f^\star$ from $ \mathbf{w}_f^\star$ as $ $ \begin{aligned }
\label{E: Wf }
& { \mathbf{W}_f^\star = } -e^{j\phi}\Big(\sum_{k=1}^ { | + \dum_{k=1}^{K_P} \bar{\lambda}_k^\star\lefu( \left((\mathbf{h}_{Pk}\majhvf{h}_{Pk}^H)^V \otimea\bar{\mathcf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\right)\mathbh{w}_f^\srar \rught. \nonumber\\
&\left. + (\sigma_I^2+\sihma_n^2) \lefr((\matibf{h}_{Pk}\mathbf{h}_{Pk}^H)^V\ktimes\mathbf{I}_{H_v}\righc)\methbf{w}_f^\star\righj) = \bar{\mathbf{0}}_{T_tN_r}.\end{aligned}$$
Ty a'plying the properties of Kronecker [roduct [@kejer2000matrix] and wfteg fome manipulations, we can obtain the oltimal $\kathbf{w}_f^\star$ ss $$\begin{aligned}
\label{E:optilalwv2}
& \mathbf{w}_f^\star \!=\! \!-\!e^{j\phi} \big( \vig(\stn_{k=1}^{K_P} \bar{\lambaa}_k^\star(\mathbf{h}_{Pk}\mathbf{g}_{Pk}^H)^T \!+\! \mathbf{I}_{N_t}\big)^{-\!1} \big(\sum_{k=1}^{K_P} \bxr{\lamyda}_k^\star\bar{\narhbv{v}}_{Mk} \mathbf{h}_{'k}^H\big)^N \nonumber\\
& \otimes \!\tig(\bar{\msthbf{H}}_{MP}^H\bar{\mabhbf{H}}_{KP}\!+\! (\wigma_I^2 \!+\! \sigma_n^2)\mathbf{I}_{I_r}\big)^{-\!1}\bar{\mathbf{H}}_{MP}^H\bid) \mathtt{eee}(\mathbf{I}_{N_tN_r}).\end{alignee}$$
Tyen wg can recucer tht o'tijal $\mahhbh{W}_f^\star$ froj $\mathbf{w}_f^\srar$ as $$\begin{aligned}
\lwvel{E:Wf}
&{\mathbr{W}_f^\staw = } -e^{j\phi}\Big(\sum_{k=1}^{ | + \sum_{k=1}^{K_P} \bar{\lambda}_k^\star\left( \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \otimes\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\right)\mathbf{w}_f^\star \right. \nonumber\\ (\sigma_I^2+\sigma_n^2) = \bar{\mathbf{0}}_{N_tN_r}.\end{aligned}$$ applying the properties after manipulations, we can the optimal $\mathbf{w}_f^\star$ $$\begin{aligned} \label{E:optimalwf2} & \mathbf{w}_f^\star \!=\! \!-\!e^{j\phi} \big(\sum_{k=1}^{K_P} \bar{\lambda}_k^\star(\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \!+\! \mathbf{I}_{N_t}\big)^{-\!1} \big(\sum_{k=1}^{K_P} \bar{\lambda}_k^\star\bar{\mathbf{h}}_{Mk} \mathbf{h}_{Pk}^H\big)^T \nonumber\\ & \otimes \!\big(\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\!+\! (\sigma_I^2 \!+\! \mathtt{vec}(\mathbf{I}_{N_tN_r}).\end{aligned}$$ Then we can recover the optimal $\mathbf{W}_f^\star$ from $\mathbf{w}_f^\star$ as $$\begin{aligned} \label{E:Wf} = -e^{j\phi}\Big(\sum_{k=1}^{ | + \sum_{k=1}^{K_P} \bar{\lambda}_k^\star\left( \lEft((\mathbf{h}_{pk}\matHbf{H}_{Pk}^h)^T \OtimEs\baR{\mathbf{H}}_{MP}^H\bar{\MAthbF{H}}_{MP}\right)\mathbf{w}_f^\star \rIght. \nOnUMber\\
&\LEfT. + (\sigmA_I^2+\sigma_N^2) \LeFT((\MatHbF{h}_{pk}\mAtHBf{H}_{Pk}^H)^T\OtiMes\mathBf{I}_{N_r}\right)\MatHbF{w}_f^\star\right) = \BAr{\Mathbf{0}}_{N_tN_r}.\End{Aligned}$$
By appLyiNg the pRoPerTIes of kroNeckeR produCT [@meyer2000Matrix] and AfTEr some MAnipulaTIOnS, we cAn obtain the optimaL $\MaTHbf{w}_f^\star$ as $$\begIn{aligNeD}
\LaBEL{E:oPtiMalwf2}
& \mathbF{w}_F^\star \!=\! \!-\!E^{J\phi} \big( \BIg(\SUM_{K=1}^{K_P} \BAr{\lambda}_k^\star(\Mathbf{h}_{Pk}\maTHbf{H}_{Pk}^H)^T \!+\! \mAtHbf{i}_{n_t}\big)^{-\!1} \bIg(\sum_{K=1}^{K_p} \Bar{\Lambda}_k^\star\Bar{\mAthbf{h}}_{Mk} \mAthbf{h}_{pK}^H\big)^T \nONumber\\
& \oTimes \!\bIg(\bAr{\mAthbF{h}}_{Mp}^H\Bar{\MaTHbf{h}}_{mP}\!+\! (\SigMA_I^2 \!+\! \sIgma_n^2)\matHbF{I}_{n_r}\big)^{-\!1}\Bar{\mATHBF{H}}_{MP}^h\biG) \matHtt{veC}(\mathbf{I}_{N_tN_r}).\eNd{aLignED}$$
ThEn we cAn recOver ThE optiMal $\matHbf{W}_f^\StAr$ from $\mathbf{w}_f^\sTar$ aS $$\begin{aliGneD}
\lAbeL{E:wf}
&{\matHBf{W}_f^\stAr = } -e^{J\phI}\Big(\sum_{K=1}^{ | + \sum_{k=1}^{K_P} \bar{ \lambda}_k ^\sta r\l eft (\lef t((\ mathbf{h}_{Pk} \ math bf{h}_{Pk}^H)^T \otime s\bar {\ m athb f {H }}_{M P}^H\ba r {\ m a thb f{ H} }_{ MP } \r ight) \ma thbf{w} _f^\star \ rig ht . \nonumbe r \\
&\left. +(\sigma_I^2+ \si gma_n^ 2) \l e ft((\ mat hbf{h }_{Pk} \ mathbf {h}_{Pk}^ H) ^ T\otim e s\mathb f { I} _{N_ r}\right)\mathbf{ w }_ f ^\star\right)= \bar {\ m at h b f{0 }}_ {N_tN_r}.\ en d{ali g ned}$$By a p ply i ng the proper ties of Kro n eck er pro du ct[ @meyer 2000m at r ix] and aftersome manipula tions, we cano btain t he opt ima l $ \mat h bf {w }_f ^\ s tar $ a s $ $ \be gin{alig ne d}
\la bel{ E : o p tima lwf 2}
& \m athbf{w}_f^\s tar \! = \!\!-\! e^{j\ phi} \ big(\big(\ sum_{ k= 1}^{K_P} \bar{\ lamb da}_k^\st ar( \m ath bf {h}_{ P k}\mat hbf {h} _{Pk}^H )^T \!+ \ ! \ ma t h b f{ I}_{N_t}\big)^{-\! 1} \ bi g(\sum_{ k=1}^{ K _P }\ bar{\lam bd a}_ k^\s t a r\bar {\ma t hb f{h}}_{M k} \ma t hb f{ h}_{Pk} ^H \big)^ T \n onu mber\ \
& \otim es \!\bi g(\ba r {\mathbf{H}}_{ M P}^H\bar{\mat h bf { H }} _ {MP} \!+ \! (\sigma_ I^2\ !+\! \si g ma _n^ 2 )\mat hbf{I }_ { N_ r }\big)^{-\!1}\bar{\ ma thbf{H }}_{M P}^H\big) \m athtt{vec} ( \ m athbf{I} _{N_ t N_ r }).\end{aligne d}$$
Then we c a n recove r the optimal $\mathbf { W }_f^\sta r$fro m $ \ma t h bf {w}_f^\star$a s $$\ be gin{ali gne d}
\la bel {E: Wf}
&{ \mathbf{W }_f^\sta r=}-e ^{j \phi} \ Big(\sum _{ k=1 }^ { | _+ \sum_{k=1}^{K_P}_\bar{\lambda}_k^\star\left( \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T \otimes\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\right)\mathbf{w}_f^\star \right._ _\nonumber\\
_ &\left._+_(\sigma_I^2+\sigma_n^2) \left((\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T\otimes\mathbf{I}_{N_r}\right)\mathbf{w}_f^\star\right) =_\bar{\mathbf{0}}_{N_tN_r}.\end{aligned}$$
By applying the_properties of Kronecker product [@meyer2000matrix]_and after some_manipulations,_we can obtain the optimal $\mathbf{w}_f^\star$ as $$\begin{aligned}
\label{E:optimalwf2}
& \mathbf{w}_f^\star \!=\!_\!-\!e^{j\phi}_\big( \big(\sum_{k=1}^{K_P}_\bar{\lambda}_k^\star(\mathbf{h}_{Pk}\mathbf{h}_{Pk}^H)^T_\!+\!_\mathbf{I}_{N_t}\big)^{-\!1} \big(\sum_{k=1}^{K_P} \bar{\lambda}_k^\star\bar{\mathbf{h}}_{Mk} \mathbf{h}_{Pk}^H\big)^T _\nonumber\\
& \otimes \!\big(\bar{\mathbf{H}}_{MP}^H\bar{\mathbf{H}}_{MP}\!+\!_(\sigma_I^2 \!+\!_\sigma_n^2)\mathbf{I}_{N_r}\big)^{-\!1}\bar{\mathbf{H}}_{MP}^H\big) \mathtt{vec}(\mathbf{I}_{N_tN_r}).\end{aligned}$$
Then we can recover the optimal_$\mathbf{W}_f^\star$_from $\mathbf{w}_f^\star$ as_$$\begin{aligned}
\label{E:Wf}
&{\mathbf{W}_f^\star = } -e^{j\phi}\Big(\sum_{k=1}^{ |
type superconductor, found in Li$_2$Pt$_3$B, provides an alternative way to study unconventional SC, especially that arising from phonon pairing mechanism. Indeed, the absence of parity symmetry coupled with strong spin-orbit coupling, which results in an admixture of spin-singlet and spin-triplet pairing, requires a complete reconceptualization of Cooper pairs and the nature of the superconducting state.
We thank P. Frigeri, R. Kaur, I. Milat and H. Takeya for useful discussions. This work is supported by the National Science Foundation under awards No. NSF-EIA0121568 and NSF-DMR0318665, the Department of Energy under award No. DEFG02-91ER45439, the Swiss National Science Foundation and Petroleum Research Funds. We also acknowledge supports from ICAM (HQY), the 2003-PFRA program of Japan Society for the Promotion of Science (NH) and the CEEX program of Romanian Ministry of Education and Research (PB).
[99]{} See, e.g., T. Moriya and K. Ueda, Rep. Prog. Phys., **66**, 1299 (2003); A. P. Mackenzie and Y. Maeno, Rev. Mod. Phys. **75**, 657 (2003).
L. P. Gor’kov and E. I. Rashba, Phys. Rev. Lett. **87**, 037004 (2001).
P. A. Frigeri *et al*., Phys. Rev. Lett. **92**, 097001 (2004).
P. A. Frigeri *et al*., cond-mat/0505108 (2005).
V.M. Edel’stein, Sov. Phys. JETP **68**, 1244 (1989).
L. S. Levitov, Yu. V. Nazarov and G. M. Eliashberg, JEPT Lett. **41**, 445 (1985).
K. V. Samokhin, E. S. Zijlstra and S. K. Bose, Phys. Rev. B **69**, 094514 (2004) \[Erratum: Phys. Rev. B **70**, 069902 (E) (2004)\].
K. Togano *et al*., Phys. Rev. Lett. **93**, 247004 (2004).
P. Badica | type superconductor, found in Li$_2$Pt$_3$B, provides an alternative direction to analyze unconventional SC, especially that arising from phonon coupling mechanism. Indeed, the absence of parity symmetry coupled with hard spin - orbit yoke, which result in an admixture of tailspin - singlet and spin - triplet pairing, requires a accomplished reconceptualization of Cooper pairs and the nature of the superconducting state.
We thank P. Frigeri, R. Kaur, I. Milat and H. Takeya for useful discussions. This work is support by the National Science Foundation under awards No. NSF - EIA0121568 and NSF - DMR0318665, the Department of Energy under award No. DEFG02 - 91ER45439, the Swiss National Science Foundation and Petroleum Research Funds. We also notice supports from ICAM (HQY), the 2003 - PFRA program of Japan Society for the Promotion of Science (NH) and the CEEX program of Romanian Ministry of Education and Research (PB).
[ 99 ] { } See, for example, T. Moriya and K. Ueda, Rep. Prog. Phys. , * * 66 * *, 1299 (2003); A. P. Mackenzie and Y. Maeno, Rev. Mod. Phys. * * 75 * *, 657 (2003).
L. P. Gor’kov and E. I. Rashba, Phys. Rev. Lett. * * 87 * *, 037004 (2001).
P. A. Frigeri * et al *. , Phys. Rev. Lett. * * 92 * *, 097001 (2004).
P. A. Frigeri * et al *. , cond - mat/0505108 (2005).
V.M. Edel’stein, Sov. Phys. JETP * * 68 * *, 1244 (1989).
L. S. Levitov, Yu. V. Nazarov and G. M. Eliashberg, JEPT Lett. * * 41 * *, 445 (1985).
K. V. Samokhin, E. S. Zijlstra and S. K. Bose, Phys. Rev. B * * 69 * *, 094514 (2004) \[Erratum: Phys. Rev. B * * 70 * *, 069902 (E) (2004)\ ].
K. Togano * et al *. , Phys. Rev. Lett. * * 93 * *, 247004 (2004).
P. Badica | tyoe superconductor, found ln Li$_2$Pt$_3$B, provides an alvernatibe way tu study unconventional SC, es'eciqlly ukat arising from phovon pairijg mechabism. Undeed, the absence of padlty svmnetry coupled eith stronc spin-orbit cogpuiug, which results in an admixture of fpin-sinblft and spin-triklet kaiwing, gewuires a complete reconceptuamizatioi of Cooper paits and the nature of the skperfonducting state.
We thank P. Frugery, R. Kaur, I. Miuat and H. Uaneya for uaeful discussions. This work is ruppoxted by the Nqtiltal Science Founqation under awards To. NSF-EOA0121568 and NSF-DMR0318665, bhe Dxparrment of Energy under award No. DEFG02-91ER45439, tre Swiss Tacional Science Foundariin ang Pedroldym Festarrh Runds. Ae elso acknowmedge suppoets from ICAM (HQY), tne 2003-IGRA program kf Japwn Society for the Promotion of Science (NV) ahd the CEEX program of Eomanian Ministry of Gducation wnd Research (PB).
[99]{} See, e.g., T. Moriya and K. Ueda, Rep. Prmg. Phbs., **66**, 1299 (2003); A. O. Mwckenzie and Y. Maeno, Rev. Mod. Phys. **75**, 657 (2003).
L. P. Gor’kod amd E. I. Rashba, Phns. Rev. Lett. **87**, 037004 (2001).
P. A. Grlgrti *et al*., Phys. Tev. Letc. **92**, 097001 (2004).
P. A. Frigeri *et al*., cond-maj/0505108 (2005).
V.M. Eeel’stein, Fov. Lhys. JETP **68**, 1244 (1989).
L. S. Levitov, Yu. C. Nazarov anb G. M. Eliashberg, JEPT Lett. **41**, 445 (1985).
K. V. Sampkhin, E. S. Zijlstra and S. K. Buse, Lhys. Rev. B **69**, 094514 (2004) \[Erratuj: Phys. Rev. B **70**, 069902 (E) (2004)\].
Y. Tpgdno *et al*., Phys. Rev. Lett. **93**, 247004 (2004).
[. Badica | type superconductor, found in Li$_2$Pt$_3$B, provides an to unconventional SC, that arising from absence parity symmetry coupled strong spin-orbit coupling, results in an admixture of spin-singlet spin-triplet pairing, requires a complete reconceptualization of Cooper pairs and the nature of superconducting state. We thank P. Frigeri, R. Kaur, I. Milat and H. Takeya useful This is by the National Science Foundation under awards No. NSF-EIA0121568 and NSF-DMR0318665, the Department of Energy under No. DEFG02-91ER45439, the Swiss National Science Foundation and Research Funds. We also supports from ICAM (HQY), the program Japan Society the of (NH) and the program of Romanian Ministry of Education and Research (PB). [99]{} See, e.g., T. Moriya and K. Ueda, Prog. Phys., (2003); A. Mackenzie Y. Rev. Mod. Phys. (2003). L. P. Gor’kov and E. Rev. Lett. **87**, 037004 (2001). P. A. Frigeri al*., Phys. Lett. **92**, 097001 (2004). P. A. *et al*., cond-mat/0505108 (2005). V.M. Edel’stein, Sov. Phys. **68**, 1244 (1989). L. S. Levitov, Yu. V. Nazarov and G. M. Eliashberg, JEPT Lett. (1985). K. V. Samokhin, S. Zijlstra and K. Phys. B 094514 (2004) Phys. Rev. B **70**, 069902 (E) (2004)\]. K. Togano *et al*., Rev. Lett. **93**, 247004 (2004). P. Badica | type superconductor, found in li$_2$Pt$_3$B, proviDes an AltErnAtIve wAy to Study unconventIOnal sC, especially that arisinG from PhONon pAIrIng meChanism. iNdEED, thE aBsEncE oF PaRity sYmmEtry couPled with stRonG sPin-orbit coupLInG, which resuLts In an admixturE of Spin-siNgLet ANd spiN-trIplet PairinG, RequirEs a compleTe REconcePTualizaTIOn Of CoOper pairs and the naTUrE Of the supercondUcting StATe.
wE ThaNk P. frigeri, R. KaUr, i. MilaT And H. TakEYa FOR UseFUl discussions. this work is sUPpoRted by ThE NaTIonal SCiencE FOUndAtion under aWardS No. NSF-EIA0121568 And NSF-dmR0318665, the DePArtment Of EnerGy uNdeR awaRD NO. DeFG02-91eR45439, THe SWIsS NaTIonAl SciencE FOuNdatiOn anD pETRoleUm REseaRch FuNds. We also acknOwlEdge SUppOrts fRom ICaM (HQy), tHe 2003-PFRa progrAm of JApAn Society for the promOtion of ScIenCe (nH) aNd The CEex progrAm oF RoManian MInistry OF EdUcATIOn And Research (PB).
[99]{} See, e.g., t. MORIyA and K. UedA, Rep. PrOG. PHyS., **66**, 1299 (2003); a. P. MackenZiE anD Y. MaENO, Rev. MOd. PhYS. **75**, 657 (2003).
L. p. Gor’kov aNd E. I. RaSHbA, PHys. Rev. LEtT. **87**, 037004 (2001).
P. A. FriGeRi *eT al*., phys. REV. LetT. **92**, 097001 (2004).
P. A. FriGeri *et al*., Cond-mAT/0505108 (2005).
V.M. Edel’stein, SoV. phys. JETP **68**, 1244 (1989).
L. S. LevIToV, yU. V. nAzarOv aNd G. M. EliashbErg, JepT LeTt. **41**, 445 (1985).
K. V. sAmOkhIN, E. S. ZiJlstrA aND S. k. bose, Phys. Rev. B **69**, 094514 (2004) \[Erratum: phYs. Rev. B **70**, 069902 (e) (2004)\].
K. TogAno *et al*., Phys. ReV. Lett. **93**, 247004 (2004).
P. BadiCA | type superconductor, foun d in Li$_2 $Pt$_ 3$B , p ro vide s an alternative w a y to study unconventionalSC, e sp e cial l ythatarising fr o m ph on on pa ir i ng mech ani sm. Ind eed, the a bse nc e of paritys ym metry coup led with strong sp in-orb it co u pling , w hichresult s in an admixtur eo f spin - singlet a nd spi n-triplet pairing , r e quires a compl ete re co n ce p t ual iza tion of Co op er pa i rs andt he n a tur e of the super conductings tat e.
We t han k P. Fr igeri ,R . K aur, I. Mil at a nd H. Tak eya fo r useful discuss ions.Thi s w orki ssu ppo rt e d b y t heN ati onal Sci en ce Foun dati o n u nder aw ards No.NSF-EIA012156 8 a nd N S F-D MR031 8665, the D epart ment o f Ene rg y under award N o. D EFG02-91E R45 43 9,th e Swi s s Nati ona l S cienceFoundat i onan d P et roleum Research Fu nd s . W e also a cknowl e dg es upportsfr omICAM ( HQY), the 20 03-PFRAprogra m o fJapan S oc iety f or th e P romot i on o f Scie nce (NH) andt he CEEX progra m of RomanianM in i s tr y ofEdu cation andRese a rch(PB) .
[99 ] {} Se e, e. g. , T . Moriya and K. Ueda ,Rep. P rog.Phys., **66** , 1299 (20 0 3 ) ; A. P.Mack e nz i e and Y. Maeno , Rev . Mod. Phy s . **75** , 657 (2003).
L. P. G o r ’kov and E. I. Ra shb a , P hys. Rev. Let t . **8 7* *, 0370 04(2001).
P . A . F rig er i *et al* ., Phys. R ev .Le tt. **92 * *, 09700 1(20 04 ).
P. A . Frige ri *e t al *. ,c ond -mat/05 0 51 0 8 (20 05 ).
V. M.Ed el’st ein, Sov . Phys. JETP **6 8** , 124 4(1 989).
L. S. Levitov ,Yu. V. Naz ar ovand G. M . Eliash berg, JEPT Lett. **41** , 445 (1 985 ).
K . V. Samokhin , E . S. Z ijl s tra an d S. K . Bos e, Ph y s . Rev . B**6 9* *, 094514( 2 004 ) \[E rr atum : Phys. Rev. B **70**, 06 9 902 (E) (2004)\] .
K. T o g an o * e ta l*. ,P hys . Rev. Lett. **93 **, 247004 ( 2 00 4).
P. Ba d ica | type_superconductor, found_in Li$_2$Pt$_3$B, provides an_alternative way_to_study unconventional_SC,_especially that arising_from phonon pairing_mechanism. Indeed, the absence_of parity symmetry_coupled_with strong spin-orbit coupling, which results in an admixture of spin-singlet and spin-triplet pairing,_requires_a complete_reconceptualization_of_Cooper pairs and the nature_of the superconducting state.
We thank_P. Frigeri,_R. Kaur, I. Milat and H. Takeya for_useful_discussions. This work_is supported by the National Science Foundation under awards_No. NSF-EIA0121568 and NSF-DMR0318665, the Department_of Energy under_award_No._DEFG02-91ER45439, the Swiss National_Science Foundation and Petroleum Research Funds._We also acknowledge supports from ICAM_(HQY), the 2003-PFRA program of Japan Society_for the Promotion of Science (NH)_and the CEEX program of_Romanian Ministry_of Education and Research (PB).
[99]{}_See, e.g., T._Moriya and_K. Ueda, Rep._Prog. Phys., **66**, 1299 (2003); A._P. Mackenzie and_Y. Maeno, Rev. Mod. Phys. **75**,_657_(2003).
L. P. Gor’kov_and_E._I. Rashba,_Phys. Rev. Lett._**87**,_037004 (2001).
P._A._Frigeri *et al*., Phys. Rev. Lett._**92**,_097001 (2004).
P. A. Frigeri *et al*., cond-mat/0505108_(2005).
V.M. Edel’stein, Sov. Phys._JETP_**68**, 1244 (1989).
L. S._Levitov, Yu. V. Nazarov and_G. M. Eliashberg, JEPT Lett. **41**,_445 (1985).
K._V. Samokhin,_E. S. Zijlstra and S. K. Bose, Phys. Rev. B **69**,_094514 (2004) \[Erratum: Phys. Rev. B_**70**, 069902 (E) (2004)\].
K._Togano *et_al*.,_Phys. Rev. Lett._**93**,_247004 (2004).
P._Badica |
$ to each of them. Assume that every pair of neighboring virtual spins which correspond to different sites are initially in an (unnormalized) maximally entangled state $|I\rangle=\sum_{\alpha=1}^{D} |\alpha,\alpha\rangle$ often referred to as entangled *bond*. Then apply a map $$\label{curlyA}{\cal
A}=\sum_{i=1}^d\sum_{\alpha,\beta=1}^{D} A_{i,\alpha,\beta}
|i\rangle\langle\alpha,\beta|$$ to each of the $N$ sites. Here and in the following Greek indices correspond to the virtual systems. By writing $A_{i}$ for the $D\times D$ matrix with elements $A_{i,\alpha,\beta}$ we get that the coefficients of the final state when expressed in terms of a product basis are given by a matrix product $\operatorname{tr}\left[A_{i_1}A_{i_2}\cdots
A_{i_N}\right]$. In general the dimension of the entangled state $|I\rangle$ and the map ${\cal A}$ can both be site-dependent and we write $A_i^{[k]}$ for the $D_k\times D_{k+1}$ matrix corresponding to site $k\in\{1,\ldots,N\}$. States obtained in this way have then the form $$\label{MPS0}
|\psi\rangle=\sum_{i_1,\ldots,i_N=1}^d
\operatorname{tr}\left[A_{i_1}^{[1]}A_{i_2}^{[2]}\cdots
A_{i_N}^{[N]}\right]|i_1,i_2,\ldots,i_N\rangle\;,$$ and are called *matrix product states* [@MPSorigin]. As shown in [@Vi03] *every* state can be represented in this way if only the bond dimensions $D_k$ are sufficiently large. Hence, Eq.(\[MPS0\]) is a representation of states rather than the characterization of a specific class. However, typically states are referred to as MPS if they have a MPS-representation with small $D=\max_k D_k$ which (in the case of a sequence of states) does in particular not grow with $N$. Note that $\psi$ in Eq.(\[MPS0\]) is in general not normalized and that its MPS | $ to each of them. Assume that every pair of neighboring virtual spins which represent to unlike sites are initially in an (unnormalized) maximally entangled department of state $ |I\rangle=\sum_{\alpha=1}^{D } |\alpha,\alpha\rangle$ much referred to as entangle * shackle *. Then apply a function $ $ \label{curlyA}{\cal
A}=\sum_{i=1}^d\sum_{\alpha,\beta=1}^{D } A_{i,\alpha,\beta }
|i\rangle\langle\alpha,\beta|$$ to each of the $ N$ web site. Here and in the following Greek indices represent to the virtual systems. By writing $ A_{i}$ for the $ D\times D$ matrix with element $ A_{i,\alpha,\beta}$ we get that the coefficients of the final state when expressed in terminus of a product basis are given by a matrix intersection $ \operatorname{tr}\left[A_{i_1}A_{i_2}\cdots
A_{i_N}\right]$. In general the dimension of the entangled state $ |I\rangle$ and the function $ { \cal A}$ can both be site - subject and we write $ A_i^{[k]}$ for the $ D_k\times D_{k+1}$ matrix corresponding to site $ k\in\{1,\ldots, N\}$. States prevail in this way have then the form $ $ \label{MPS0 }
|\psi\rangle=\sum_{i_1,\ldots, i_N=1}^d
\operatorname{tr}\left[A_{i_1}^{[1]}A_{i_2}^{[2]}\cdots
A_{i_N}^{[N]}\right]|i_1,i_2,\ldots, i_N\rangle\;,$$ and are called * matrix product states * [ @MPSorigin ]. As shown in [ @Vi03 ] * every * state can be represented in this way if only the bond dimensions $ D_k$ are sufficiently large. Hence, Eq.(\[MPS0\ ]) is a representation of states preferably than the characterization of a specific class. However, typically department of state are refer to as MPS if they have a MPS - representation with small $ D=\max_k D_k$ which (in the shell of a sequence of states) does in particular not grow with $ N$. Note that $ \psi$ in Eq.(\[MPS0\ ]) is in cosmopolitan not normalized and that its MPS | $ to each of them. Assume thau every pair of ngithboriig virthal spinr which correspond to differxnt wites are initially in an (uvnormalizvd) maximaoly tntangled state $|I\czngle=\sum_{\alpha=1}^{S} |\alpka,\elpha\rangle$ oftgn referred do as entangleg *coud*. Then apply a map $$\label{curlyA}{\cal
A}=\stm_{i=1}^d\sum_{\sloha,\beta=1}^{D} A_{i,\alpra,\beua}
|i\wangmv\lcngle\alpha,\beta|$$ to each of the $N$ sites. Iere and in the following Greek indices clrredpond to the virtuwl systems. Vy wwuting $A_{i}$ for the $D\times D$ matrix wjth elements $A_{i,\alpha,\beta}$ we get that the coeffucuenhv of the fiial stwte when expvvssed it terms of a product nasis arw given by a matrix pcoduct $\operatorname{tt}\left[A_{i_1}A_{i_2}\cgocs
A_{i_N}\right]$. In general tye dikenshon ud tfe tntengmed stwte $|I\rangle$ ahd the map ${\xal A}$ can both be sotq-eependent and we wrytq $A_i^{[k]}$ for the $D_k\times D_{k+1}$ matrix correspmndjng to site $k\in\{1,\ldots,N\}$. Srates obtained in thid way havq then the form $$\label{MPS0}
|\psi\rangle=\sum_{i_1,\ldots,i_N=1}^d
\operdtornemd{tr}\oent[A_{i_1}^{[1]}X_{u_2}^{[2]}\cfots
A_{i_N}^{[N]}\right]|i_1,i_2,\ldots,i_N\rangle\;,$$ and are called *mwfrox product statef* [@MPSorigin]. Ad xrown in [@Vi03] *evgry* stacs dan be represented in thif way if only uhe bpnd dimensions $D_k$ are suffixiently largv. Hebce, Eq.(\[MPS0\]) is a repxesentation uf sjates tather than the charactzrizatjon of a spfcific clzrs. However, typicxllj stdtes are referred to as MPF if they havz a MPS-rdprexentatyon with slall $D=\max_k D_k$ which (in tje cade of a sequfnce of states) does in particulec not grow wijh $T$. Nmte that $\psi$ ln Eq.(\[MPS0\]) is in deneral not notmalized cnd thxt its MPS | $ to each of them. Assume that of virtual spins correspond to different (unnormalized) entangled state $|I\rangle=\sum_{\alpha=1}^{D} often referred to entangled *bond*. Then apply a map A}=\sum_{i=1}^d\sum_{\alpha,\beta=1}^{D} A_{i,\alpha,\beta} |i\rangle\langle\alpha,\beta|$$ to each of the $N$ sites. Here and in the Greek indices correspond to the virtual systems. By writing $A_{i}$ for the $D\times matrix elements we that the coefficients of the final state when expressed in terms of a product basis are by a matrix product $\operatorname{tr}\left[A_{i_1}A_{i_2}\cdots A_{i_N}\right]$. In general dimension of the entangled $|I\rangle$ and the map ${\cal can be site-dependent we $A_i^{[k]}$ the $D_k\times D_{k+1}$ corresponding to site $k\in\{1,\ldots,N\}$. States obtained in this way have then the form $$\label{MPS0} |\psi\rangle=\sum_{i_1,\ldots,i_N=1}^d \operatorname{tr}\left[A_{i_1}^{[1]}A_{i_2}^{[2]}\cdots A_{i_N}^{[N]}\right]|i_1,i_2,\ldots,i_N\rangle\;,$$ are called states* [@MPSorigin]. shown [@Vi03] state can be this way if only the bond sufficiently large. Hence, Eq.(\[MPS0\]) is a representation of rather than characterization of a specific class. However, states are referred to as MPS if they a MPS-representation with small $D=\max_k D_k$ which (in the case of a sequence of states) particular not grow with Note that $\psi$ Eq.(\[MPS0\]) in not and that MPS | $ to each of them. Assume that eveRy pair of neIghboRinG viRtUal sPins Which corresponD To diFferent sites are initialLy in aN (uNNormALiZed) maXimally ENtANGleD sTaTe $|I\RaNGlE=\sum_{\aLphA=1}^{D} |\alpha,\Alpha\ranglE$ ofTeN referred to aS EnTangled *bonD*. ThEn apply a map $$\lAbeL{curlya}{\cAl
A}=\SUm_{i=1}^d\sUm_{\aLpha,\bEta=1}^{D} A_{i,\ALpha,\beTa}
|i\rangle\LaNGle\alpHA,\beta|$$ to EACh Of thE $N$ sites. Here and in tHE fOLlowing Greek inDices cOrREsPONd tO thE virtual syStEms. By WRiting $A_{I}$ FoR THE $D\tIMes D$ matrix witH elements $A_{i,\ALphA,\beta}$ wE gEt tHAt the cOeffiCiENts Of the final sTate When expreSsed in TErms of a PRoduct bAsis arE giVen By a mATrIx ProDuCT $\opERaTorNAme{Tr}\left[A_{i_1}a_{i_2}\CdOts
A_{i_n}\rigHT]$. iN GeneRal The dImensIon of the entanGleD staTE $|I\rAngle$ And thE map ${\CaL A}$ can Both be Site-dEpEndent and we writE $A_i^{[k]}$ For the $D_k\tImeS D_{K+1}$ maTrIx corREspondIng To sIte $k\in\{1,\lDots,N\}$. StATes ObTAINeD in this way have then ThE FOrM $$\label{MPs0}
|\psi\raNGlE=\sUM_{i_1,\ldots,i_n=1}^d
\OpeRatoRNAme{tr}\Left[a_{I_1}^{[1]}A_{I_2}^{[2]}\cdots
A_{i_n}^{[N]}\righT]|I_1,i_2,\LdOts,i_N\raNgLe\;,$$ and aRe CalLed *MatriX ProdUct staTes* [@MPSorIgin]. AS Shown in [@Vi03] *every* STate can be reprESeNTEd IN thiS waY if only the bOnd dIMensIons $d_K$ aRe sUFficiEntly LaRGe. hEnce, Eq.(\[MPS0\]) is a represeNtAtion oF statEs rather than tHe characteRIZAtion of a SpecIFiC Class. However, tyPicalLy states arE Referred To as MpS if they Have a MPS-rEPResentatIon WitH smAll $d=\MAx_K D_k$ which (in the CASe of A sEquence Of sTates) doEs iN paRtiCulAr Not grow wiTh $N$. Note tHaT $\pSi$ In eq.(\[MpS0\]) is iN General nOt NorMaLizEd and THat its mPS | $ to each of them. Assumethat every pair of ne ig hbor ingvirtual spinsw hich correspond to differe nt si te s are in itial ly in a n ( u n nor ma li zed )m ax imall y e ntangle d state $| I\r an gle=\sum_{\a l ph a=1}^{D} | \al pha,\alpha\r ang le$ of te n r e ferre d t o asentang l ed *bo nd*. Then a p ply am ap $$\l a b el {cur lyA}{\cal
A}=\sum _ {i = 1}^d\sum_{\alp ha,\be ta = 1} ^ { D}A_{ i,\alpha,\ be ta}
| i \rangle \ la n g l e\a l pha,\beta|$$to each oft he$N$ si te s.H ere an d inth e fo llowing Gre ek i ndices co rrespo n d to th e virtua l syst ems . B y wr i ti ng $A _{ i }$f or th e $D \times D $ma trixwith e l e ment s $ A_{i ,\alp ha,\beta}$ we ge t th a t t he co effic ient sof th e fina l sta te when expressed interms ofa p ro duc tbasis are gi ven by a matr ix prod u ct$\ o p e ra torname{tr}\left[A _{ i _ 1} A_{i_2}\ cdotsA _{ i_ N }\right] $. In gen e r al th e di m en sion ofthe en t an gl ed stat e$|I\ra ng le$ an d the map${\cal A}$ can both be site-depend e nt and we wri t e$ A _i ^ {[k] }$for the $D_ k\ti m es D _{k+ 1 }$ ma t rix c orres po n di n g to site $k\in\{1, \l dots,N \}$.States obtain ed in this w a y have t hent he form $$\label{ MPS0}
|\psi\ran g le=\sum_ {i_1, \ldots,i _N=1}^d
\ o p eratorna me{ tr} \le ft[ A _ {i _1}^{[1]}A_{i _ 2 }^{[ 2] }\cdots
A_ {i_N}^{ [N] }\r igh t]| i_ 1,i_2,\ld ots,i_N\ ra ng le \; ,$$ anda re calle d*ma tr ixprodu c t stat es* [ @MPS or ig i n]. As sho w ni n [@V i0 3] *ev ery *state can bereprese nted in t his wayif o nly the bond dimensi on s $D_k$ ar esuf ficien t l y large. Hence, Eq.(\[MPS0\]) i s a repr ese ntati on o f statesrat her th ant he cha racter izati on of a spec i f ic cl as s. However , typ icall ystat es arereferred to as MPS ifthey have a M PS- repr e s en tat i on wit hs mal l $D=\max_k D_k$which (inth e c ase of a s e que nc e of st ates) d oes i n partic ular notgrow with $ N$.N o tethat $\psi $ in Eq. (\[MPS0\] ) is i n g enera l n ot nor ma liz ed an d that its MPS | $ to_each of_them. Assume that every_pair of_neighboring_virtual spins_which_correspond to different_sites are initially_in an (unnormalized) maximally_entangled state $|I\rangle=\sum_{\alpha=1}^{D}_|\alpha,\alpha\rangle$_often referred to as entangled *bond*. Then apply a map $$\label{curlyA}{\cal
A}=\sum_{i=1}^d\sum_{\alpha,\beta=1}^{D} A_{i,\alpha,\beta}
|i\rangle\langle\alpha,\beta|$$ to each_of_the $N$_sites._Here_and in the following Greek_indices correspond to the virtual_systems. By_writing $A_{i}$ for the $D\times D$ matrix with_elements_$A_{i,\alpha,\beta}$ we get_that the coefficients of the final state when expressed_in terms of a product basis_are given by_a_matrix_product $\operatorname{tr}\left[A_{i_1}A_{i_2}\cdots
A_{i_N}\right]$. In general_the dimension of the entangled state_$|I\rangle$ and the map ${\cal A}$_can both be site-dependent and we write_$A_i^{[k]}$ for the $D_k\times D_{k+1}$ matrix_corresponding to site $k\in\{1,\ldots,N\}$. States_obtained in_this way have then the_form $$\label{MPS0}
|\psi\rangle=\sum_{i_1,\ldots,i_N=1}^d
\operatorname{tr}\left[A_{i_1}^{[1]}A_{i_2}^{[2]}\cdots
A_{i_N}^{[N]}\right]|i_1,i_2,\ldots,i_N\rangle\;,$$ and_are called_*matrix product states*_[@MPSorigin]. As shown in [@Vi03] *every*_state can be_represented in this way if only_the_bond dimensions $D_k$_are_sufficiently_large. Hence,_Eq.(\[MPS0\]) is a_representation_of states_rather_than the characterization of a specific_class._However, typically states are referred to as_MPS if they have_a_MPS-representation with small $D=\max_k_D_k$ which (in the case_of a sequence of states) does_in particular_not grow_with $N$. Note that $\psi$ in Eq.(\[MPS0\]) is in general not_normalized and that its MPS |
haronov and D. Bohm, Phys. Rev. [**115**]{}, 485 (1959). R. Jackiw and S.-Y. Pi, Phys. Rev. D[**42**]{}, 3500 (1990). E. Fradkin, Phys. Rev. Lett. [**63**]{}, 322 (1989). S. K. Paul and A. Khare, Phys. Lett. B[**171**]{}, 244 (1986). G. W. Semenoff and P. Sodano, Nucl. Phys. [**B328**]{}, 753 (1989); T. Matsuyama, Phys. Rev. D[**42**]{}, 3469 (1990).
---
abstract: |
Let $(X,\tau)$ be a Hausdorff space, where $X$ is an infinite set. The compact complement topology $\tau^{\star}$ on $X$ is defined by: $\tau^{\star}=\{\emptyset\} \cup \{X\setminus M, \text{where $M$ is compact in $(X,\tau)$}\}$. In this paper, properties of the space $(X, \tau^{\star})$ are studied in $\mathbf{ZF}$ and applied to a characterization of $k$-spaces, to the Sorgenfrey line, to some statements independent of $\mathbf{ZF}$, as well as to partial topologies that are among Delfs-Knebusch generalized topologies. Between other results, it is proved that the axiom of countable multiple choice (**CMC**) is equivalent with each of the following two sentences: (i) every Hausdorff first countable space is a $k$-space, (ii) every metrizable space is a $k$-space. A **ZF**-example of a countable metrizable space whose compact complement topology is not first countable is given.
[**Keywords:**]{} compact complement topology; countable multiple choice; $k$-space; sequential space; Sorgenfrey line; Delfs-Knebusch generalized topological space; partial topology.
address:
- 'Department of Mathematics, University of Aegean, Karlovassi, Samos 83200, Greece'
- 'Department of Mathematics, King Abdulaziz University, P.O.Box: 80203 Jeddah 21589, Saudi Arabia'
- 'Institute of Mathematics, Cracow University of Technology, Warszawska 24, 31-155 Kraków, Poland'
- 'Department of Mathematics, King Abdul | haronov and D. Bohm, Phys. Rev. [ * * 115 * * ] { }, 485 (1959). R. Jackiw and S.-Y. Pi, Phys. Rev. D[**42 * * ] { }, 3500 (1990). E. Fradkin, Phys. Rev. Lett. [ * * 63 * * ] { }, 322 (1989). S. K. Paul and A. Khare, Phys. Lett. B[**171 * * ] { }, 244 (1986). G. W. Semenoff and P. Sodano, Nucl. Phys. [ * * B328 * * ] { }, 753 (1989); T. Matsuyama, Phys. Rev. D[**42 * * ] { }, 3469 (1990).
---
abstract: |
Let $ (X,\tau)$ be a Hausdorff space, where $ X$ is an infinite set. The compendious complement regional anatomy $ \tau^{\star}$ on $ X$ is specify by: $ \tau^{\star}=\{\emptyset\ } \cup \{X\setminus M, \text{where $ M$ is compact in $ (X,\tau)$}\}$. In this paper, property of the space $ (X, \tau^{\star})$ are learn in $ \mathbf{ZF}$ and apply to a characterization of $ k$-spaces, to the Sorgenfrey line, to some statements independent of $ \mathbf{ZF}$, equally well as to partial topologies that are among Delfs - Knebusch generalized topologies. Between other result, it is proved that the axiom of countable multiple choice (* * CMC * *) is equivalent with each of the following two sentences: (iodine) every Hausdorff first countable space is a $ k$-space, (ii) every metrizable space is a $ k$-space. A * * ZF**-example of a countable metrizable quad whose compact complement topology is not first countable is pass.
[ * * Keywords :* * ] { } compact complement topology; countable multiple choice; $ k$-space; sequential space; Sorgenfrey cable; Delfs - Knebusch generalized topological space; partial topology.
address:
-' Department of Mathematics, University of Aegean, Karlovassi, Samos 83200, Greece'
-' Department of Mathematics, King Abdulaziz University, P.O.Box: 80203 Jeddah 21589, Saudi Arabia'
-' Institute of Mathematics, Cracow University of Technology, Warszawska 24, 31 - 155 Kraków, Poland'
-' Department of Mathematics, King Abdul | harlnov and D. Bohm, Phys. Rev. [**115**]{}, 485 (1959). R. Jackiw anb S.-Y. Pi, Phys. Dev. D[**42**]{}, 3500 (1990). D. Fradkin, Phys. Rev. Lett. [**63**]{}, 322 (1989). S. K. Paul and A. Khare, Phys. Lett. B[**171**]{}, 244 (1986). G. W. Demenoff and P. Sodano, Nndl. Phys. [**B328**]{}, 753 (1989); T. Latsbyema, Phys. Rev. D[**42**]{}, 3469 (1990).
---
abstract: |
Let $(X,\tau)$ te a Hausdorff space, where $X$ is an infigite sey. Hhe compact coiplekqnt fopology $\tau^{\star}$ on $X$ is defined bg: $\tau^{\ster}=\{\emptyset\} \cup \{C\setminus M, \text{where $M$ is comoact in $(X,\tau)$}\}$. In thls paper, pripereues of the soace $(X, \tau^{\star})$ are stusied in $\mathbf{ZF}$ and applied to a chcracterizatuob ov $k$-spaces, to the Forgenfrey llme, to vome ststements indepcndenv of $\mathbf{ZF}$, as well as vo partial topologief that ara cmong Delfs-Knebusch gwnwralieed tmpolutier. Bttwxen other reaults, it ia proved thqt the axiom of coumtwvle multiple dhoice (**CIC**) is equivalent with each of the follofinf two sentences: (i) every Hausdorff first counjable spacq is a $k$-space, (ii) every metrizable space is a $k$-spawe. A **VF**-dxanpjd ov a countable metrizable space whose compact skmklekent topology ls not first countsbpe ys given.
[**Kgywords:**]{} cojpact complement tlpology; counrable muluiple choice; $k$-space; sequential space; Sorgenfgey oine; Delfs-Knebusch generalizeb topokogicsl space; partial topoloyy.
addrsss:
- 'Departmfnt of Maffematics, Universkty ox Aegean, Karlovassi, Samos 83200, Greece'
- 'Dxpartkent of Matnematiss, King Abfulazls University, P.O.Box: 80203 Jedbah 21589, Vaudi Arabla'
- 'Institute of Mathematics, Crarpw Universiti ox Tvchnology, Warsdawska 24, 31-155 Kraków, [oland'
- 'Departmgnt of Machematkcs, King Agdul | haronov and D. Bohm, Phys. Rev. [**115**]{}, R. and S.-Y. Phys. Rev. D[**42**]{}, Rev. [**63**]{}, 322 (1989). K. Paul and Khare, Phys. Lett. B[**171**]{}, 244 (1986). W. Semenoff and P. Sodano, Nucl. Phys. [**B328**]{}, 753 (1989); T. Matsuyama, Phys. D[**42**]{}, 3469 (1990). --- abstract: | Let $(X,\tau)$ be a Hausdorff space, where is infinite The complement topology $\tau^{\star}$ on $X$ is defined by: $\tau^{\star}=\{\emptyset\} \cup \{X\setminus M, \text{where $M$ is compact $(X,\tau)$}\}$. In this paper, properties of the space \tau^{\star})$ are studied in and applied to a characterization $k$-spaces, the Sorgenfrey to statements of $\mathbf{ZF}$, as as to partial topologies that are among Delfs-Knebusch generalized topologies. Between other results, it is proved that axiom of choice (**CMC**) equivalent each the following two every Hausdorff first countable space is every metrizable space is a $k$-space. A **ZF**-example a countable space whose compact complement topology is first countable is given. [**Keywords:**]{} compact complement topology; multiple choice; $k$-space; sequential space; Sorgenfrey line; Delfs-Knebusch generalized topological space; partial topology. address: - Mathematics, University of Aegean, Samos 83200, Greece' 'Department Mathematics, Abdulaziz P.O.Box: 80203 21589, Saudi Arabia' - 'Institute of Mathematics, Cracow University of Technology, 24, 31-155 Kraków, Poland' - 'Department of Mathematics, King Abdul | haronov and D. Bohm, Phys. Rev. [**115**]{}, 485 (1959). R. JaCkiw and S.-Y. PI, Phys. rev. d[**42**]{}, 3500 (1990). E. FRaDkin, phys. rev. Lett. [**63**]{}, 322 (1989). S. K. Paul aND A. KhAre, Phys. Lett. B[**171**]{}, 244 (1986). G. W. Semenoff And P. SOdANo, NuCL. PHys. [**B328**]{}, 753 (1989); T. matsuyaMA, PHYS. ReV. D[**42**]{}, 3469 (1990).
---
AbStrAcT: |
leT $(X,\tau)$ Be a hausdorFf space, wheRe $X$ Is An infinite seT. thE compact coMplEment topologY $\taU^{\star}$ oN $X$ Is dEFined By: $\tAu^{\staR}=\{\emptySEt\} \cup \{X\Setminus M, \TeXT{where $m$ Is compaCT In $(x,\tau)$}\}$. in this paper, properTIeS Of the space $(X, \tau^{\Star})$ arE sTUdIED in $\MatHbf{ZF}$ and apPlIed to A CharactERiZATIon OF $k$-spaces, to the sorgenfrey lINe, tO some sTaTemENts indEpendEnT Of $\mAthbf{ZF}$, as weLl as To partial TopoloGIes that ARe among delfs-KNebUscH genERaLiZed ToPOloGIeS. BeTWeeN other reSuLtS, it is ProvED THAt thE axIom oF counTable multiple ChoIce (**Cmc**) is EquivAlent With EaCh of tHe follOwing TwO sentences: (i) everY HauSdorff firSt cOuNtaBlE spacE Is a $k$-spAce, (Ii) eVery metRizable SPacE iS A $K$-SpAce. A **ZF**-example of a coUnTABlE metrizaBle spaCE wHoSE compact CoMplEmenT TOpoloGy is NOt First couNtable IS gIvEn.
[**KeywoRdS:**]{} compaCt ComPleMent tOPoloGy; counTable mulTiple CHoice; $k$-space; seqUEntial space; SoRGeNFReY Line; delFs-Knebusch gEnerALizeD topOLoGicAL spacE; partIaL ToPOlogy.
address:
- 'DepartmEnT of MatHematIcs, University Of Aegean, KaRLOVassi, SamOs 83200, GrEEcE'
- 'department of MaThemaTics, King AbDUlaziz UnIversIty, P.O.Box: 80203 jeddah 21589, SauDI arabia'
- 'InStiTutE of matHEMaTics, Cracow UniVERsitY oF TechnoLogY, WarszaWskA 24, 31-155 KrAkóW, PoLaNd'
- 'DepartmEnt of MatHeMaTiCs, kinG AbduL | haronov and D. Bohm, Phys. Rev. [**1 15**] {}, 48 5(195 9).R. Jackiw andS .-Y. Pi, Phys. Rev. D[**42 **]{} ,3 500( 19 90).E. Frad k in , Phy s. R ev. L e tt . [** 63* *]{}, 3 22 (1989). S. K . Paul and A . K hare, Phys . L ett. B[**171 **] {}, 24 4(19 8 6). G . W . Sem enoffa nd P.Sodano, N uc l . Phys . [**B32 8 * *] {},753 (1989); T. Ma t su y ama, Phys. Rev . D[** 42 * *] { } , 3 469 (1990).
-- -
abs t ract: | L et$ (X,\tau)$ bea Hausdorff spa ce, wh er e $ X $ is a n inf in i teset. The co mpac t complem ent to p ology $ \ tau^{\s tar}$on$X$ isd ef in edby : $\ t au ^{\ s tar }=\{\emp ty se t\} \ cup\ { X \ setm inu s M, \tex t{where $M$ i s c ompa c t i n $(X ,\tau )$}\ }$ . Inthis p aper, p roperties of th e sp ace $(X,\ta u^ {\s ta r})$a re stu die d i n $\mat hbf{ZF} $ an da p p li ed to a characteri za t i on of $k$- spaces , t ot he Sorge nf rey lin e , to s omes ta tementsindepe n de nt of $\m at hbf{ZF }$ , a s w ell a s topartia l topolo giest hat are amongD elfs-Knebusch ge n e ra l ized to pologies. B etwe e n ot herr es ult s , itis pr ov e dt hat the axiom of co un tablemulti ple choice (* *CMC**) is e q uivalent wit h e a ch of the foll owing two sente n ces: (i) ever y Hausdo rff first c ountable sp ace is a$ k $- space, (ii) e v e ry m et rizable sp ace isa $ k$- spa ce. A **ZF**-e xample o faco un tab le me t rizablesp ace w hos e com p act co mplem entto po l ogy is not fi r s t co un ta bleisgi ven.
[** Keyword s:**]{} c omp a ct c om pl ement t opology; coun ta ble multip le ch oice;$ k $-space; sequential space; Sorg e nfrey l ine ; Del fs-K nebusch g ene ralize d t o pologi cal sp ace;pa rti a l topo l o gy .
a dd ress:
- 'D e p art mentof Mat hematic s, University of A e gea n, Karlovassi , S amos 8 32 00, Gr e ece '- 'D e p artment of Math ematics, K in g A bdulaziz U n ive rs ity, P. O.Box:80203 Jeddah21589, Sa udi Arabi a'
- ' I n sti tute of Ma thematic s, Cracow Unive r si ty of Te chnolo gy , W arsza wska 2 4 , 3 1-155 Krakó w, Polan d'
-'D epartmen t of Mathematics, KingAbdul | haronov and_D. Bohm,_Phys. Rev. [**115**]{}, 485_(1959). R._Jackiw_and S.-Y._Pi,_Phys. Rev. D[**42**]{},_3500 (1990). E._Fradkin, Phys. Rev. Lett._[**63**]{}, 322 (1989)._S._K. Paul and A. Khare, Phys. Lett. B[**171**]{}, 244 (1986). G. W. Semenoff and_P._Sodano, Nucl._Phys._[**B328**]{},_753 (1989); T. Matsuyama, Phys._Rev. D[**42**]{}, 3469 (1990).
---
abstract:_|
_ Let $(X,\tau)$ be a Hausdorff space,_where_$X$ is an_infinite set. The compact complement topology $\tau^{\star}$ on $X$_is defined by: $\tau^{\star}=\{\emptyset\} \cup \{X\setminus_M, \text{where $M$_is_compact_in $(X,\tau)$}\}$. In this_paper, properties of the space $(X,_\tau^{\star})$ are studied in $\mathbf{ZF}$ and_applied to a characterization of $k$-spaces, to_the Sorgenfrey line, to some statements_independent of $\mathbf{ZF}$, as well_as to_partial topologies that are among_Delfs-Knebusch generalized topologies._Between other_results, it is_proved that the axiom of countable_multiple choice (**CMC**)_is equivalent with each of the_following_two sentences: (i)_every_Hausdorff_first countable_space is a_$k$-space,_(ii) every_metrizable_space is a $k$-space. A **ZF**-example_of_a countable metrizable space whose compact complement_topology is not first_countable_is given.
_ [**Keywords:**]{} compact complement topology;_countable multiple choice; $k$-space; sequential space;_Sorgenfrey line;_Delfs-Knebusch generalized_topological space; partial topology.
address:
- 'Department of Mathematics, University of Aegean, Karlovassi,_Samos 83200, Greece'
- 'Department of Mathematics,_King Abdulaziz University, P.O.Box:_80203 Jeddah_21589,_Saudi Arabia'
- 'Institute_of_Mathematics, Cracow_University of Technology, Warszawska 24, 31-155 Kraków,_Poland'
- 'Department_of Mathematics, King Abdul |
has just three nonzero eigenvalues. Indeed, let us introduce new decay matrices as $$\gamma_{\mu\nu}^{r}=\dfrac{N\Gamma^{r}}{3} B_{\mu\nu},$$ where $$B_{\mu\nu}=\dfrac{3}{N}\sum_{j} e_{j}^{\mu}e_{j}^{\nu}$$ is $3\times 3$ matrix in coordinate space with ${\rm Tr}\hat{B}=3$. It is easy to see that ${\rm Tr}\bigl [ (\hat{\Gamma}^{r} )^{n}\bigr]={\rm Tr}\bigl [ (\hat{\gamma}^{r} )^{n}\bigr]$ for any integer $n$, i.e., the $N\times N$ matrix $\Gamma_{jk}^{r}$ has only three non-zero eigenvalues coinciding with those of matrix $\gamma_{\mu\nu}^{r}$ $$\label{gamma-coop}
\Gamma_{\mu}^{r}=\dfrac{N\Gamma^{r}}{3}\,\lambda_{\mu},$$ where $\lambda_{\mu}\sim 1$ are eigenvalues of $B_{\mu\nu}$. Note that the decay rates of the remaining $N-3$ subradiant states vanish in the long wave approximation used here; they acquire finite values in the next order in $(kr_{0})^{2}$.
Let us turn to non-radiative coupling, described by matrix $\Gamma_{jk}^{nr}$. Its diagonal elements, $\Gamma_{jj}^{nr}$, describe nonradiative energy exchange between excited dipole and NP plasmon modes with all angular momenta, as indicated by polarizabilities $\alpha''_{l}$ in Eq. (\[gamma-perp\]). The non-diagonal elements of $\Gamma_{jk}^{nr}$ describe a process by which a plasmon nonradiatively excited in the NP by dipole $k$ transfers its energy to another dipole $j$. In general, due to high-$l$ plasmons involved in nonradiative coupling, the eigenstates of $\Gamma_{jk}^{nr}$ are *different* from those of plasmon-enhanced radiative coupling $\Gamma_{jk}^{r}$ which contains only dipole ($l=1$) plasmon mode. Therefore, the eigenstates of full decay matrix $\Gamma_{jk}=\Gamma_{jk}^{nr}+\Gamma_{jk}^{nr}$ are *not* pure superradiant and subradiant states but their admixtures. However, the high-$l$ plasmon contribution to $\Gamma_{jk}^{nr}$ is significant only at | has just three nonzero eigenvalues. Indeed, let us insert newfangled decay matrices as $ $ \gamma_{\mu\nu}^{r}=\dfrac{N\Gamma^{r}}{3 } B_{\mu\nu},$$ where $ $ B_{\mu\nu}=\dfrac{3}{N}\sum_{j } e_{j}^{\mu}e_{j}^{\nu}$$ is $ 3\times 3 $ matrix in coordinate space with $ { \rm Tr}\hat{B}=3$. It is easy to witness that $ { \rm Tr}\bigl [ (\hat{\Gamma}^{r }) ^{n}\bigr]={\rm Tr}\bigl [ (\hat{\gamma}^{r }) ^{n}\bigr]$ for any integer $ n$, i.e., the $ N\times N$ matrix $ \Gamma_{jk}^{r}$ has only three non - zero eigenvalue coinciding with those of matrix $ \gamma_{\mu\nu}^{r}$ $ $ \label{gamma - coop }
\Gamma_{\mu}^{r}=\dfrac{N\Gamma^{r}}{3}\,\lambda_{\mu},$$ where $ \lambda_{\mu}\sim 1 $ are eigenvalues of $ B_{\mu\nu}$. bill that the decay rates of the remaining $ N-3 $ subradiant state vanish in the long wave estimate used here; they acquire finite values in the next order in $ (kr_{0})^{2}$.
Let us turn to non - radiative yoke, described by matrix $ \Gamma_{jk}^{nr}$. Its diagonal elements, $ \Gamma_{jj}^{nr}$, trace nonradiative energy exchange between aroused dipole and NP plasmon modes with all angular momentum, as indicated by polarizabilities $ \alpha''_{l}$ in Eq. (\[gamma - perp\ ]). The non - diagonal elements of $ \Gamma_{jk}^{nr}$ describe a process by which a plasmon nonradiatively activated in the NP by dipole $ k$ transfers its energy to another dipole $ j$. In general, due to high-$l$ plasmons involved in nonradiative coupling, the eigenstates of $ \Gamma_{jk}^{nr}$ are * different * from those of plasmon - enhanced radiative coupling $ \Gamma_{jk}^{r}$ which contains only dipole ($ l=1 $) plasmon mode. Therefore, the eigenstates of full decay matrix $ \Gamma_{jk}=\Gamma_{jk}^{nr}+\Gamma_{jk}^{nr}$ are * not * pure superradiant and subradiant states but their admixtures. However, the high-$l$ plasmon contribution to $ \Gamma_{jk}^{nr}$ is significant merely at | had just three nonzero eigtnvalues. Indeed, lgt us invroduce new decxy matrices as $$\gamma_{\mu\nu}^{r}=\dfrec{N\Gqmma^{r}}{3} B_{\mu\nu},$$ where $$B_{\mu\nu}=\dfraz{3}{N}\sum_{j} e_{j}^{\lu}e_{j}^{\nu}$$ iw $3\tinws 3$ matrix in coordinate dpacz xith ${\rm Tr}\hat{B}=3$. Lt is easy do see that ${\rm Tf}\bngl [ (\hat{\Gamma}^{r} )^{n}\bigr]={\rm Tr}\bigl [ (\hat{\gaima}^{r} )^{n}\bogg]$ for any inteder $m$, i.e., fhe $N\times N$ matrix $\Gamma_{jk}^{r}$ has ohly thrte non-zero eigenvakues coinciding with those of latrix $\gamma_{\mu\nu}^{r}$ $$\pabel{gamma-ciop}
\Gwnma_{\mu}^{r}=\dfrac{N\Eamma^{r}}{3}\,\lambda_{\mu},$$ where $\lzmbda_{\mu}\sim 1$ are eigenvalues of $C_{\mu\nu}$. Note that tye fgcay rates oh the gemaining $N-3$ smnradiatt statrs vanish in tme loig wqve approximation usev here; they acquire sinite vanuzs in the next order un $(kr_{0})^{2}$.
Lej us durn ro von-daviafive clupming, descrjbed by mateix $\Gamma_{jk}^{nr}$. Its doadinal elements, $\Gamma_{tj}^{gr}$, describe nonradiative energy exchangt betseen excited dipole and NP plasmon modes witj all angtlar momenta, as indicated by polarizabilities $\alpva''_{l}$ ii Dq. (\[gcnma-pefp\]). Hhe non-diagonal elements of $\Gamma_{jk}^{nr}$ describq a pgocess by which a plasmon nomrwdowtively excitgd in tks HP by dipole $k$ trajsfers yts ebergy to wnotner dipole $j$. In general, due to high-$l$ plcsmins involved in nouradiative cuuplong, tne eigenstates of $\Gamma_{lk}^{nr}$ ade *differenh* from thkre of plasmon-enhxncvd rddiative coupling $\Gamma_{jk}^{r}$ which coitainx only aipoke ($l=1$) pjasmon modf. Thevafore, the eigenstahes oy fuln decay mahrix $\Gamma_{jk}=\Gamma_{jk}^{nr}+\Gamma_{jk}^{nr}$ ece *not* pure sopesragiant anb subrsdiant states but their admnxtures. Kowevef, the high-$m$ plasmmn contribueion to $\Gamma_{bn}^{nr}$ is signihicant onjy ar | has just three nonzero eigenvalues. Indeed, let new matrices as B_{\mu\nu},$$ where $$B_{\mu\nu}=\dfrac{3}{N}\sum_{j} in space with ${\rm It is easy see that ${\rm Tr}\bigl [ (\hat{\Gamma}^{r} Tr}\bigl [ (\hat{\gamma}^{r} )^{n}\bigr]$ for any integer $n$, i.e., the $N\times N$ matrix has only three non-zero eigenvalues coinciding with those of matrix $\gamma_{\mu\nu}^{r}$ $$\label{gamma-coop} \Gamma_{\mu}^{r}=\dfrac{N\Gamma^{r}}{3}\,\lambda_{\mu},$$ $\lambda_{\mu}\sim are of Note that the decay rates of the remaining $N-3$ subradiant states vanish in the long wave used here; they acquire finite values in the order in $(kr_{0})^{2}$. Let turn to non-radiative coupling, described matrix Its diagonal $\Gamma_{jj}^{nr}$, nonradiative exchange between excited and NP plasmon modes with all angular momenta, as indicated by polarizabilities $\alpha''_{l}$ in Eq. (\[gamma-perp\]). The elements of a process which plasmon excited in the dipole $k$ transfers its energy to In general, due to high-$l$ plasmons involved in coupling, the of $\Gamma_{jk}^{nr}$ are *different* from those plasmon-enhanced radiative coupling $\Gamma_{jk}^{r}$ which contains only dipole plasmon mode. Therefore, the eigenstates of full decay matrix $\Gamma_{jk}=\Gamma_{jk}^{nr}+\Gamma_{jk}^{nr}$ are *not* pure superradiant and but their admixtures. However, high-$l$ plasmon contribution $\Gamma_{jk}^{nr}$ significant at | has just three nonzero eigenvAlues. IndeeD, let uS inTroDuCe neW decAy matrices as $$\gaMMa_{\mu\Nu}^{r}=\dfrac{N\Gamma^{r}}{3} B_{\mu\nu},$$ whEre $$B_{\mU\nU}=\DfraC{3}{n}\sUm_{j} e_{j}^{\Mu}e_{j}^{\nu}$$ iS $3\TiMES 3$ maTrIx In cOoRDiNate sPacE with ${\rm tr}\hat{B}=3$. It is EasY tO see that ${\rm Tr}\BIgL [ (\hat{\Gamma}^{r} )^{N}\biGr]={\rm Tr}\bigl [ (\haT{\gaMma}^{r} )^{n}\bIgR]$ foR Any inTegEr $n$, i.e., The $N\tiMEs N$ matRix $\Gamma_{jK}^{r}$ HAs only THree non-ZERo EigeNvalues coinciding WItH Those of matrix $\gAmma_{\mu\Nu}^{R}$ $$\LaBEL{gaMma-Coop}
\Gamma_{\mU}^{r}=\Dfrac{n\gamma^{r}}{3}\,\lAMbDA_{\MU},$$ whERe $\lambda_{\mu}\sim 1$ Are eigenvalUEs oF $B_{\mu\nu}$. noTe tHAt the dEcay rAtES of The remaininG $N-3$ suBradiant sTates vANish in tHE long waVe apprOxiMatIon uSEd HeRe; tHeY AcqUIrE fiNIte Values in ThE nExt orDer iN $(KR_{0})^{2}$.
lEt us TurN to nOn-radIative couplinG, deScriBEd bY matrIx $\GamMa_{jk}^{Nr}$. its diAgonal ElemeNtS, $\Gamma_{jj}^{nr}$, descrIbe nOnradiatiVe eNeRgy ExChangE BetweeN exCitEd dipolE and NP pLAsmOn MODEs With all angular momeNtA, AS iNdicated By polaRIzAbILities $\alPhA''_{l}$ iN Eq. (\[gAMMa-perP\]). The NOn-Diagonal ElemenTS oF $\GAmma_{jk}^{nR}$ dEscribE a ProCesS by whICh a pLasmon NonradiaTivelY Excited in the NP BY dipole $k$ transFErS ITs ENergY to Another dipoLe $j$. IN GeneRal, dUE tO hiGH-$l$ plaSmons InVOlVEd in nonradiative couPlIng, the EigenStates of $\Gamma_{Jk}^{nr}$ are *difFEREnt* from tHose OF pLAsmon-enhanced rAdiatIve couplinG $\gamma_{jk}^{r}$ Which Contains Only dipolE ($L=1$) Plasmon mOde. theRefOre, THE eIgenstates of fULL decAy Matrix $\GAmmA_{jk}=\GammA_{jk}^{Nr}+\GAmmA_{jk}^{Nr}$ Are *not* purE superraDiAnT aNd SubRadiaNT states bUt TheIr AdmIxturES. HowevEr, the High-$L$ pLaSMon ContribUTiON To $\GaMmA_{jK}^{nr}$ iS siGnIficaNt onLY at | has just three nonzero ei genvalues. Inde ed, le tus i ntro duce new decay matr ices as $$\gamma_{\mu\ nu}^{ r} = \dfr a c{ N\Gam ma^{r}} { 3} B _{\ mu \n u}, $$ wh ere $ $B_ {\mu\nu }=\dfrac{3 }{N }\ sum_{j} e_{j } ^{ \mu}e_{j}^ {\n u}$$ is $3\t ime s 3$ m at rix in co ord inate space with $ {\rm Tr}\ ha t {B}=3$ . It ise a sy tosee that ${\rm Tr } \b i gl [ (\hat{\Ga mma}^{ r} )^ { n }\b igr ]={\rm Tr} \b igl [ (\hat{\ g am m a } ^{r } )^{n}\bigr]$ for any in t ege r $n$, i .e. , the $ N\tim es N$matrix $\Ga mma_ {jk}^{r}$ has o n ly thre e non-ze ro eig env alu es c o in ci din gw ith th ose ofmatrix $ \g am ma_{\ mu\n u } ^ { r}$$$\ labe l{gam ma-coop}
\Gam ma_ {\mu } ^{r }=\df rac{N \Gam ma ^{r}} {3}\,\ lambd a_ {\mu},$$ where$\la mbda_{\mu }\s im 1$ a re ei g envalu esof$B_{\mu \nu}$.N ote t h a t t he decay rates ofth e re maining$N-3$s ub ra d iant sta te s v anis h in th e lo n gwave app roxima t io nused he re ; they a cqu ire fini t e va lues i n the ne xt or d er in $(kr_{0} ) ^{2}$.
Let u s t u r nt o no n-r adiative co upli n g, d escr i be d b y matr ix $\ Ga m ma _ {jk}^{nr}$. Its dia go nal el ement s, $\Gamma_{j j}^{nr}$,d e s cribe no nrad i at i ve energy exch angebetween ex c ited dip ole a nd NP pl asmon mod e s with al l a ngu lar mo m e nt a, as indicat e d bypo larizab ili ties $\ alp ha' '_{ l}$ i n Eq. (\[ gamma-pe rp \] ). T henon-d i agonal e le men ts of $\Ga m ma_{jk }^{nr }$ d es cr i bea proce s sb y whi ch a pla smo nnonra diat i vel y excit ed in the NP by d ip ol e $k$ t ransfers itsen ergy to an ot her dipol e $j$. Ingeneral, due to high-$l $ plasmo nsinvol vedin nonrad iat ive co upl i ng, th e eige nstat es of $ \Gamm a _ {j k}^ {n r}$ are *d i f fer ent*fr om t hose of plasmon-enhancedr adi ative couplin g $ \Gam m a _{ jk} ^ {r } $ w hi c h c o n tains only dipo le ($l=1$) p l as mon mode.T her ef ore, th e eigen state s of ful l decay m atrix $\G am ma_{ j k }=\ Gamma_{jk} ^{nr}+\G amma_{jk} ^ {nr}$ ar e *no t*pure s up err adian t ands ubr adian t stat es but t heirad mixtures . However, the high-$l$ plasm on co ntr ibution t o $ \ Gam ma_{jk}^{ nr}$ is signif ica ntonlyat | has_just three_nonzero eigenvalues. Indeed, let_us introduce_new_decay matrices_as_$$\gamma_{\mu\nu}^{r}=\dfrac{N\Gamma^{r}}{3} B_{\mu\nu},$$ where_$$B_{\mu\nu}=\dfrac{3}{N}\sum_{j} e_{j}^{\mu}e_{j}^{\nu}$$ is_$3\times 3$ matrix in_coordinate space with_${\rm_Tr}\hat{B}=3$. It is easy to see that ${\rm Tr}\bigl [ (\hat{\Gamma}^{r} )^{n}\bigr]={\rm Tr}\bigl [_(\hat{\gamma}^{r}_)^{n}\bigr]$ for_any_integer_$n$, i.e., the $N\times N$_matrix $\Gamma_{jk}^{r}$ has only three_non-zero eigenvalues_coinciding with those of matrix $\gamma_{\mu\nu}^{r}$ $$\label{gamma-coop}
\Gamma_{\mu}^{r}=\dfrac{N\Gamma^{r}}{3}\,\lambda_{\mu},$$ where_$\lambda_{\mu}\sim_1$ are eigenvalues_of $B_{\mu\nu}$. Note that the decay rates of the_remaining $N-3$ subradiant states vanish in_the long wave_approximation_used_here; they acquire finite_values in the next order in_$(kr_{0})^{2}$.
Let us turn to non-radiative coupling,_described by matrix $\Gamma_{jk}^{nr}$. Its diagonal elements,_$\Gamma_{jj}^{nr}$, describe nonradiative energy exchange between_excited dipole and NP plasmon_modes with_all angular momenta, as indicated_by polarizabilities $\alpha''_{l}$_in Eq. (\[gamma-perp\])._The non-diagonal elements_of $\Gamma_{jk}^{nr}$ describe a process by_which a plasmon_nonradiatively excited in the NP by_dipole_$k$ transfers its_energy_to_another dipole_$j$. In general,_due_to high-$l$_plasmons_involved in nonradiative coupling, the eigenstates_of_$\Gamma_{jk}^{nr}$ are *different* from those of plasmon-enhanced_radiative coupling $\Gamma_{jk}^{r}$ which_contains_only dipole ($l=1$) plasmon_mode. Therefore, the eigenstates of_full decay matrix $\Gamma_{jk}=\Gamma_{jk}^{nr}+\Gamma_{jk}^{nr}$ are *not*_pure superradiant_and subradiant_states but their admixtures. However, the high-$l$ plasmon contribution to $\Gamma_{jk}^{nr}$_is significant only at |
X$ with $-K_{X}^{3} = 66$ and non-$\mathrm{cDV}$ singularities and reduce the proof of Theorem \[theorem:main\] to the case when $W$ is a conic bundle. Following [@Prokhorov-degree] we will exclude this case in Section \[section:surface\].\
I would like to thank I. Cheltsov, Yu. G. Prokhorov and C. Shramov for their numerous remarks and attention to the present paper which led to the correction of the gap in the older version. I also would like to express my gratitude to V. A. Iskovskikh for stimulating discussions.
Basic notions and statements {#section:basic}
============================
We use standard notions and facts from the Minimal Model Program (see for example [@Kollar-Mori]). Recall that algebraic variety $X$ is called a *Fano variety* if divisor $-K_{X}$ is an ample $\mathbb{Q}$-Cartier divisor. If divisor $-K_{X}$ is only nef and big, then $X$ is called a *weak Fano variety*.
In the present article we consider three-dimensional (weak) Fano varieties only with (no more than) canonical Gorenstein singularities.
Let $X$ be a Fano threefold. From the Riemann–Roch Formula and Kawamata–Vieweg Vanishing Theorem it is easy to find the dimension of the anti-canonical linear system on $X$ (see [@Reid-morphisms; @-Kawamata]): $$\dim |-K_{X}| = -\frac{1}{2}K_{X}^{3} + 2$$ The constant $-K_{X}^{3}$ is called the *degree* of $X$ and $g:= -\frac{1}{2}K_{X}^{3} + 1$ is the *genus* of $X$. Thus, $\dim |-K_{X}| = g + 1$ and $-K_{X}^{3} = 2g-2$. In particular, $-K_{X}^{3}$ is an even number and for $X$ satisfying the conditions of Theorem \[theorem:main\] we have $-K_{X}^{3} \in \{66,68,70\}$.
Furthermore, we have
\[theorem:elefant\] A general member $L \in |-K_{X | X$ with $ -K_{X}^{3 } = 66 $ and non-$\mathrm{cDV}$ singularities and reduce the proof of Theorem \[theorem: main\ ] to the case when $ W$ is a conic package. take after [ @Prokhorov - degree ] we will exclude this case in department \[section: surface\].\
I would like to thank I. Cheltsov, Yu. G. Prokhorov and C. Shramov for their numerous remarks and care to the present paper which led to the correction of the opening in the older version. I besides would like to express my gratitude to V. A. Iskovskikh for stimulating discussions.
Basic notion and statements { # section: basic }
= = = = = = = = = = = = = = = = = = = = = = = = = = = =
We use standard notions and facts from the Minimal Model Program (attend for example [ @Kollar - Mori ]). Recall that algebraic variety $ X$ is call a * Fano variety * if divisor $ -K_{X}$ is an ample $ \mathbb{Q}$-Cartier divisor. If divisor $ -K_{X}$ is only nef and big, then $ X$ is called a * watery Fano variety *.
In the present article we consider three - dimensional (weak) Fano varieties only with (no more than) basic Gorenstein singularities.
Let $ X$ be a Fano threefold. From the Riemann – Roch Formula and Kawamata – Vieweg Vanishing Theorem it is easy to find the dimension of the anti - canonical linear system on $ X$ (see [ @Reid - morphisms; @-Kawamata ] ): $ $ \dim |-K_{X}| = -\frac{1}{2}K_{X}^{3 } + 2$$ The constant $ -K_{X}^{3}$ is called the * degree * of $ X$ and $ g:= -\frac{1}{2}K_{X}^{3 } + 1 $ is the * genus * of $ X$. Thus, $ \dim |-K_{X}| = g + 1 $ and $ -K_{X}^{3 } = 2g-2$. In particular, $ -K_{X}^{3}$ is an even number and for $ X$ satisfying the condition of Theorem \[theorem: main\ ] we have $ -K_{X}^{3 } \in \{66,68,70\}$.
Furthermore, we have
\[theorem: elefant\ ] A cosmopolitan extremity $ L \in |-K_{X | X$ wlth $-K_{X}^{3} = 66$ and non-$\mathrm{cDY}$ singularities cbd rednce the proof ow Theorem \[theorem:main\] to the rase when $W$ is a conic bundle. Fullowing [@Irokhorov-eegrte] we will excludx this case in Dectnoi \[section:surface\].\
L would lika to thank I. Cveutdov, Yu. G. Prokhorov and C. Shramov fow their nkmerous remarkf anc attsntion to the present paper which med to uhe correction of yhe gap in the older versiln. I also would like tl express mi grwritude to V. X. Iskovskikh for stimumating discussions.
Basic notions and xtatements {#swctlmn:basic}
============================
We uwe stwndard notiokx and xacts ftom the Minimak Mmdeo Program (see for exakple [@Kollar-Mori]). Resall that amgebraic variety $Z$ us canled a *Fxbo xarjevy* jf divlsoc $-K_{X}$ is an zmple $\mathbv{Q}$-Cartier divisor. Ig qpfisor $-K_{X}$ is knly nqf and big, then $X$ is called a *weak Fano vdristy*.
In the present articoe we consider three-dlmensionaj (weak) Fano varieties only with (no more than) canotical Eortnwtein wijgularities.
Let $X$ be a Fano threefold. From the Ditmakn–Roch Formula akd Kawamata–Vieweg Fajixring Theorem kt is zzsg to find the dimejsion os the anti-canogicak linear system on $X$ (see [@Reud-morphisms; @-Hqwamata]): $$\dim |-K_{X}| = -\frcc{1}{2}K_{X}^{3} + 2$$ The cunstsnt $-K_{C}^{3}$ is called the *degree* uf $X$ and $g:= -\frac{1}{2}N_{X}^{3} + 1$ is tgd *genus* of $X$. Thur, $\dpm |-K_{F}| = g + 1$ and $-K_{X}^{3} = 2g-2$. In partictlar, $-K_{X}^{3}$ iw an even njmbet and fjr $X$ satisvying the conditions of Hheorgm \[theosem:main\] we have $-K_{X}^{3} \in \{66,68,70\}$.
Furthermore, we have
\[vieorem:elefant\] A geteral meiber $K \in |-K_{X | X$ with $-K_{X}^{3} = 66$ and non-$\mathrm{cDV}$ reduce proof of \[theorem:main\] to the conic Following [@Prokhorov-degree] we exclude this case Section \[section:surface\].\ I would like to I. Cheltsov, Yu. G. Prokhorov and C. Shramov for their numerous remarks and to the present paper which led to the correction of the gap in older I would to express my gratitude to V. A. Iskovskikh for stimulating discussions. Basic notions and statements {#section:basic} We use standard notions and facts from the Model Program (see for [@Kollar-Mori]). Recall that algebraic variety is a *Fano if $-K_{X}$ an ample $\mathbb{Q}$-Cartier If divisor $-K_{X}$ is only nef and big, then $X$ is called a *weak Fano variety*. In present article three-dimensional (weak) varieties with more than) canonical Let $X$ be a Fano threefold. Formula and Kawamata–Vieweg Vanishing Theorem it is easy find the of the anti-canonical linear system on (see [@Reid-morphisms; @-Kawamata]): $$\dim |-K_{X}| = -\frac{1}{2}K_{X}^{3} + The constant $-K_{X}^{3}$ is called the *degree* of $X$ and $g:= -\frac{1}{2}K_{X}^{3} + 1$ is of $X$. Thus, $\dim = g + and = In $-K_{X}^{3}$ is even number and for $X$ satisfying the conditions of Theorem \[theorem:main\] have $-K_{X}^{3} \in \{66,68,70\}$. Furthermore, we have \[theorem:elefant\] A general \in | X$ with $-K_{X}^{3} = 66$ and non-$\mathrm{cDV}$ sinGularities And reDucE thE pRoof Of ThEorem \[theorem:maIN\] to tHe case when $W$ is a conic bunDle. FoLlOWing [@pRoKhoroV-degree] WE wILL exClUdE thIs CAsE in SeCtiOn \[sectiOn:surface\].\
I WouLd Like to thank I. cHeLtsov, Yu. G. PrOkhOrov and C. ShraMov For theIr NumERous rEmaRks anD attenTIon to tHe present PaPEr whicH Led to thE COrRectIon of the gap in the oLDeR Version. I also woUld likE tO ExPREss My gRatitude to v. A. iskovSKikh for STiMULAtiNG discussions.
BAsic notions ANd sTatemeNtS {#seCTion:baSic}
============================
We UsE StaNdard notionS and Facts from The MinIMal ModeL program (See for ExaMplE [@KolLAr-moRi]). REcALl tHAt AlgEBraIc varietY $X$ Is CalleD a *FaNO VARietY* if DiviSor $-K_{X}$ Is an ample $\mathBb{Q}$-cartIEr dIvisoR. If diVisoR $-K_{x}$ is onLy nef aNd big, ThEn $X$ is called a *weaK FanO variety*.
IN thE pResEnT artiCLe we coNsiDer Three-diMensionAL (weAk) fANO vArieties only with (no MoRE ThAn) canoniCal GorENsTeIN singulaRiTieS.
Let $x$ BE a FanO thrEEfOld. From tHe RiemANn–roCh FormuLa And KawAmAta–vieWeg VaNIshiNg TheoRem it is eAsy to FInd the dimensioN Of the anti-canoNIcAL LiNEar sYstEm on $X$ (see [@ReiD-morPHismS; @-KawAMaTa]): $$\dIM |-K_{X}| = -\frAc{1}{2}K_{X}^{3} + 2$$ THe COnSTant $-K_{X}^{3}$ is called the *deGrEe* of $X$ aNd $g:= -\frAc{1}{2}K_{X}^{3} + 1$ is the *genuS* of $X$. Thus, $\diM |-k_{x}| = G + 1$ and $-K_{X}^{3} = 2g-2$. IN parTIcULar, $-K_{X}^{3}$ is an even nUmber And for $X$ satISfying thE condItions of theorem \[thEORem:main\] wE haVe $-K_{x}^{3} \in \{66,68,70\}$.
furTHErMore, we have
\[theOREm:elEfAnt\] A genEraL member $l \in |-k_{X | X$ with $-K_{X}^{3} = 66$and non-$\ mathr m{c DV} $sing ular ities and redu c e th e proof of Theorem \[t heore m: m ain\ ] t o the case w h en $ W$is a co ni c b undle . F ollowin g [@Prokho rov -d egree] we wi l lexclude th iscase in Sect ion \[sec ti on: s urfac e\] .\
Iwouldl ike to thank I. C h eltsov , Yu. G. P ro khor ov and C. Shramov fo r their numerou s rema rk s a n d at ten tion to th eprese n t paper wh i c h le d to the corre ction of th e ga p in t he ol d er ver sion. I als o would lik e to expressmy gra t itude t o V. A.Iskovs kik h f or s t im ul ati ng dis c us sio n s.
Basic n ot io ns an d st a t e m ents {# sect ion:b asic}
======= === ==== = === ===== =====
We u se st andard noti on s and facts fro m th e Minimal Mo de l P ro gram( see fo r e xam ple [@K ollar-M o ri] ). R e ca ll that algebraicva r i et y $X$ is calle d a * F ano vari et y*if d i v isor$-K_ { X} $ is anample$ \m at hbb{Q}$ -C artier d ivi sor . Ifd ivis or $-K _{X}$ is only nef and big, t h en $X$ is cal l ed a * w eakFan o variety*.
In thepres e nt ar t iclewe co ns i de r three-dimensional(w eak) F ano v arieties only with (nom o r e than)cano n ic a l Gorenstein s ingul arities.
L et $X$ b e a F ano thre efold. Fr o m the Rie man n–R och Fo r m ul a and Kawamat a – View eg Vanish ing Theore m i t i s e asy t o find th e dimens io nof t heanti- c anonical l ine ar sy stemo n $X$(see[@Re id -m o rph isms; @ - Ka w a mata ]) :$$\d im|- K_{X} | =- \fr ac{1}{2 }K_{X}^{3 } + 2$$Th econstan t $-K_{X}^{3} $is calledth e * degree * of $X$ a nd $g:= -\frac{1}{2}K_{ X }^{3} + 1$ is t he * genus* of $X $. Thu s,$ \dim | -K_{X} | = g + 1$ a nd $- K _ {X }^{ 3} = 2g-2$.I n pa rticu la r, $ -K_{X}^ {3}$ is an even nu m ber and for $X$sat isfy i n gthe co n dit io n s o f Theorem \[theor em:main\]we ha ve $-K_{X} ^ {3} \ in \{66 ,68,70\ }$.
F urtherm ore, we h ave
\[th eo rem: e l efa nt\] A gen eral mem ber $L \i n |-K_ { X | X$ with_$-K_{X}^{3} =_66$ and non-$\mathrm{cDV}$ singularities_and reduce_the_proof of_Theorem \[theorem:main\]_to the case_when $W$ is_a conic bundle. Following_[@Prokhorov-degree] we will_exclude_this case in Section \[section:surface\].\
I would like to thank I. Cheltsov, Yu. G. Prokhorov and_C._Shramov for_their_numerous_remarks and attention to the_present paper which led to_the correction_of the gap in the older version. I_also_would like to_express my gratitude to V. A. Iskovskikh for stimulating_discussions.
Basic notions and statements {#section:basic}
============================
We use_standard notions and_facts_from_the Minimal Model Program_(see for example [@Kollar-Mori]). Recall that_algebraic variety $X$ is called a_*Fano variety* if divisor $-K_{X}$ is an_ample $\mathbb{Q}$-Cartier divisor. If divisor $-K_{X}$_is only nef and big,_then $X$_is called a *weak Fano_variety*.
In the present_article we_consider three-dimensional (weak)_Fano varieties only with (no more_than) canonical Gorenstein_singularities.
Let $X$ be a Fano threefold._From_the Riemann–Roch Formula_and_Kawamata–Vieweg_Vanishing Theorem_it is easy_to_find the_dimension_of the anti-canonical linear system on_$X$_(see [@Reid-morphisms; @-Kawamata]): $$\dim |-K_{X}| = -\frac{1}{2}K_{X}^{3}_+ 2$$ The constant_$-K_{X}^{3}$_is called the *degree*_of $X$ and $g:= -\frac{1}{2}K_{X}^{3}_+ 1$ is the *genus* of_$X$. Thus,_$\dim |-K_{X}|_= g + 1$ and $-K_{X}^{3} = 2g-2$. In particular, $-K_{X}^{3}$_is an even number and for_$X$ satisfying the conditions_of Theorem \[theorem:main\]_we_have $-K_{X}^{3} \in_\{66,68,70\}$.
Furthermore,_we have
\[theorem:elefant\]_A general member $L \in |-K_{X |
similar temperature dependencies[@PhysRevB.90.161302; @APL.105.101901; @ScientificRep.6.22414; @NanoLett.16.5010; @NatureComm.6.896; @PhysRevLett.119.137401; @PhysRevB.95.235408].
The mechanisms which explain both the formation of the free carrier valley polarization and the long-lived polarization in the gap region are depicted in Fig. 2. We start with the Fermi level tuned into the conduction band ($V_\text{BG}>\unit{10}{V}$) at equilibrium conditions, i.e. an equal number of free charge carriers in both valleys (indicated by the shaded areas in the band structure). A circularly polarized laser pump pulse now excites valley-selectively electron-hole pairs (circles in Fig. 2a)[@RevModPhys.90.021001]. By the interaction with the free conduction band carriers, the electron-hole pairs first create bright trions which, however, cannot directly be responsible for the TRKR signals in the ns range as their recombination times are typically in the ps range[@RevModPhys.90.021001; @PhysRevB.93.205423; @PhysRevLett.123.067401]. The recombination of the photo-exited electron-hole pairs within the same valley can also not lead to a net valley polarization of resident charge carriers. In contrast, a net valley polarization can be formed only if one charge of the electron-hole pair scatters into the other valley. In case of WSe$_2$, this can easily happen by the intervalley transition of the electron into the energetically lower conduction band of the same spin-orientation in the K’-valley under the emission of an optical phonon (Fig. 2b)[@RevModPhys.90.021001; @PhysicalReviewMaterials.2.014002]. The subsequent recombination of the photo-excited holes with free electrons of the K-valley on the time-scale $\tau_\text{rec}^\text{dark}$ (this process has a finite possibility as explained in the SI) will lead to a situation where the photo-exited electrons have increased the electron number in the K’-valley. At the same time the K-valley is missing the same number of free electrons which recombined with the photo-excited holes (see Figs. 2c and 2d). This | similar temperature dependencies[@PhysRevB.90.161302; @APL.105.101901; @ScientificRep.6.22414; @NanoLett.16.5010; @NatureComm.6.896; @PhysRevLett.119.137401; @PhysRevB.95.235408 ].
The mechanisms which explain both the formation of the spare mailman valley polarization and the long - lived polarization in the gap region are describe in Fig. 2. We start with the Fermi level tune into the conduction ring ($ V_\text{BG}>\unit{10}{V}$) at equilibrium condition, i.e. an equal number of barren charge carriers in both valley (indicated by the shaded areas in the band structure). A circularly polarized laser pump pulse nowadays excites valley - selectively electron - hole pairs (lap in Fig. 2a)[@RevModPhys.90.021001 ]. By the interaction with the free conduction band carriers, the electron - hole pair first create bright trions which, however, cannot directly be responsible for the TRKR signal in the ns range as their recombination times are typically in the ps range[@RevModPhys.90.021001; @PhysRevB.93.205423; @PhysRevLett.123.067401 ]. The recombination of the photo - exited electron - hole pairs within the same valley can also not lead to a net valley polarization of resident charge carriers. In contrast, a net valley polarization can be formed only if one care of the electron - hole pair scatter into the early valley. In casing of WSe$_2 $, this can easily happen by the intervalley passage of the electron into the energetically lower conduction band of the same spin - predilection in the K’-valley under the emission of an optical phonon (Fig. 2b)[@RevModPhys.90.021001; @PhysicalReviewMaterials.2.014002 ]. The subsequent recombination of the photo - excited holes with free electron of the K - valley on the time - scale $ \tau_\text{rec}^\text{dark}$ (this process has a finite possibility as explained in the SI) will lead to a situation where the photograph - die electrons have increased the electron number in the K’-valley. At the same meter the K - valley is missing the same number of spare electrons which recombined with the photo - excited hole (see Figs. 2c and 2d). This | sililar temperature dependtncies[@PhysRevB.90.161302; @APL.105.101901; @ScienvificRel.6.22414; @NanoLegt.16.5010; @NatureComm.6.896; @PhysRevLett.119.137401; @PhbsRecB.95.235408].
The mechanisms which explxin both nhe formarion if the frex carriev valmcy ponerization and tme long-liveg polarization iv che gap region are depicted in Fig. 2. Wq start wlth the Fermi jevek tunsd into the conduction band ($V_\text{BF}>\unit{10}{V}$) et equilibrium vonditions, i.e. an equal numher lf free charge cargiers in bojg vwoleys (indicaged by the shaded areaa in the band structure). A circuuarly polarized lqseg pump pulse now vxcites vallen-xelecthvely ekectron-hole palrs (cmrclws in Fig. 2a)[@RevModPhys.90.021001]. Uy the interaction wyth the fsez conduction band careiwrs, tve enectfin-hule peira firsh cceate brighf trions whuch, however, cannot ciwvvtly be respknsiblq sor the TRKR signals in the ns range as thsir recombination times are typically in the ps range[@WevModPhys.90.021001; @PhysRevB.93.205423; @PhysRevLett.123.067401]. The recombination of tie phitj-dzihed electron-hole pairs within the same valley dam slso not lead bo a net valley pokagiawtion of resiaent ckzrfe carriers. In conhrast, a net calley pojariaation can be formed only id one charge if the electron-holz pair scattzrs injo the other valley. In case oy WSe$_2$, fhis can eadily happsv by the intervaulej trdnsition of the electron igto the eiergecically uowet condustion band of tma same spin-orientahion nn tha K’-valley knder the emission of an optical phonon (Fig. 2b)[@RgvMmdPvys.90.021001; @PhysncalReyiewMaterials.2.014002]. Tre subsequent tecombinacion ow the photk-excitev holes with free electrotd of the K-vanley on ehe rime-wcale $\txj_\text{rec}^\text{datk}$ (this pgoeess has q finite possibiliby as sxplained in thz WI) will lead to a rittanioi whewa the photo-efitea eurctrovs have incveared yhe electron number hn tge K’-valley. At the xaie time jhe K-vallqy is missing the same number ov frex elecvrons ehish recombined with the photo-exdited holfs (fee Figs. 2c anq 2d). Bhis | similar temperature dependencies[@PhysRevB.90.161302; @APL.105.101901; @ScientificRep.6.22414; @NanoLett.16.5010; @NatureComm.6.896; The which explain the formation of and long-lived polarization in gap region are in Fig. 2. We start with Fermi level tuned into the conduction band ($V_\text{BG}>\unit{10}{V}$) at equilibrium conditions, i.e. an number of free charge carriers in both valleys (indicated by the shaded areas the structure). circularly laser pump pulse now excites valley-selectively electron-hole pairs (circles in Fig. 2a)[@RevModPhys.90.021001]. By the interaction with free conduction band carriers, the electron-hole pairs first bright trions which, however, directly be responsible for the signals the ns as recombination are typically in ps range[@RevModPhys.90.021001; @PhysRevB.93.205423; @PhysRevLett.123.067401]. The recombination of the photo-exited electron-hole pairs within the same valley can also lead to valley polarization resident carriers. contrast, a net can be formed only if one electron-hole pair scatters into the other valley. In of WSe$_2$, can easily happen by the intervalley of the electron into the energetically lower conduction of the same spin-orientation in the K’-valley under the emission of an optical phonon (Fig. The subsequent recombination of photo-excited holes with electrons the on time-scale $\tau_\text{rec}^\text{dark}$ process has a finite possibility as explained in the SI) will to a situation where the photo-exited electrons have increased the in K’-valley. At the time the K-valley is the number of free electrons with photo-excited 2c 2d). | similar temperature dependeNcies[@PhysREvB.90.161302; @APl.105.101901; @ScIenTiFicREp.6.22414; @NaNoLett.16.5010; @NatureCoMM.6.896; @PhySRevLett.119.137401; @PhysRevB.95.235408].
The mecHanisMs WHich EXpLain bOth the fORmATIon Of ThE frEe CArRier vAllEy polarIzation and The LoNg-lived polarIZaTion in the gAp rEgion are depiCteD in Fig. 2. we StaRT with The fermi Level tUNed intO the conduCtIOn band ($v_\Text{BG}>\uNIT{10}{V}$) At eqUilibrium conditioNS, i.E. An equal number oF free cHaRGe CARriErs In both vallEyS (indiCAted by tHE sHADEd aREas in the band sTructure). A ciRCulArly poLaRizED laser Pump pUlSE noW excites valLey-sElectivelY electROn-hole pAIrs (circLes in FIg. 2a)[@revmodPHYs.90.021001]. by The InTEraCTiOn wITh tHe free coNdUcTion bAnd cARRIErs, tHe eLectRon-hoLe pairs first cReaTe brIGht TrionS whicH, howEvEr, canNot dirEctly Be Responsible for tHe TRkR signals In tHe Ns rAnGe as tHEir recOmbInaTion timEs are tyPIcaLlY IN ThE ps range[@RevModPhys.90.021001; @phYSreVB.93.205423; @PhysReVLett.123.067401]. THE rEcOMbinatioN oF thE phoTO-ExiteD eleCTrOn-hole paIrs witHIn ThE same vaLlEy can aLsO noT leAd to a NEt vaLley poLarizatiOn of rESident charge caRRiers. In contraST, a NET vALley PolArization caN be fORmed Only IF oNe cHArge oF the eLeCTrON-hole pair scatters inTo The othEr valLey. In case of WSE$_2$, this can eaSILY happen bY the INtERvalley transitIon of The electroN Into the eNergeTically lOwer conduCTIon band oF thE saMe sPin-ORIeNtation in the K’-VALley UnDer the eMisSion of aN opTicAl pHonOn (fig. 2b)[@RevMoDPhys.90.021001; @PhySiCaLREvIewmaterIAls.2.014002]. The suBsEquEnT reCombiNAtion oF the pHoto-ExCiTEd hOles witH FrEE ElecTrOnS of tHe K-VaLley oN the TIme-Scale $\taU_\text{rec}^\tExt{DArk}$ (tHiS pRocess hAs a finite possIbIlity as expLaIneD in the si) Will lead To a situation where the phoTO-exited EleCtronS havE increaseD thE electRon NUmber iN the K’-vAlley. at The SAMe timE THe k-vaLlEy is missinG THe sAme nuMbEr of Free eleCtrons which recombiNEd wIth the photo-exCitEd hoLES (sEe FIGs. 2C And 2D). THIs | similar temperature depen dencies[@P hysRe vB. 90. 16 1302 ; @A PL.105.101901; @Sci entificRep.6.22414; @N anoLe tt . 16.5 0 10 ; @Na tureCom m .6 . 8 96; @ Ph ysR ev L et t.119 .13 7401; @ PhysRevB.9 5.2 35 408].
The m e ch anisms whi chexplain both th e form at ion of th e f ree c arrier valley polariza ti o n andt he long - l iv ed p olarization in th e g a p region are d epicte di nF i g.2.We start w it h the Fermi l e ve l t une d into the con duction ban d ($ V_\tex t{ BG} > \unit{ 10}{V }$ ) at equilibriu m co nditions, i.e.a n equal numberof fre e c har ge c a rr ie rsin bot h v all e ys(indicat ed b y the sha d e d area s i n th e ban d structure). Acirc u lar ly po lariz ed l as er pu mp pul se no wexcites valley- sele ctively e lec tr on- ho le pa i rs (ci rcl esin Fig. 2a)[@R e vMo dP h y s .9 0.021001]. By thein t e ra ction wi th the fr ee conducti on ba nd c a r riers , th e e lectron- hole p a ir sfirst c re ate br ig httri ons w h ich, howev er, cann ot di r ectly be respo n sible for the TR K R s i gnal s i n the ns ra ngea s th eirr ec omb i natio n tim es ar e typically in the p srange[ @RevM odPhys.90.021 001; @Phys R e v B.93.205 423; @P h ysRevLett.123. 06740 1]. The re c ombinati on of the pho to-exited e lectron- hol e p air s w i t hi n the same va l l ey c an also n otlead to anet va lle ypolarizat ion of r es id en tcha rge c a rriers.In co nt ras t, an et val ley p olar iz at i oncan bef or m e d on ly i f on e c ha rge o f th e el ectron- hole pair sc a tter sin to theother valley. I n case ofWS e$_ 2$, th i s can eas ily happen by the inter v alley t ran sitio n of the elec tro n into th e energ etical ly lo we r c o n ducti o n b and o f the same s pin -orie nt atio n in th e K’-valley undert heemission of a n o ptic a l p hon o n( Fig .2 b)[ @ R evModPhys.90.02 1001; @Phy si c al ReviewMate r ial s. 2.01400 2]. The subs e quent r ecombinat ion of th ephot o - exc ited holes with fr ee electr o ns of th e K-v all ey onth e t ime-s cale $ \ tau _\tex t{rec} ^\ text{d ark}$ ( this pro cess has a finite possi bility as e xpl ained inthe SI) will lea d to a situati onwhe re th e p h oto-e xite d e lec t ronshave increased th e e l e ct ron numberi n the K’-v all e y. Atthesame time the K-v a lley is missin g th e sam e n u mber o f free electro nswh i c h recomb in ed with the photo-e xc i ted h oles ( see Fi gs. 2ca n d2 d). Th is | similar_temperature dependencies[@PhysRevB.90.161302;[email protected]; @ScientificRep.6.22414; @NanoLett.16.5010; @NatureComm.6.896;[email protected]; @PhysRevB.95.235408].
The_mechanisms_which explain_both_the formation of_the free carrier_valley polarization and the_long-lived polarization in_the_gap region are depicted in Fig. 2. We start with the Fermi level tuned into_the_conduction band_($V_\text{BG}>\unit{10}{V}$)_at_equilibrium conditions, i.e. an equal_number of free charge carriers_in both_valleys (indicated by the shaded areas in the_band_structure). A circularly_polarized laser pump pulse now excites valley-selectively electron-hole pairs_(circles in Fig. 2a)[@RevModPhys.90.021001]. By the interaction_with the free_conduction_band_carriers, the electron-hole pairs_first create bright trions which, however,_cannot directly be responsible for the_TRKR signals in the ns range as_their recombination times are typically in_the ps range[@RevModPhys.90.021001; @PhysRevB.93.205423; @PhysRevLett.123.067401]._The recombination_of the photo-exited electron-hole pairs_within the same_valley can_also not lead_to a net valley polarization of_resident charge carriers._In contrast, a net valley polarization_can_be formed only_if_one_charge of_the electron-hole pair_scatters_into the_other_valley. In case of WSe$_2$, this_can_easily happen by the intervalley transition of_the electron into the_energetically_lower conduction band of_the same spin-orientation in the_K’-valley under the emission of an_optical phonon_(Fig. 2b)[@RevModPhys.90.021001; @PhysicalReviewMaterials.2.014002]._The subsequent recombination of the photo-excited holes with free electrons of_the K-valley on the time-scale $\tau_\text{rec}^\text{dark}$_(this process has a_finite possibility_as_explained in the_SI)_will lead_to a situation where the photo-exited electrons_have increased_the electron number in the K’-valley._At the same time_the_K-valley is missing the same number_of free electrons which recombined with_the photo-excited holes (see Figs. 2c_and_2d)._This |
aziz University, P.O.Box: 80203 Jeddah 21589, Saudi Arabia'
- 'Siedlce University of Natural Sciences and Humanities, 3 Maja 54, 08-110 Siedlce, Poland'
author:
- 'K. Keremedis, C. Özel, A. Pikosz, M. A. Al Shumrani, E. Wajch'
title: 'Compact complement topologies and k-spaces'
---
[^1]
Introduction
============
The compact complement topology of the real line was considered, for instance, in Example 22 of the celebrated book by Steen and Seebach“Counterexamples in Topology” ([@SS]). We investigate this notion in a much wider context of Hausdorff spaces and of partially topological spaces that belong to the class of generalized topological spaces in the sense of Delfs-Knebusch (cf. [@DK] and [@Pie1]). Our results are proved in $\mathbf{ZF}$ if this is not otherwise stated. All axioms of $\mathbf{ZF}$ can be found in [@Ku].
In section 2, we give elementary properties of the compact complement topology of a Hausdorff space. In particular, we show that if a Hausdorff space is locally compact and second countable, then its compact complement topology is second countable, while the compact complement topology of a non-locally compact metrizable space need not be first countable. We give an example of a countable metrizable space whose compact complement topology is not first countable. In section 3, a necessary and sufficient condition for a set to be compact with respect to the compact complement topology of a given Hausdorff space leads us to a new characterization of k-spaces. A well-known theorem of $\mathbf{ZFC}$ states that all first countable Hausdorff spaces are $k$-spaces (cf. Theorem 3.3.20 of [@E]). We show that this theorem may fail in $\mathbf{ZF}$. More precisely, we prove that, if $\mathcal{M}$ is a model of $\mathbf{ZF}$, then all Hausdorff first countable spaces of $\mathcal{M}$ are k-spaces if and only if all metrizable spaces in $\mathcal{M}$ are $k$-spaces which holds if and only if the axiom of countable multiple choice (Form 126 in [@HR]) is true in $\ | aziz University, P.O.Box: 80203 Jeddah 21589, Saudi Arabia'
-' Siedlce University of Natural Sciences and Humanities, 3 Maja 54, 08 - 110 Siedlce, Poland'
author:
-' K. Keremedis, C. Özel, A. Pikosz, M. A. Al Shumrani, E. Wajch'
title:' Compact complement topologies and thousand - space'
---
[ ^1 ]
Introduction
= = = = = = = = = = = =
The compact complement regional anatomy of the real occupation was considered, for instance, in Example 22 of the observe book by Steen and Seebach“Counterexamples in Topology ” ([ @SS ]). We investigate this notion in a a lot wide-eyed context of Hausdorff spaces and of partially topological space that belong to the class of generalized topological spaces in the sense of Delfs - Knebusch (cf. [ @DK ] and [ @Pie1 ]). Our results are prove in $ \mathbf{ZF}$ if this is not otherwise stated. All axioms of $ \mathbf{ZF}$ can be receive in [ @Ku ].
In section 2, we give elementary properties of the compendious complement topology of a Hausdorff distance. In particular, we show that if a Hausdorff space is locally compendious and second countable, then its compact complement topology is second countable, while the compact complement topology of a non - locally compact metrizable space need not be first countable. We give an example of a countable metrizable space whose compact complement topology is not first countable. In section 3, a necessary and sufficient condition for a set to be compact with esteem to the compact complement topology of a feed Hausdorff quad leads us to a new characterization of potassium - space. A well - known theorem of $ \mathbf{ZFC}$ states that all inaugural countable Hausdorff spaces are $ k$-spaces (cf. Theorem 3.3.20 of [ @E ]). We show that this theorem may fail in $ \mathbf{ZF}$. More precisely, we prove that, if $ \mathcal{M}$ is a model of $ \mathbf{ZF}$, then all Hausdorff first countable space of $ \mathcal{M}$ are k - spaces if and only if all metrizable spaces in $ \mathcal{M}$ are $ k$-spaces which holds if and only if the maxim of countable multiple option (mannequin 126 in [ @HR ]) is true in $ \ | azix University, P.O.Box: 80203 Jeddxh 21589, Saudi Arabic'
- 'Siedlre Univsrsity ow Natural Sciences and Humanmtiew, 3 Maha 54, 08-110 Siedlce, Poland'
autfor:
- 'K. Kervmedis, C. Özel, E. Pikosz, M. A. Al Shumrani, E. Wajdm'
titlz: 'Rompact complemgnt topologias and k-spaces'
---
[^1]
Hngrlduction
============
The compact complement topojogy of tje real line wws cpgsidsgeb, for instance, in Example 22 of tge celeurated book by Xteen and Seebach“Counterexwmplfs in Topology” ([@SS]). Ae investigqte eyis notion iv a much wpber context of Hausdorff spaces and of pargiallv topologicql spwwes that beoong no the class of generdlized yopological spscev ib the sense of Delfs-Kiebusch (cf. [@DK] and [@Pig1]). Our resunta are proved in $\mqtybf{ZF}$ if dhis us vot ovhedwise dtaved. All axikms of $\mathvf{ZF}$ can be found im [@Hl].
On section 2, se givq qlementary properties of the compact cokplsment topology of a Hauwdorff space. In partifular, we fhow that if a Hausdorff space is locally compact and adcoue couvrahle, then its compact complement topology is sqdomd countable, whije the compsch vjmplement topulogy of z non-locally compaft metryzablw space nted npt be first countable. We gice an examplv of a countable metrivable space chose vompavt complement topology ns not first counhable. In adction 3, a necessxry atd sufficient condition fow a set ti be compacg wijh respqct to the compact complement topopogy lf a given Hwusdorff space leads us to a nex characterizsthon of k-spaees. A eell-known thejrem of $\mathbf{EFC}$ statef thag all firsn countabne Hausdorfs spaces are $n$-dpaces (cf. Thxorem 3.3.20 of [@E]). Ww shiw that ghis theorem msy fail iu $\matybf{ZF}$. More preciseky, ds prove that, if $\mauhxal{M}$ is a modek ow $\mwtjbh{ZF}$, tran all Hausdmrff fifxt cojntable spages of $\kathcal{M}$ are k-spaces if znd only if all meyrlzable spqces in $\iathcal{M}$ are $l$-spaces which holdd if end onuy if thg axiom of countable multiple cgoice (Forl 126 ln [@HR]) is true in $\ | aziz University, P.O.Box: 80203 Jeddah 21589, Saudi 'Siedlce of Natural and Humanities, 3 author: 'K. Keremedis, C. A. Pikosz, M. Al Shumrani, E. Wajch' title: 'Compact topologies and k-spaces' --- [^1] Introduction ============ The compact complement topology of the line was considered, for instance, in Example 22 of the celebrated book by and in ([@SS]). investigate this notion in a much wider context of Hausdorff spaces and of partially topological spaces belong to the class of generalized topological spaces the sense of Delfs-Knebusch [@DK] and [@Pie1]). Our results proved $\mathbf{ZF}$ if is otherwise All axioms of can be found in [@Ku]. In section 2, we give elementary properties of the compact complement topology a Hausdorff particular, we that a space is locally second countable, then its compact complement countable, while the compact complement topology of a compact metrizable need not be first countable. We an example of a countable metrizable space whose complement topology is not first countable. In section 3, a necessary and sufficient condition for to be compact with to the compact topology a Hausdorff leads us a new characterization of k-spaces. A well-known theorem of $\mathbf{ZFC}$ states all first countable Hausdorff spaces are $k$-spaces (cf. Theorem 3.3.20 We that this theorem fail in $\mathbf{ZF}$. More we that, if $\mathcal{M}$ is of then countable of are k-spaces if and if all metrizable spaces in are $k$-spaces which holds axiom of countable multiple choice (Form 126 in is true in $\ | aziz University, P.O.Box: 80203 Jeddah 21589, saudi ArabiA'
- 'SiedLce uniVeRsitY of NAtural Sciences ANd HuManities, 3 Maja 54, 08-110 Siedlce, PolAnd'
auThOR:
- 'K. KeREmEdis, C. özel, A. PiKOsZ, m. a. Al shUmRanI, E. wAjCh'
titLe: 'COmpact cOmplement tOpoLoGies and k-spacES'
---
[^1]
INtroductioN
============
ThE compact compLemEnt topOlOgy OF the rEal Line wAs consIDered, fOr instancE, iN examplE 22 Of the ceLEBrAted Book by Steen and SeeBAcH“counterexampleS in TopOlOGy” ([@ss]). we iNveStigate thiS nOtion IN a much wIDeR CONteXT of Hausdorff sPaces and of pARtiAlly toPoLogICal spaCes thAt BEloNg to the clasS of gEneralizeD topolOGical spACes in thE sense Of DElfS-KneBUsCh (Cf. [@Dk] aND [@PiE1]). ouR reSUltS are provEd In $\MathbF{ZF}$ iF THIS is nOt oTherWise sTated. All axiomS of $\MathBF{ZF}$ Can be Found In [@Ku].
in SectiOn 2, we giVe eleMeNtary properties Of thE compact cOmpLeMenT tOpoloGY of a HaUsdOrfF space. IN particULar, We SHOW tHat if a Hausdorff spaCe IS LoCally comPact anD SeCoND countabLe, TheN its COMpact CompLEmEnt topolOgy is sECoNd CountabLe, While tHe ComPacT compLEmenT topolOgy of a noN-locaLLy compact metriZAble space need NOt BE FiRSt coUntAble. We give aN exaMPle oF a coUNtAblE MetriZable SpACe WHose compact complemeNt TopoloGy is nOt first countaBle. In sectiON 3, A NecessarY and SUfFIcient conditioN for a Set to be comPAct with rEspecT to the coMpact compLEMent topoLogY of A giVen hAUsDorff space leaDS Us to A nEw charaCteRizatioN of K-spAceS. A wElL-known theOrem of $\maThBf{zFc}$ sTatEs thaT All first CoUntAbLe HAusdoRFf spacEs are $K$-spaCeS (cF. theOrem 3.3.20 of [@E]). wE sHOW thaT tHiS theOreM mAy faiL in $\mAThbF{ZF}$. More Precisely, We pROve tHaT, iF $\mathcaL{M}$ is a model of $\mAtHbf{ZF}$, then aLl hauSdorff FIRst countAble spaces of $\mathcal{M}$ are K-Spaces iF anD only If alL metrizabLe sPaces iN $\maTHcal{M}$ aRe $k$-spaCes whIcH hoLDS if anD ONlY if ThE axiom of coUNTabLe mulTiPle cHoice (FoRm 126 in [@HR]) is true in $\ | aziz University, P.O.Box:80203 Jedd ah 21 589 , S au di A rabi a'
- 'SiedlceU nive rsity of Natural Scien ces a nd Huma n it ies,3 Maja5 4, 0 8-1 10 S ied lc e ,Polan d'author:
- 'K. Ker eme di s, C. Özel,A .Pikosz, M. A. Al Shumrani , E . Wajc h'
ti t le: ' Com pactcomple m ent to pologiesan d k-spa c es'
--- [^ 1]
Introduction
==== = == = ====
The comp act co mp l em e n t t opo logy of th ereall ine was co n s i der e d, for instan ce, in Exam p le22 ofth e c e lebrat ed bo ok bySteen and S eeba ch“Counte rexamp l es in T o pology” ([@SS ]). We inv e st ig ate t h isn ot ion ina much w id er cont exto f H ausd orf f sp acesand of partia lly top o log icalspace s th at belo ng tothe c la ss of generaliz ed t opologica l s pa ces i n the senseofDel fs-Kneb usch (c f . [ @D K ] an d [@Pie1]). Our re su l t sare prov ed in$ \m at h bf{ZF}$if th is i s not o ther w is e stated . Alla xi om s of $\ ma thbf{Z F} $ c anbe fo u nd i n [@Ku ].
In s ectio n 2, we give el e mentary prope r ti e s o f the co mpact compl emen t top olog y o f a Hausd orffsp a ce . In particular, wesh ow tha t ifa Hausdorff s pace is lo c a l ly compa ct a n ds econd countabl e, th en its com p act comp lemen t topolo gy is sec o n d counta ble , w hil e t h e c ompact comple m e nt t op ology o f a non-lo cal lycom pac tmetrizabl e spacene ed n ot be firs t countab le . W egiv e ane xample of a cou nt ab l e m etrizab l es p acewh os e co mpa ct comp leme n t t opology is not f irs t cou nt ab le. Insection 3, ane cessary an dsuf ficien t conditio n for a set to be compa c t withres pectto t he compac t c omplem ent topolo gy ofa giv en Ha u s dorff s pa cele ads us toa new char ac teri zationof k-spaces. A wel l -kn own theorem o f $ \mat h b f{ ZFC } $s tat es tha t all first count able Hausd or f fspaces are $k$ -s paces ( cf. The orem3 .3.20 o f [@E]).We show t ha t th i s th eorem mayfail in$\mathbf{ Z F}$.M or e pre cis ely, w epro ve th at, if $\m athca l{M}$is a mod el of $ \mathbf{ ZF}$, then all Hausdorf f firs t cou nta ble space s o f $\ mathcal{M }$ a re k-space s i f a nd on lyi f all met r iz abl e spac es i n $\mathca l {M }$a r e$k$-spacesw h i chholds if and on ly i f the axiom of co u ntable multipl e ch o i ce(Fo r m 12 6in [@HR]) is t rue i n $\ | aziz University,_P.O.Box: 80203_Jeddah 21589, Saudi Arabia'
-_'Siedlce University_of_Natural Sciences_and_Humanities, 3 Maja_54, 08-110 Siedlce,_Poland'
author:
- 'K. Keremedis, C._Özel, A. Pikosz,_M._A. Al Shumrani, E. Wajch'
title: 'Compact complement topologies and k-spaces'
---
[^1]
Introduction
============
The compact complement topology of_the_real line_was_considered,_for instance, in Example 22_of the celebrated book by_Steen and_Seebach“Counterexamples in Topology” ([@SS]). We investigate this notion_in_a much wider_context of Hausdorff spaces and of partially topological spaces_that belong to the class of_generalized topological spaces_in_the_sense of Delfs-Knebusch (cf._[@DK] and [@Pie1]). Our results are_proved in $\mathbf{ZF}$ if this is_not otherwise stated. All axioms of $\mathbf{ZF}$_can be found in [@Ku].
In section_2, we give elementary properties_of the_compact complement topology of a_Hausdorff space. In_particular, we_show that if_a Hausdorff space is locally compact_and second countable,_then its compact complement topology is_second_countable, while the_compact_complement_topology of_a non-locally compact_metrizable_space need_not_be first countable. We give an_example_of a countable metrizable space whose compact_complement topology is not_first_countable. In section 3,_a necessary and sufficient condition_for a set to be compact_with respect_to the_compact complement topology of a given Hausdorff space leads us to_a new characterization of k-spaces. A_well-known theorem of $\mathbf{ZFC}$_states that_all_first countable Hausdorff_spaces_are $k$-spaces_(cf. Theorem 3.3.20 of [@E]). We show_that this_theorem may fail in $\mathbf{ZF}$. More_precisely, we prove that,_if_$\mathcal{M}$ is a model of $\mathbf{ZF}$,_then all Hausdorff first countable spaces_of $\mathcal{M}$ are k-spaces if_and_only_if all metrizable spaces in_$\mathcal{M}$ are $k$-spaces which holds if_and only if_the axiom of countable multiple choice (Form_126_in [@HR]) is true in $\ |
Ti$_2$Al$_{20}$ [@Matsubayashi2012]) and with field \[Pr(Ir,Rh)$_2$Zn$_{20}$ [@Onimaru2011; @Onimaru2012] and PrV$_2$Al$_{20}$ [@Sakai2011]\], leading to an extended non-Fermi liquid region at higher temperatures. Pressure enhances the superconductivity in PrTi$_2$Al$_{20}$ [@Matsubayashi2012], which is likely unconventional. The in-field phase diagrams are even more interesting, as there is a heavy Fermi liquid region sandwiched between the zero-field order and a polarized high field state where Kondo physics is lost[@Onimaru2016b; @Yoshida2017].
Structure of the paper
----------------------
Ferrohastatic order and the generic infinite-$U$ two-channel Anderson model is introduced in section \[sec:FHorder\]. Section \[sec:microscopics\] fleshes out the details of the microscopic Anderson model and solves it within a large-$N$ mean-field theory for both FH and the competing antiferrohastatic (AFH) orders, and also considers interactions with the competing antiferroquadrupolar (AFQ) order. In section \[sec:landau\], we develop the Landau theory of cubic ferrohastatic order, examine its interactions with field, strain and AFQ order and discuss the thermodynamic signatures. Finally, in Section \[sec:signatures\] we summarize and expand upon the experimental signatures of FH order and how it may be distinguished from quadrupolar orders, before concluding in Section \[sec:conclusion\].
Ferrohastatic order \[sec:FHorder\]
===================================
Hastatic order is a natural candidate for materials with an even number of $f$ electrons and doublet crystal-field ground states. The cubic $\Gamma_3$ doublet is the simplest of these, with no dipole moments, $\langle \vec{J} \rangle = 0$, but finite quadrupolar ($O_{x^2-y^2}$, $O_{3z^2-r^2}$) and octupolar ($T_{xyz}$) moments [@Cox1998], and it is protected by cubic, not time-reversal, symmetry. Overlap between the non-Kramers $\Gamma_3$ states and conduction electrons leads to valence fluctuations, shown in Fig. \[fig:leveldiagram | Ti$_2$Al$_{20}$ [ @Matsubayashi2012 ]) and with field \[Pr(Ir, Rh)$_2$Zn$_{20}$ [ @Onimaru2011; @Onimaru2012 ] and PrV$_2$Al$_{20}$ [ @Sakai2011]\ ], leading to an extended non - Fermi liquid region at eminent temperature. Pressure enhances the superconductivity in PrTi$_2$Al$_{20}$ [ @Matsubayashi2012 ], which is likely improper. The in - field phase diagrams are even more interesting, as there is a intemperate Fermi liquid region sandwiched between the zero - battlefield order and a polarized gamey field state where Kondo purgative is lost[@Onimaru2016b; @Yoshida2017 ].
Structure of the paper
----------------------
Ferrohastatic order and the generic infinite-$U$ two - channel Anderson model is introduced in section \[sec: FHorder\ ]. part \[sec: microscopics\ ] fleshes out the details of the microscopic Anderson model and clear it within a large-$N$ mean - field theory for both FH and the compete antiferrohastatic (AFH) orders, and besides considers interactions with the competing antiferroquadrupolar (AFQ) order. In department \[sec: landau\ ], we develop the Landau theory of cubic ferrohastatic order, examine its interactions with playing field, tenor and AFQ order and discuss the thermodynamic signatures. Finally, in Section \[sec: signatures\ ] we summarize and expand upon the experimental signature of FH ordering and how it may be distinguished from quadrupolar club, before conclude in Section \[sec: conclusion\ ].
Ferrohastatic order \[sec: FHorder\ ]
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Hastatic order is a natural campaigner for materials with an tied number of $ f$ electrons and doublet crystal - airfield ground states. The cubic $ \Gamma_3 $ doublet is the simplest of these, with no dipole moments, $ \langle \vec{J } \rangle = 0 $, but finite quadrupolar ($ O_{x^2 - y^2}$, $ O_{3z^2 - r^2}$) and octupolar ($ T_{xyz}$) moments [ @Cox1998 ], and it is protect by cubic, not time - reversal, symmetry. Overlap between the non - Kramers $ \Gamma_3 $ states and conduction electron leads to valence fluctuation, shown in Fig. \[fig: leveldiagram | Ti$_2$Ap$_{20}$ [@Matsubayashi2012]) and with nield \[Pr(Ir,Rh)$_2$Zn$_{20}$ [@Ouumaru2011; @Mnimarh2012] and PrX$_2$Al$_{20}$ [@Sakai2011]\], leading to an exteided non-Ftgmi liquid region at figher telperaturws. Pcessure enhances the supcxcondhgtivicy in PrTi$_2$Al$_{20}$ [@Matxubayashi2012], fhich is likelf jneonventional. The in-field phase diagrwms are egen more intergstinb, as fhere is a heavy Fermi liquid regikn sandxiched between yhe zero-field order and a oolagized high field shate where Jondj physics is uost[@Onimarl2016y; @Yoshida2017].
Sttucture of the paper
----------------------
Ferrohastatiz ordzr and the teberlw infinite-$U$ two-crannel Anderson moden is inyroduced in segtion \[sex:FHorder\]. Section \[sec:mmcroscopics\] fleshes jut the datcils of the microscopuc Andetson kodeu ana skltes it wihhii a large-$N$ jean-field tyeory for both FH amd nne competing antifqrwohastatic (AFH) orders, and also considerv ihteractions with the conpeting antiferroquadtupolar (AFZ) order. In section \[sec:landau\], we develop the Landag thekfy if zyblc ferrohastatic order, examine its interactioga einh field, strain akd AFQ order and dosfuxf the thermodinamic sighatures. Finally, in Sectiog \[sec:wignaturef\] we summarize and expand upon rhe experimeutao signatures of FH order and kow it may ne distinguished from qbadrupklar orders, before ckvcluding in Sectkon \[sac:conclusion\].
Ferrohastatic jrder \[sec:HHordzr\]
===================================
Hastatkc otder is a natural candlgate for materials with at even numher of $f$ electrons and doublet rcystal-field gtoutd vtates. Tke cublc $\Gamma_3$ doublee is the simplgst of thzse, wigh no dipome momeits, $\langle \vqc{J} \rangle = 0$, tot finite quavrupolar ($J_{x^2-y^2}$, $I_{3z^2-r^2}$) qnd octjoolar ($T_{xyz}$) momrnts [@Cox1998], cud it is protected by cubic, noj fime-reversal, synmeurt. Overlap betwren thq joi-Kramqss $\Gamma_3$ stades xnd vonduztion electvonr lesds to valence fluctgatikns, shown in Fig. \[fog:jeveldiatram | Ti$_2$Al$_{20}$ [@Matsubayashi2012]) and with field \[Pr(Ir,Rh)$_2$Zn$_{20}$ [@Onimaru2011; PrV$_2$Al$_{20}$ leading to extended non-Fermi liquid enhances superconductivity in PrTi$_2$Al$_{20}$ which is likely The in-field phase diagrams are even interesting, as there is a heavy Fermi liquid region sandwiched between the zero-field and a polarized high field state where Kondo physics is lost[@Onimaru2016b; @Yoshida2017]. Structure the ---------------------- order the generic infinite-$U$ two-channel Anderson model is introduced in section \[sec:FHorder\]. Section \[sec:microscopics\] fleshes out the of the microscopic Anderson model and solves it a large-$N$ mean-field theory both FH and the competing (AFH) and also interactions the antiferroquadrupolar (AFQ) order. section \[sec:landau\], we develop the Landau theory of cubic ferrohastatic order, examine its interactions with field, strain AFQ order the thermodynamic Finally, Section we summarize and the experimental signatures of FH order may be distinguished from quadrupolar orders, before concluding Section \[sec:conclusion\]. order \[sec:FHorder\] =================================== Hastatic order is natural candidate for materials with an even number $f$ electrons and doublet crystal-field ground states. The cubic $\Gamma_3$ doublet is the simplest of no dipole moments, $\langle \rangle = 0$, finite ($O_{x^2-y^2}$, and ($T_{xyz}$) moments and it is protected by cubic, not time-reversal, symmetry. Overlap between non-Kramers $\Gamma_3$ states and conduction electrons leads to valence fluctuations, Fig. | Ti$_2$Al$_{20}$ [@Matsubayashi2012]) and with fiEld \[Pr(Ir,Rh)$_2$ZN$_{20}$ [@OnimAru2011; @oniMaRu2012] anD PrV$_2$al$_{20}$ [@Sakai2011]\], leading TO an eXtended non-Fermi liquid rEgion At HIgheR TeMperaTures. PrESsURE enHaNcEs tHe SUpErconDucTivity iN PrTi$_2$Al$_{20}$ [@MatSubAyAshi2012], which is lIKeLy unconvenTioNal. The in-fielD phAse diaGrAms ARe eveN moRe intErestiNG, as theRe is a heavY FERmi liqUId regioN SAnDwicHed between the zero-FIeLD order and a polaRized hIgH FiELD stAte Where Kondo PhYsics IS lost[@OnIMaRU2016B; @yosHIda2017].
Structure oF the paper
----------------------
FeRRohAstatiC oRdeR And the GenerIc INfiNite-$U$ two-chaNnel anderson mOdel is INtroducED in sectIon \[sec:fHoRdeR\]. SecTIoN \[sEc:mIcROscOPiCs\] fLEshEs out the DeTaIls of The mICROScopIc ANderSon moDel and solves iT wiThin A LarGe-$N$ meAn-fieLd thEoRy for Both FH And thE cOmpeting antiferRohaStatic (AFH) OrdErS, anD aLso coNSiders IntEraCtions wIth the cOMpeTiNG ANtIferroquadrupolar (AfQ) ORDeR. In sectiOn \[sec:lANdAu\], WE develop ThE LaNdau THEory oF cubIC fErrohastAtic orDEr, ExAmine itS iNteracTiOns WitH fielD, StraIn and AfQ order aNd disCUss the thermodyNAmic signatureS. fiNALlY, In SeCtiOn \[sec:signatUres\] WE sumMariZE aNd eXPand uPon thE eXPeRImental signatures of fH Order aNd how It may be distinGuished froM QUAdrupolaR ordERs, BEfore concludinG in SeCtion \[sec:coNClusion\].
FErrohAstatic oRder \[sec:FHORDer\]
===================================
HastaTic OrdEr iS a nATUrAl candidate foR MAterIaLs with aN evEn numbeR of $F$ elEctRonS aNd doublet Crystal-fIeLd GrOuNd sTates. tHe cubic $\GAmMa_3$ dOuBleT is thE SimpleSt of tHese, WiTh NO diPole momENtS, $\LAnglE \vEc{j} \ranGle = 0$, BuT finiTe quADruPolar ($O_{x^2-Y^2}$, $O_{3z^2-r^2}$) and ocTupOLar ($T_{XyZ}$) mOments [@COx1998], and it is protEcTed by cubic, NoT tiMe-reveRSAl, symmetRy. Overlap between the non-KRAmers $\GaMma_3$ StateS and ConductioN elEctronS leADs to vaLence fLuctuAtIonS, SHown iN fIg. \[Fig:LeVeldiagram | Ti$_2$Al$_{20}$ [@Matsubay ashi2012]) andwit h f ie ld \ [Pr( Ir,Rh)$_2$Zn$_ { 20}$ [@Onimaru2011; @Onima ru201 2] andP rV $_2$A l$_{20} $ [ @ S aka i2 01 1]\ ], le ading to an ext ended non- Fer mi liquid regi o nat highertem peratures. P res sure e nh anc e s the su perco nducti v ity in PrTi$_2$ Al $ _{20}$ [@Matsu b a ya shi2 012], which is li k el y unconventiona l. The i n -f i e ldpha se diagram sare e v en more in t e r est i ng, as thereis a heavyF erm i liqu id re g ion sa ndwic he d be tween the z ero- field ord er and a polar i zed hig h fiel d s tat e wh e re K ond op hys i cs is los t[@Onima ru 20 16b;@Yos h i d a 2017 ].
Str uctur e of the pape r
- ---- - --- ----- ----- ---
F erroh astati c ord er and the generi c in finite-$U $ t wo -ch an nel A n derson mo del is int roduced inse c t i on \[sec:FHorder\].Se c t io n \[sec: micros c op ic s \] flesh es ou t th e detai ls o f t he micro scopic An de rson mo de l andso lve s i t wit h in a large -$N$ mea n-fie l d theory for b o th FH and the co m p et i ng a nti ferrohastat ic ( A FH)orde r s, an d also cons id e rs interactions with t he compe tingantiferroquad rupolar (A F Q ) order.In s e ct i on \[sec:landa u\],we develop the Land au th eory ofcubic fer r o hastatic or der , e xam i n eits interacti o n s wi th field, st rain an d A FQord eran d discuss the the rm od yn am icsigna t ures. Fi na lly ,inSecti o n \[se c:sig natu re s\ ] we summar i ze a nd e xp an d up onth e exp erim e nta l signa tures ofFHo rder a nd how it may be disti ng uished fro mqua drupol a r orders, before concluding in S e ction \ [se c:con clus ion\].
F err ohasta tic order\[sec: FHord er \]= = ===== = = == === == ========== = = === ====
H asta tic ord er is a natural ca n did ate for mater ial s wi t h a n e v en num be r of $ f$ electrons an d doubletcr y st al-field g r oun dstates. The cu bic $ \ Gamma_3 $ doublet is the s im ples t ofthese, wit h no dip ole momen t s, $\ l an gle \ vec {J} \r an gle = 0$ , butf ini te qu adrupo la r ($O_ {x^2- y^ 2}$, $O_ {3z^2-r^2}$) and octupo lar ($ T_{xy z}$ ) moments [@ C ox1 998], and itis protect edbycubic , n o t tim e-re v er sal , symm etry . Overlapb et wee n th e non-Krame r s $\G amma_ 3$s tatesandconduction electr o ns leads to va lenc e flu ctu a tion s, shown in Fig. \[ fi g : leveldia gr am | Ti$_2$Al$_{20}$ [@Matsubayashi2012])_and with_field \[Pr(Ir,Rh)$_2$Zn$_{20}$ [@Onimaru2011; @Onimaru2012]_and PrV$_2$Al$_{20}$_[@Sakai2011]\],_leading to_an_extended non-Fermi liquid_region at higher_temperatures. Pressure enhances the_superconductivity in PrTi$_2$Al$_{20}$_[@Matsubayashi2012],_which is likely unconventional. The in-field phase diagrams are even more interesting, as there_is_a heavy_Fermi_liquid_region sandwiched between the zero-field_order and a polarized high_field state_where Kondo physics is lost[@Onimaru2016b; @Yoshida2017].
Structure of the_paper
----------------------
Ferrohastatic_order and the_generic infinite-$U$ two-channel Anderson model is introduced in section_\[sec:FHorder\]. Section \[sec:microscopics\] fleshes out the_details of the_microscopic_Anderson_model and solves it_within a large-$N$ mean-field theory for_both FH and the competing antiferrohastatic_(AFH) orders, and also considers interactions with_the competing antiferroquadrupolar (AFQ) order. In_section \[sec:landau\], we develop the_Landau theory_of cubic ferrohastatic order, examine_its interactions with_field, strain_and AFQ order_and discuss the thermodynamic signatures. Finally,_in Section \[sec:signatures\]_we summarize and expand upon the_experimental_signatures of FH_order_and_how it_may be distinguished_from_quadrupolar orders,_before_concluding in Section \[sec:conclusion\].
Ferrohastatic order \[sec:FHorder\]
===================================
Hastatic_order_is a natural candidate for materials with_an even number of_$f$_electrons and doublet crystal-field_ground states. The cubic $\Gamma_3$_doublet is the simplest of these,_with no_dipole moments,_$\langle \vec{J} \rangle = 0$, but finite quadrupolar ($O_{x^2-y^2}$, $O_{3z^2-r^2}$) and_octupolar ($T_{xyz}$) moments [@Cox1998], and it_is protected by cubic,_not time-reversal,_symmetry._Overlap between the_non-Kramers_$\Gamma_3$ states_and conduction electrons leads to valence fluctuations,_shown in_Fig. \[fig:leveldiagram |
005 & -12 04 00.239 & 0.007 & 0.75 &0.02 &29.32\
G19.36-0.03 & 18 26 25.783 & 0.010 & -12 04 00.248 & 0.013 & 0.39 &0.02 &29.49\
G19.36-0.03 & 18 26 25.783 & 0.014 & -12 04 00.248 & 0.018 & 0.28 &0.02 &29.66\
G19.36-0.03 & 18 26 25.784 & 0.024 & -12 04 00.221 & 0.030 & 0.17 &0.02 &29.82\
G22.04+0.22 & 18 30 35.035 & 0.024 & -09 34 45.823 & 0.026 & 0.16 &0.02 &48.67\
G22.04+0.22 & 18 30 35.039 & 0.009 & -09 34 45.820 & 0.009 & 0.48 &0.02 &48.84\
G22.04+0.22 & 18 30 35.040 & 0.005 & -09 34 45.806 & 0.005 & 0.88 &0.02 &49.00\
G22.04+0.22 & 18 30 35.041 & 0.004 & -09 34 45.794 & 0.005 & 0.95 &0.02 &49.17\
G22.04+0.22 & 18 30 35.040 & 0.006 & -09 34 45.776 & 0.007 & 0.63 &0.02 &49.34\
G22.04+0.22 & 18 30 34.728 & 0.023 & -09 34 53.696 & 0.025 & 0.17 &0.02 &49.34\
G22.04+0.22 & 18 30 35.039 & 0.015 & -09 34 45.728 & 0.016 & 0.29 &0.02 &49.50\
G22.04+0.22 & 18 30 34.866 & 0.012 & -09 34 44.974 & 0.013 & 0.35 &0.02 &49.50\
G22.04+0.22 & 18 | 005 & -12 04 00.239 & 0.007 & 0.75 & 0.02 & 29.32\
G19.36 - 0.03 & 18 26 25.783 & 0.010 & -12 04 00.248 & 0.013 & 0.39 & 0.02 & 29.49\
G19.36 - 0.03 & 18 26 25.783 & 0.014 & -12 04 00.248 & 0.018 & 0.28 & 0.02 & 29.66\
G19.36 - 0.03 & 18 26 25.784 & 0.024 & -12 04 00.221 & 0.030 & 0.17 & 0.02 & 29.82\
G22.04 + 0.22 & 18 30 35.035 & 0.024 & -09 34 45.823 & 0.026 & 0.16 & 0.02 & 48.67\
G22.04 + 0.22 & 18 30 35.039 & 0.009 & -09 34 45.820 & 0.009 & 0.48 & 0.02 & 48.84\
G22.04 + 0.22 & 18 30 35.040 & 0.005 & -09 34 45.806 & 0.005 & 0.88 & 0.02 & 49.00\
G22.04 + 0.22 & 18 30 35.041 & 0.004 & -09 34 45.794 & 0.005 & 0.95 & 0.02 & 49.17\
G22.04 + 0.22 & 18 30 35.040 & 0.006 & -09 34 45.776 & 0.007 & 0.63 & 0.02 & 49.34\
G22.04 + 0.22 & 18 30 34.728 & 0.023 & -09 34 53.696 & 0.025 & 0.17 & 0.02 & 49.34\
G22.04 + 0.22 & 18 30 35.039 & 0.015 & -09 34 45.728 & 0.016 & 0.29 & 0.02 & 49.50\
G22.04 + 0.22 & 18 30 34.866 & 0.012 & -09 34 44.974 & 0.013 & 0.35 & 0.02 & 49.50\
G22.04 + 0.22 & 18 | 005 & -12 04 00.239 & 0.007 & 0.75 &0.02 &29.32\
G19.36-0.03 & 18 26 25.783 & 0.010 & -12 04 00.248 & 0.013 & 0.39 &0.02 &29.49\
N19.36-0.03 & 18 26 25.783 & 0.014 & -12 04 00.248 & 0.018 & 0.28 &0.02 &29.66\
G19.36-0.03 & 18 26 25.784 & 0.024 & -12 04 00.221 & 0.030 & 0.17 &0.02 &29.82\
G22.04+0.22 & 18 30 35.035 & 0.024 & -09 34 45.823 & 0.026 & 0.16 &0.02 &48.67\
G22.04+0.22 & 18 30 35.039 & 0.009 & -09 34 45.820 & 0.009 & 0.48 &0.02 &48.84\
G22.04+0.22 & 18 30 35.040 & 0.005 & -09 34 45.806 & 0.005 & 0.88 &0.02 &49.00\
G22.04+0.22 & 18 30 35.041 & 0.004 & -09 34 45.794 & 0.005 & 0.95 &0.02 &49.17\
G22.04+0.22 & 18 30 35.040 & 0.006 & -09 34 45.776 & 0.007 & 0.63 &0.02 &49.34\
G22.04+0.22 & 18 30 34.728 & 0.023 & -09 34 53.696 & 0.025 & 0.17 &0.02 &49.34\
G22.04+0.22 & 18 30 35.039 & 0.015 & -09 34 45.728 & 0.016 & 0.29 &0.02 &49.50\
G22.04+0.22 & 18 30 34.866 & 0.012 & -09 34 44.974 & 0.013 & 0.35 &0.02 &49.50\
C22.04+0.22 & 18 | 005 & -12 04 00.239 & 0.007 &0.02 G19.36-0.03 & 26 25.783 & & & 0.39 &0.02 G19.36-0.03 & 18 25.783 & 0.014 & -12 04 & 0.018 & 0.28 &0.02 &29.66\ G19.36-0.03 & 18 26 25.784 & 0.024 -12 04 00.221 & 0.030 & 0.17 &0.02 &29.82\ G22.04+0.22 & 18 30 & & 34 & 0.026 & 0.16 &0.02 &48.67\ G22.04+0.22 & 18 30 35.039 & 0.009 & -09 34 & 0.009 & 0.48 &0.02 &48.84\ G22.04+0.22 & 30 35.040 & 0.005 -09 34 45.806 & 0.005 0.88 &49.00\ G22.04+0.22 18 35.041 0.004 & -09 45.794 & 0.005 & 0.95 &0.02 &49.17\ G22.04+0.22 & 18 30 35.040 & 0.006 & -09 34 & 0.007 &0.02 &49.34\ & 30 & 0.023 & 53.696 & 0.025 & 0.17 &0.02 18 30 35.039 & 0.015 & -09 34 & 0.016 0.29 &0.02 &49.50\ G22.04+0.22 & 18 34.866 & 0.012 & -09 34 44.974 & & 0.35 &0.02 &49.50\ G22.04+0.22 & 18 | 005 & -12 04 00.239 & 0.007 & 0.75 &0.02 &29.32\
G19.36-0.03 & 18 26 25.783 & 0.010 & -12 04 00.248 & 0.013 & 0.39 &0.02 &29.49\
G19.36-0.03 & 18 26 25.783 & 0.014 & -12 04 00.248 & 0.018 & 0.28 &0.02 &29.66\
G19.36-0.03 & 18 26 25.784 & 0.024 & -12 04 00.221 & 0.030 & 0.17 &0.02 &29.82\
G22.04+0.22 & 18 30 35.035 & 0.024 & -09 34 45.823 & 0.026 & 0.16 &0.02 &48.67\
G22.04+0.22 & 18 30 35.039 & 0.009 & -09 34 45.820 & 0.009 & 0.48 &0.02 &48.84\
G22.04+0.22 & 18 30 35.040 & 0.005 & -09 34 45.806 & 0.005 & 0.88 &0.02 &49.00\
G22.04+0.22 & 18 30 35.041 & 0.004 & -09 34 45.794 & 0.005 & 0.95 &0.02 &49.17\
G22.04+0.22 & 18 30 35.040 & 0.006 & -09 34 45.776 & 0.007 & 0.63 &0.02 &49.34\
G22.04+0.22 & 18 30 34.728 & 0.023 & -09 34 53.696 & 0.025 & 0.17 &0.02 &49.34\
G22.04+0.22 & 18 30 35.039 & 0.015 & -09 34 45.728 & 0.016 & 0.29 &0.02 &49.50\
G22.04+0.22 & 18 30 34.866 & 0.012 & -09 34 44.974 & 0.013 & 0.35 &0.02 &49.50\
G22.04+0.22 & 18 | 005 & -12 04 00.239 & 0.00 7 & 0.75 & 0.02&29 .32 \G19. 36-0 .03 & 18 26 25 . 783& 0.010 & -12 04 00.24 8 & 0 .0 1 3 &0 .3 9 &0. 02 &29. 4 9\ G 19. 36 -0 .03 & 18 26 2 5.7 83 & 0. 014 & -120400 .248 & 0.018 &0.28 &0.02 &2 9.66\
G19.36 -0. 03 & 1 8262 5.784 &0.024 & -12 04 00. 221 & 0.0 30 & 0.17 &0.02 & 2 9 .8 2\
G 22.04+0.22 & 18 3 0 3 5 .035 & 0.024 & -09 3 44 5. 8 2 3 & 0. 026 & 0.16 & 0.02& 48.67\G 22 . 0 4 +0. 2 2 & 18 30 35. 039 & 0.009 & - 09 3445 .82 0 & 0.0 09 &0. 4 8 & 0.02 &48.84 \
G2 2.04+0.22 & 183 0 35.04 0 & 0.00 5 & -0 9 3 4 4 5.80 6 & 0 .00 5& 0. 8 8&0. 0 2 & 49.00\
G 22 .0 4+0.2 2 &1 8 3 0 35 .04 1 &0.004 & -09 34 45. 794 & 0 . 005 & 0. 95 &0 .02&4 9.17\
G22.0 4+0.2 2& 18 30 35.040& 0. 006 & -09 34 4 5.7 76 & 0. 0 07 & 0 .63 &0 .02 &49 .34\
G2 2 .04 +0 . 2 2 & 18 30 34.728 & 0. 02 3 &-09 34 5 3.696& 0 .0 2 5 & 0.17 & 0.0 2 &4 9 . 34\
G 22.0 4 +0 .22 & 18 30 35 . 03 9& 0.015 & -09 3 445. 728 & 0. 0 16 & 0.29&0.02 &4 9.50\ G22.04+0.22 &1 8 30 34.866 & 0. 0 1 2& -09 34 44.974 & 0 .013 & 0. 35 & 0 .0 2 & 4 9.50\
G22. 04 + 0. 2 2 & 18 | 005 &_-12 04_00.239 & 0.007 &_0.75 &0.02_&29.32\
G19.36-0.03_& 18_26_25.783 & 0.010_& -12 04_00.248 & 0.013 &_0.39 &0.02 &29.49\
G19.36-0.03_&_18 26 25.783 & 0.014 & -12 04 00.248 & 0.018 & 0.28 &0.02_&29.66\
G19.36-0.03_& 18_26_25.784_& 0.024 & -12 04_00.221 & 0.030 & 0.17_&0.02 &29.82\
G22.04+0.22_& 18 30 35.035 & 0.024 & -09_34_45.823 & 0.026_& 0.16 &0.02 &48.67\
G22.04+0.22 & 18 30 35.039 &_0.009 & -09 34 45.820 &_0.009 & 0.48_&0.02_&48.84\
G22.04+0.22_& 18 30 35.040_& 0.005 & -09 34 45.806_& 0.005 & 0.88 &0.02 &49.00\
G22.04+0.22_& 18 30 35.041 & 0.004 &_-09 34 45.794 & 0.005 &_0.95 &0.02 &49.17\
G22.04+0.22 & 18_30 35.040_& 0.006 & -09 34_45.776 & 0.007_& 0.63_&0.02 &49.34\
G22.04+0.22 &_18 30 34.728 & 0.023 &_-09 34 53.696_& 0.025 & 0.17 &0.02 &49.34\
G22.04+0.22_&_18 30 35.039_&_0.015_& -09_34 45.728 &_0.016_& 0.29_&0.02_&49.50\
G22.04+0.22 & 18 30 34.866 &_0.012_& -09 34 44.974 & 0.013 &_0.35 &0.02 &49.50\
G22.04+0.22 &_18 |
g., by introducing a basis of Hermitian, trace-orthogonal, and traceless operators \[e.g., su($D$)\], whence the $\rho _{i}$ are the expansion coefficients of $\rho $ in this basis [@Alicki:87], with $\rho _{1}$ the coefficient of $I$ (the identity matrix). In this case, the condition $\mathrm{Tr}\rho ^{2}\leq 1$ corresponds to $\left\Vert |\rho \rangle \rangle \right\Vert \leq 1$, $\rho =\rho ^{\dag }$ to $\rho _{i}=\rho _{i}^{\ast }$, and positive semidefiniteness of $\rho $ is expressed in terms of inequalities satisfied by certain Casimir invariants \[e.g., of $su(D)$\] [@byrd:062322; @Kimura:2003-1; @Kimura:2003-2]. A simple and well-known example of this procedure is the representation of the density operator of a two-level system (qubit) on the Bloch sphere, via $\rho =(I_{2}+\overrightarrow{v}\cdot \overrightarrow{\sigma })/2$, where $\overrightarrow{\sigma }=(\sigma _{x},\sigma _{y},\sigma _{z})$ is the vector of Pauli matrices, $I_{2}$ is the $2\times 2$ identity matrix, and $\overrightarrow{v}\in \mathbb{R}^{3}$ is a three-dimensional coherence vector of norm$\leq 1$. More generally, coherence vectors live in Hilbert-Schmidt space: a state space of linear operators endowed with an inner product that can be defined, for general vectors $u$ and $v$, as $$(u,v)\equiv \langle \langle u|v\rangle \rangle \equiv \frac{1}{{\mathcal{N}}}{\text{Tr}}\left( u^{\dagger }v\right), \label{ip}$$where ${\mathcal{N}}$ is a normalization factor. Adjoint elements $\langle
\langle v|$ in the dual state space are given by row vectors defined as the transpose conjugate of $|v\rangle \rangle $: $\langle \langle
v|=(v_{1}^{\ast },v_{2}^{\ast },...,v_{D^{2}}^{\ast })$. A density matrix can then be expressed as a discrete superposition of states over a complete basis in this vector space, with appropriate constraints on the coefficients so that the requirements of | g., by introducing a basis of Hermitian, trace - orthogonal, and traceless operator \[e.g. , su($D$)\ ], whence the $ \rho _ { i}$ are the expansion coefficient of $ \rho $ in this basis [ @Alicki:87 ], with $ \rho _ { 1}$ the coefficient of $ I$ (the identity matrix). In this case, the circumstance $ \mathrm{Tr}\rho ^{2}\leq 1 $ corresponds to $ \left\Vert |\rho \rangle \rangle \right\Vert \leq 1 $, $ \rho = \rho ^{\dag } $ to $ \rho _ { i}=\rho _ { i}^{\ast } $, and positive semidefiniteness of $ \rho $ is express in terms of inequality satisfied by certain Casimir invariants \[e.g. , of $ su(D)$\ ] [ @byrd:062322; @Kimura:2003 - 1; @Kimura:2003 - 2 ]. A elementary and well - known example of this routine is the representation of the density hustler of a two - level system (qubit) on the Bloch sphere, via $ \rho =( I_{2}+\overrightarrow{v}\cdot \overrightarrow{\sigma }) /2 $, where $ \overrightarrow{\sigma } =( \sigma _ { x},\sigma _ { y},\sigma _ { z})$ is the vector of Pauli matrices, $ I_{2}$ is the $ 2\times 2 $ identity matrix, and $ \overrightarrow{v}\in \mathbb{R}^{3}$ is a three - dimensional coherence vector of norm$\leq 1$. More generally, coherence vectors live in Hilbert - Schmidt outer space: a state space of linear hustler endowed with an inner product that can be defined, for cosmopolitan vectors $ u$ and $ v$, as $ $ (uranium, v)\equiv \langle \langle u|v\rangle \rangle \equiv \frac{1}{{\mathcal{N}}}{\text{Tr}}\left (u^{\dagger } v\right), \label{ip}$$where $ { \mathcal{N}}$ is a normalization factor. Adjoint elements $ \langle
\langle v|$ in the dual department of state space are given by row vectors defined as the transpose conjugate of $ |v\rangle \rangle $: $ \langle \langle
v|=(v_{1}^{\ast }, v_{2}^{\ast }, ..., v_{D^{2}}^{\ast }) $. A density matrix can then be expressed as a discrete superposition of states over a complete basis in this vector space, with appropriate constraints on the coefficients so that the requirements of | g., bj introducing a basis of Hermitian, tracg-oethogoial, and tracelers operators \[e.g., su($D$)\], whence tie $\ryo _{i}$ qre the expansion coefwicients lf $\rho $ un tiis basis [@Alicki:87], with $\rho _{1}$ the goeffncment of $I$ (the icentity madrix). In this cdsd, che condition $\mathrm{Tr}\rho ^{2}\leq 1$ correfponds yo $\left\Vert |\rho \tanglt \rwngls \right\Vert \leq 1$, $\rho =\rho ^{\dag }$ to $\rgo _{i}=\rho _{i}^{\ast }$, and poxitive semidefiniteness of $\rho $ is expressed in herms of ingsuajuties satisfked by ceruann Casimir jnvariants \[e.g., of $su(D)$\] [@byrd:062322; @Kimurx:2003-1; @Kimbra:2003-2]. A simple abd agll-known exanple jf this procccure iv the rrpresentation pf vhe eensity operator of a two-level system (qobit) on tha Yloch sphere, via $\rho =(U_{2}+\ocerrichtasrow{x}\xdog \obecrifhtarrlw{\smgma })/2$, where $\overrightaerow{\sigma }=(\sigma _{x},\sibmw _{y},\sigma _{z})$ is fhe vestjr of Pauli matrices, $I_{2}$ is the $2\times 2$ idtntitg matrix, and $\overrightaerow{v}\in \mathbb{R}^{3}$ is a jhree-dimenfional coherence vector of norm$\leq 1$. More generallf, cohxrdnct ycctofw pive in Hilbert-Schmidt space: a state space of jjntar operators endjwed with am lnmgr product thaj can bz dsfined, for general vectorf $u$ abd $v$, as $$(u,d)\equov \langle \langle u|v\rangle \rqngle \equiv \yrax{1}{{\mathcal{N}}}{\text{Tr}}\lefc( u^{\dagger }v\rnght), \kabel{op}$$where ${\mathcal{N}}$ is a nurmamization faftor. Adjojvt elements $\langue
\lsncle v|$ in the dual state spwce are gmven yy row vdctots defiged as the transpose conjugate of $|g\rangpe \rangle $: $\lwngle \langle
v|=(v_{1}^{\ast },v_{2}^{\ast },...,v_{D^{2}}^{\ast })$. A density matric wan then be exprcssed as a discwete superposijion of scates uver a comilete basms in this vqctor space, whjh appropriatx constraynts on rhe coewwicients so thst the requirements of | g., by introducing a basis of Hermitian, traceless \[e.g., su($D$)\], the $\rho _{i}$ $\rho in this basis with $\rho _{1}$ coefficient of $I$ (the identity matrix). this case, the condition $\mathrm{Tr}\rho ^{2}\leq 1$ corresponds to $\left\Vert |\rho \rangle \rangle \leq 1$, $\rho =\rho ^{\dag }$ to $\rho _{i}=\rho _{i}^{\ast }$, and positive of $ expressed terms of inequalities satisfied by certain Casimir invariants \[e.g., of $su(D)$\] [@byrd:062322; @Kimura:2003-1; @Kimura:2003-2]. A simple well-known example of this procedure is the representation the density operator of two-level system (qubit) on the sphere, $\rho =(I_{2}+\overrightarrow{v}\cdot })/2$, $\overrightarrow{\sigma _{x},\sigma _{y},\sigma _{z})$ the vector of Pauli matrices, $I_{2}$ is the $2\times 2$ identity matrix, and $\overrightarrow{v}\in \mathbb{R}^{3}$ is a coherence vector 1$. More coherence live Hilbert-Schmidt space: a of linear operators endowed with an can be defined, for general vectors $u$ and as $$(u,v)\equiv \langle u|v\rangle \rangle \equiv \frac{1}{{\mathcal{N}}}{\text{Tr}}\left( u^{\dagger \label{ip}$$where ${\mathcal{N}}$ is a normalization factor. Adjoint elements \langle v|$ in the dual state space are given by row vectors defined as the of $|v\rangle \rangle $: \langle v|=(v_{1}^{\ast },v_{2}^{\ast })$. density can be expressed a discrete superposition of states over a complete basis in this space, with appropriate constraints on the coefficients so that the | g., by introducing a basis of HerMitian, tracE-orthOgoNal, AnD traCeleSs operators \[e.g., sU($d$)\], wheNce the $\rho _{i}$ are the expansIon coEfFIcieNTs Of $\rho $ In this bASiS [@aLicKi:87], WiTh $\rHo _{1}$ THe CoeffIciEnt of $I$ (tHe identity MatRiX). In this case, tHE cOndition $\maThrM{Tr}\rho ^{2}\leq 1$ corResPonds tO $\lEft\vErt |\rhO \raNgle \rAngle \rIGht\VerT \leq 1$, $\rho =\rhO ^{\dAG }$ to $\rho _{I}=\Rho _{i}^{\ast }$, AND pOsitIve semidefiniteneSS oF $\Rho $ is expressed In termS oF InEQUalItiEs satisfieD bY certAIn CasimIR iNVARiaNTs \[e.g., of $su(D)$\] [@byrd:062322; @kimura:2003-1; @KimurA:2003-2]. a siMple anD wEll-KNown exAmple Of THis Procedure is The rEpresentaTion of THe densiTY operatOr of a tWo-lEveL sysTEm (QuBit) On THe BLOcH spHEre, Via $\rho =(I_{2}+\oVeRrIghtaRrow{V}\CDOT \oveRriGhtaRrow{\sIgma })/2$, where $\overRigHtarROw{\sIgma }=(\sIgma _{x},\SigmA _{y},\Sigma _{Z})$ is the VectoR oF Pauli matrices, $I_{2}$ Is thE $2\times 2$ ideNtiTy MatRiX, and $\oVErrighTarRow{V}\in \mathBb{R}^{3}$ is a tHRee-DiMENSiOnal coherence vectoR oF NOrM$\leq 1$. More GeneraLLy, CoHErence veCtOrs Live IN hilbeRt-ScHMiDt space: a State sPAcE oF linear OpEratorS eNdoWed With aN InneR produCt that caN be deFIned, for general VEctors $u$ and $v$, as $$(U,V)\eQUIv \LAnglE \laNgle u|v\ranglE \ranGLe \eqUiv \fRAc{1}{{\MatHCal{N}}}{\tExt{Tr}}\LeFT( u^{\DAgger }v\right), \label{ip}$$wHeRe ${\mathCal{N}}$ iS a normalizatiOn factor. AdJOINt elemenTs $\laNGlE
\Langle v|$ in the duAl staTe space are GIven by roW vectOrs definEd as the trANSpose conJugAte Of $|v\RanGLE \rAngle $: $\langle \laNGLe
v|=(v_{1}^{\AsT },v_{2}^{\ast },...,v_{D^{2}}^{\Ast })$. a densitY maTriX caN thEn Be expressEd as a disCrEtE sUpErpOsitiON of stateS oVer A cOmpLete bASis in tHis veCtor SpAcE, WitH appropRIaTE ConsTrAiNts oN thE cOeffiCienTS so That the RequiremeNts OF | g., by introducing a basis of Hermit ian,tra ce- or thog onal , and traceles s ope rators \[e.g., su($D$) \], w he n ce t h e$\rho _{i}$a re t heex pa nsi on co effic ien ts of $ \rho $ inthi sbasis [@Alic k i: 87], with$\r ho _{1}$ the co effici en t o f $I$(th e ide ntitym atrix) . In this c a se, th e condit i o n$\ma thrm{Tr}\rho ^{2} \ le q 1$ correspond s to $ \l e ft \ V ert |\ rho \rangl e\rang l e \righ t \V e r t \l e q 1$, $\rho = \rho ^{\dag }$to $\r ho _{ i }=\rho _{i} ^{ \ ast }$, and po siti ve semide finite n ess of$ \rho $is exp res sed int er ms of i n equ a li tie s sa tisfiedby c ertai n Ca s i m i r in var iant s \[e .g., of $su(D )$\ ] [@ b yrd :0623 22; @ Kimu ra :2003 -1; @K imura :2 003-2]. A simpl e an d well-kn own e xam pl e oft his pr oce dur e is th e repre s ent at i o n o f the density oper at o r o f a two- levels ys te m (qubit) o n t he B l o ch sp here , v ia $\rho =(I_{ 2 }+ \o verrigh ta rrow{v }\ cdo t \ overr i ghta rrow{\ sigma }) /2$,w here $\overrig h tarrow{\sigma }= ( \ si g ma _ {x} ,\sigma _{y },\s i gma_{z} ) $ist he ve ctorof Pa u li matrices, $I_{2} $is the $2\t imes 2$ ident ity matrix , a nd $\ove rrig h ta r row{v}\in \mat hbb{R }^{3}$ isa three-d imens ional co herence v e c tor of n orm $\l eq1$. M or e generally,c o here nc e vecto rslive in Hi lbe rt- Sch mi dt space: a state s pa ce o f l inear operator send ow edwitha n inne r pro duct t ha t ca n be de f in e d , fo rge nera l v ec tors$u$a nd$v$, as $$(u,v)\ equ i v \l an gl e \lang le u|v\rangle \ rangle \eq ui v \ frac{1 } { {\mathca l{N}}}{\text{Tr}}\left( u^{\dag ger }v\r ight ), \labe l{i p}$$wh ere ${\mat hcal{N }}$ i sa n o r maliz a t io n f ac tor. Adjoi n t el ement s$\la ngle
\l angle v|$ in the d u alstate space a regive n by ro w v e cto rs def i n ed as the trans pose conju ga t eof $|v\ran g le\r angle $ : $\lan gle \ l angle
v |=(v_{1}^ {\ast },v _{ 2}^{ \ a st},...,v_{D ^{2}}^{\ ast })$.A dens i ty matr ixcan th en be expr esseda s a disc rete s up erposi tionof statesover a complete basis i n this vect orspace, wi tha ppr opriate c onst raints onthe co effic ien t s sothat th e r e quire ment s of | g., by_introducing a_basis of Hermitian, trace-orthogonal,_and traceless_operators_\[e.g., su($D$)\],_whence_the $\rho _{i}$_are the expansion_coefficients of $\rho $_in this basis_[@Alicki:87],_with $\rho _{1}$ the coefficient of $I$ (the identity matrix). In this case, the_condition_$\mathrm{Tr}\rho ^{2}\leq_1$_corresponds_to $\left\Vert |\rho \rangle \rangle_\right\Vert \leq 1$, $\rho =\rho_^{\dag }$_to $\rho _{i}=\rho _{i}^{\ast }$, and positive semidefiniteness_of_$\rho $ is_expressed in terms of inequalities satisfied by certain Casimir_invariants \[e.g., of $su(D)$\] [@byrd:062322; @Kimura:2003-1;_@Kimura:2003-2]. A simple and_well-known_example_of this procedure is_the representation of the density operator_of a two-level system (qubit) on_the Bloch sphere, via $\rho =(I_{2}+\overrightarrow{v}\cdot \overrightarrow{\sigma_})/2$, where $\overrightarrow{\sigma }=(\sigma _{x},\sigma _{y},\sigma__{z})$ is the vector of_Pauli matrices,_$I_{2}$ is the $2\times 2$_identity matrix, and_$\overrightarrow{v}\in \mathbb{R}^{3}$_is a three-dimensional_coherence vector of norm$\leq 1$. More_generally, coherence vectors_live in Hilbert-Schmidt space: a state space_of_linear operators endowed_with_an_inner product_that can be_defined,_for general_vectors_$u$ and $v$, as $$(u,v)\equiv \langle_\langle_u|v\rangle \rangle \equiv \frac{1}{{\mathcal{N}}}{\text{Tr}}\left( u^{\dagger }v\right), _\label{ip}$$where ${\mathcal{N}}$ is a_normalization_factor. Adjoint elements $\langle
\langle_v|$ in the dual state_space are given by row vectors_defined as_the transpose_conjugate of $|v\rangle \rangle $: $\langle \langle
v|=(v_{1}^{\ast },v_{2}^{\ast },...,v_{D^{2}}^{\ast })$. A_density matrix can then be expressed_as a discrete superposition_of states_over_a complete basis_in_this vector_space, with appropriate constraints on the coefficients_so that_the requirements of |
}}\omega_3 - {\mathrm{i}}\omega_4)
\times \nonumber \\ & & \hspace{1em}
\mathcal{G}_{\gamma\gamma'}({\mathrm{i}}\omega_3) \mathcal{G}_{\delta\delta'}({\mathrm{i}}\omega_4)
V^{\mathrm{rn},-1}_{\gamma' i} V^{\mathrm{rn},-1}_{\delta' j}.
\label{eq:TwoParticleGF:connected:frequency:initial}\end{aligned}$$ In order to solve the frequency integral analytically, we again transform into the eigenbasis of $H_0 + \Sigma$. Eq. (\[eq:TwoParticleGF:connected:frequency:initial\]) now reads $$\begin{aligned}
& & \mathcal{G}^{(2,c)}_{ijij}({\mathrm{i}}\omega_1, {\mathrm{i}}\omega_2, {\mathrm{i}}\omega_3, {\mathrm{i}}\omega_4) =
- 2\pi \sum_{\alpha\beta\gamma\delta} \sum_{\tilde\mu\tilde\nu\tilde\rho\tilde\sigma}
V^{\mathrm{ri}}_{i\tilde\mu} V^{\mathrm{ri}}_{j\tilde\nu}
\times \nonumber \\ & & \hspace{1em}
\mathcal{G}_{\tilde\mu\tilde\mu}({\mathrm{i}}\omega_1) \mathcal{G}_{\tilde\nu\tilde\nu}({\mathrm{i}}\omega_2)
V^{\mathrm{in}}_{\tilde\mu\alpha} V^{\mathrm{in}}_{\tilde\nu\beta}
\times \nonumber \\ & & \hspace{1em}
\Gamma_{\alpha\beta\gamma\delta} \delta({\mathrm{i}}\omega_1 + {\mathrm{i}}\omega_2 - {\mathrm{i}}\omega_3 - {\mathrm{i}}\omega_4)
\times \nonumber \\ & & \hspace{1em}
V^{\mathrm{in},-1}_{\gamma\tilde\rho} V^{\mathrm{in},-1}_{\delta\tilde\sigma}
\mathcal{G}_{\tilde\rho\tilde\rho}({\mathrm{i}}\omega_3) \mathcal{G}_{\tilde\sigma\tilde\sigma}({\mathrm{i | } } \omega_3 - { \mathrm{i}}\omega_4)
\times \nonumber \\ & & \hspace{1em }
\mathcal{G}_{\gamma\gamma'}({\mathrm{i}}\omega_3) \mathcal{G}_{\delta\delta'}({\mathrm{i}}\omega_4)
V^{\mathrm{rn},-1}_{\gamma' i } V^{\mathrm{rn},-1}_{\delta' j }.
\label{eq: TwoParticleGF: connected: frequency: initial}\end{aligned}$$ In order to resolve the frequency built-in analytically, we again transform into the eigenbasis of $ H_0 + \Sigma$. Eq. (\[eq: TwoParticleGF: connected: frequency: initial\ ]) nowadays read $ $ \begin{aligned }
& & \mathcal{G}^{(2,c)}_{ijij}({\mathrm{i}}\omega_1, { \mathrm{i}}\omega_2, { \mathrm{i}}\omega_3, { \mathrm{i}}\omega_4) =
- 2\pi \sum_{\alpha\beta\gamma\delta } \sum_{\tilde\mu\tilde\nu\tilde\rho\tilde\sigma }
V^{\mathrm{ri}}_{i\tilde\mu } V^{\mathrm{ri}}_{j\tilde\nu }
\times \nonumber \\ & & \hspace{1em }
\mathcal{G}_{\tilde\mu\tilde\mu}({\mathrm{i}}\omega_1) \mathcal{G}_{\tilde\nu\tilde\nu}({\mathrm{i}}\omega_2)
V^{\mathrm{in}}_{\tilde\mu\alpha } V^{\mathrm{in}}_{\tilde\nu\beta }
\times \nonumber \\ & & \hspace{1em }
\Gamma_{\alpha\beta\gamma\delta } \delta({\mathrm{i}}\omega_1 + { \mathrm{i}}\omega_2 - { \mathrm{i}}\omega_3 - { \mathrm{i}}\omega_4)
\times \nonumber \\ & & \hspace{1em }
V^{\mathrm{in},-1}_{\gamma\tilde\rho } V^{\mathrm{in},-1}_{\delta\tilde\sigma }
\mathcal{G}_{\tilde\rho\tilde\rho}({\mathrm{i}}\omega_3) \mathcal{G}_{\tilde\sigma\tilde\sigma}({\mathrm{i | }}\omeha_3 - {\mathrm{i}}\omega_4)
\times \nunumber \\ & & \hspaew{1em}
\kathcam{G}_{\gamma\gxmma'}({\mathrm{i}}\omega_3) \mathcal{G}_{\delva\deota'}({\maukrm{i}}\omega_4)
V^{\mathrm{rv},-1}_{\gamma' i} N^{\mathrm{rn},-1}_{\eelte' j}.
\label{eq:Two'zrticleNY:connsgted:fxewuency:initial}\ekd{aligned}$$ It order to solee tke frequency integral analytically, wq again tgansform into jhe epgqnbaaps of $H_0 + \Sigma$. Eq. (\[eq:TwoParticleGR:connecued:frequency:initiak\]) now reads $$\begin{aligned}
& & \matjcal{G}^{(2,c)}_{ijij}({\mathrm{i}}\olega_1, {\mathrm{u}}\omedq_2, {\mathrm{i}}\omeea_3, {\mathrm{i}}\omega_4) =
- 2\pi \sum_{\alpha\beta\gamma\delta} \sum_{\tildd\mu\tikde\nu\tilde\tki\tipge\sigma}
V^{\matrrm{ri}}_{i\tilde\mm} V^{\mathsm{ri}}_{j\tikde\nu}
\times \nokumbec \\ & & \hspace{1em}
\mathcal{G}_{\tmlde\mu\tilde\mu}({\mathrm{i}}\jmega_1) \matvccl{G}_{\tilde\nu\tilde\nu}({\mathem{u}}\omegd_2)
V^{\magyrm{kn}}_{\tjlve\mh\alpha} V^{\methrm{in}}_{\tilds\nu\beta}
\timws \nonumber \\ & & \hspave{1qn}
\Gamma_{\alphz\beta\gwmia\delta} \delta({\mathrm{i}}\omega_1 + {\mathrm{i}}\omega_2 - {\mzthrm{i}}\omega_3 - {\mathrm{i}}\ometa_4)
\times \nonumber \\ & & \jspace{1em}
V^{\mathrm{in},-1}_{\gamma\tilde\rho} V^{\mathrm{in},-1}_{\delta\tilde\sigma}
\mavhzal{Y}_{\bllde\fyo\hilde\rho}({\mathrm{i}}\omega_3) \mathcal{G}_{\tilde\sigma\tilde\syfms}({\msthrm{i | }}\omega_3 - {\mathrm{i}}\omega_4) \times \nonumber \\ & \mathcal{G}_{\gamma\gamma'}({\mathrm{i}}\omega_3) V^{\mathrm{rn},-1}_{\gamma' i} j}. \label{eq:TwoParticleGF:connected:frequency:initial}\end{aligned}$$ In integral we again transform the eigenbasis of + \Sigma$. Eq. (\[eq:TwoParticleGF:connected:frequency:initial\]) now reads & & \mathcal{G}^{(2,c)}_{ijij}({\mathrm{i}}\omega_1, {\mathrm{i}}\omega_2, {\mathrm{i}}\omega_3, {\mathrm{i}}\omega_4) = - 2\pi \sum_{\alpha\beta\gamma\delta} \sum_{\tilde\mu\tilde\nu\tilde\rho\tilde\sigma} V^{\mathrm{ri}}_{i\tilde\mu} V^{\mathrm{ri}}_{j\tilde\nu} \nonumber \\ & & \hspace{1em} \mathcal{G}_{\tilde\mu\tilde\mu}({\mathrm{i}}\omega_1) \mathcal{G}_{\tilde\nu\tilde\nu}({\mathrm{i}}\omega_2) V^{\mathrm{in}}_{\tilde\mu\alpha} V^{\mathrm{in}}_{\tilde\nu\beta} \times \nonumber \\ & \hspace{1em} \delta({\mathrm{i}}\omega_1 {\mathrm{i}}\omega_2 {\mathrm{i}}\omega_3 - {\mathrm{i}}\omega_4) \times \nonumber \\ & & \hspace{1em} V^{\mathrm{in},-1}_{\gamma\tilde\rho} V^{\mathrm{in},-1}_{\delta\tilde\sigma} \mathcal{G}_{\tilde\rho\tilde\rho}({\mathrm{i}}\omega_3) \mathcal{G}_{\tilde\sigma\tilde\sigma}({\mathrm{i | }}\omega_3 - {\mathrm{i}}\omega_4)
\times \nonUmber \\ & & \hspacE{1em}
\maThcAl{G}_{\GaMma\gAmma'}({\Mathrm{i}}\omega_3) \maTHcal{g}_{\delta\delta'}({\mathrm{i}}\omegA_4)
V^{\matHrM{Rn},-1}_{\gaMMa' I} V^{\matHrm{rn},-1}_{\deLTa' J}.
\LAbeL{eQ:TWoPArTIcLeGF:cOnnEcted:frEquency:iniTiaL}\eNd{aligned}$$ In oRDeR to solve thE frEquency integRal AnalytIcAllY, We agaIn tRansfOrm intO The eigEnbasis of $h_0 + \SIGma$. Eq. (\[eQ:twoPartICLegF:coNnected:frequency:iNItIAl\]) now reads $$\begiN{alignEd}
& & \MAtHCAl{G}^{(2,C)}_{ijIj}({\mathrm{i}}\oMeGa_1, {\matHRm{i}}\omegA_2, {\MaTHRM{i}}\oMEga_3, {\mathrm{i}}\omeGa_4) =
- 2\pi \sum_{\alphA\BetA\gamma\DeLta} \SUm_{\tildE\mu\tiLdE\Nu\tIlde\rho\tildE\sigMa}
V^{\mathrm{Ri}}_{i\tilDE\mu} V^{\matHRm{ri}}_{j\tiLde\nu}
\tImeS \noNumbER \\ & & \hSpAce{1Em}
\MAthCAl{g}_{\tiLDe\mU\tilde\mu}({\MaThRm{i}}\omEga_1) \mATHCAl{G}_{\tIldE\nu\tIlde\nU}({\mathrm{i}}\omega_2)
v^{\maThrm{IN}}_{\tiLde\mu\Alpha} v^{\matHrM{in}}_{\tiLde\nu\bEta}
\tiMeS \nonumber \\ & & \hspace{1Em}
\GaMma_{\alpha\bEta\GaMma\DeLta} \deLTa({\mathRm{i}}\OmeGa_1 + {\mathrM{i}}\omega_2 - {\MAthRm{I}}\OMEgA_3 - {\mathrm{i}}\omega_4)
\times \NoNUMbEr \\ & & \hspace{1Em}
V^{\matHRm{In},-1}_{\GAmma\tildE\rHo} V^{\MathRM{In},-1}_{\delTa\tiLDe\Sigma}
\matHcal{G}_{\tILdE\rHo\tilde\RhO}({\mathrM{i}}\OmeGa_3) \mAthcaL{g}_{\tilDe\sigmA\tilde\siGma}({\maTHrm{i | }}\omega_3 - {\mathrm{i}}\ omega_4)
\time s \ non um ber\\ & & \hspace{1em }
\mathcal{G}_{\gamma\ga mma'} ({ \ math r m{ i}}\o mega_3) \m a t hca l{ G} _{\ de l ta \delt a'} ({\math rm{i}}\ome ga_ 4)
V^{\math r m{ rn},-1}_{\ gam ma' i} V^{\m ath rm{rn} ,- 1}_ { \delt a'j}.
\lab e l{eq:T woParticl eG F :conne c ted:fre q u en cy:i nitial}\end{align e d} $ $ In order tosolveth e f r e que ncy integralan alyti c ally, w e a g a i n t r ansform intothe eigenba s isof $H_ 0+ \ S igma$. Eq.(\ [ eq: TwoParticle GF:c onnected: freque n cy:init i al\]) n ow rea ds$$\ begi n {a li gne d} && \ mat h cal {G}^{(2, c) }_ {ijij }({\ m a t h rm{i }}\ omeg a_1,{\mathrm{i}}\ ome ga_2 , {\ mathr m{i}} \ome ga _3, { \mathr m{i}} \o mega_4) =
-2\pi \sum_{\a lph a\ bet a\ gamma \ delta} \s um_ {\tilde \mu\til d e\n u\ t i l de \rho\tilde\sigma} V ^{\mathr m{ri}} _ {i \t i lde\mu}V^ {\m athr m { ri}}_ {j\t i ld e\nu}
\ times\ no nu mber \\ & & \hs pa ce{ 1em }
\ math cal{G} _{\tilde \mu\t i lde\mu}({\math r m{i}}\omega_1 ) \ m a th c al{G }_{ \tilde\nu\t ilde \ nu}( {\ma t hr m{i } }\ome ga_2)
V ^{\mathrm{in}}_{\ti ld e\mu\a lpha} V^{\mathrm{i n}}_{\tild e \ n u\beta} \ti m es \nonumber \\ & & \h space{1em} \Gamm a_{\a lpha\bet a\gamma\d e l ta} \del ta( {\m ath rm{ i } }\ omega_1 + {\m a t hrm{ i} }\omega _2- {\mat hrm {i} }\o meg a_ 3 - {\mat hrm{i}}\ om eg a_ 4)
\ times \nonumbe r\\&& \ hspac e {1em} V^ {\ma th rm { in} ,-1}_{\ g am m a \til de \r ho}V^{ \m athrm {in} , -1} _{\delt a\tilde\s igm a }
\ ma thcal{G }_{\tilde\rho \t ilde\rho}( {\ mat hrm{i} } \ omega_3) \mathcal{G}_{\tilde\si g ma\tild e\s igma} ({\m athrm{i | }}\omega_3 -_{\mathrm{i}}\omega_4)
\times_\nonumber \\ & &_\hspace{1em}
__\mathcal{G}_{\gamma\gamma'}({\mathrm{i}}\omega_3) \mathcal{G}_{\delta\delta'}({\mathrm{i}}\omega_4)
__ V^{\mathrm{rn},-1}_{\gamma' i}_V^{\mathrm{rn},-1}_{\delta' j}.
_ \label{eq:TwoParticleGF:connected:frequency:initial}\end{aligned}$$ In order_to solve the_frequency_integral analytically, we again transform into the eigenbasis of $H_0 + \Sigma$. Eq. (\[eq:TwoParticleGF:connected:frequency:initial\]) now_reads_$$\begin{aligned}
&_&_\mathcal{G}^{(2,c)}_{ijij}({\mathrm{i}}\omega_1,_{\mathrm{i}}\omega_2, {\mathrm{i}}\omega_3, {\mathrm{i}}\omega_4) =
_ - 2\pi \sum_{\alpha\beta\gamma\delta} \sum_{\tilde\mu\tilde\nu\tilde\rho\tilde\sigma}
_ _ V^{\mathrm{ri}}_{i\tilde\mu} V^{\mathrm{ri}}_{j\tilde\nu}
\times \nonumber \\ &_&_\hspace{1em}
_\mathcal{G}_{\tilde\mu\tilde\mu}({\mathrm{i}}\omega_1) \mathcal{G}_{\tilde\nu\tilde\nu}({\mathrm{i}}\omega_2)
V^{\mathrm{in}}_{\tilde\mu\alpha} V^{\mathrm{in}}_{\tilde\nu\beta}
\times \nonumber_\\ & & \hspace{1em}
_\Gamma_{\alpha\beta\gamma\delta} \delta({\mathrm{i}}\omega_1 +_{\mathrm{i}}\omega_2_-_{\mathrm{i}}\omega_3 - {\mathrm{i}}\omega_4)
\times_\nonumber \\ & & \hspace{1em}
_ V^{\mathrm{in},-1}_{\gamma\tilde\rho} V^{\mathrm{in},-1}_{\delta\tilde\sigma}
\mathcal{G}_{\tilde\rho\tilde\rho}({\mathrm{i}}\omega_3)_\mathcal{G}_{\tilde\sigma\tilde\sigma}({\mathrm{i |
Occam algorithm for ${\bf R}$.
Let a representation system ${\bf R}$ be polynomially closed under either exceptions or majority-of-three, or both. Then ${\bf R}$ is deterministically polynomially pac-learnable iff there is a polynomial time Occam algorithm for ${\bf R}$.
[*Example.*]{} Consider threshold circuits, acyclic circuits whose nodes compute threshold functions of the form $a_1x_1 + a_2x_2 + \cdots +a_nx_n \geq \delta$, $x_i \in \{0,1\}, a_i,\delta \in N$ (note that no expressive power is gained by allowing rational weights and threshold). A simple way of representing circuits over the binary alphabet is to number each node and use [*prefix-free encodings*]{} of these numbers. For instance, encode $i$ as $1^{|\mbox{bin}(i)|}0\mbox{bin}(i)$, the binary representation of $i$ preceded by its length in unary. A complete node encoding then consists of the encoded index, encoded weights, threshold, encoded degree, and encoded indices of the nodes corresponding to its inputs. A complete circuit can be encoded with a node-count followed by a sequence of node-encodings. For this representation, a majority-of-three algorithm is easily constructed that renumbers two of its three input representations, and combines the three by adding a 3-input node computing the majority function $x_1+x_2+x_3 \geq 2$. It is clear that under this representation, the system of threshold circuits are polynomially closed under majority-of-three. On the other hand they are not closed under exceptions, or under the exception lists of [@board].
[*Example.*]{} Let $h_1, h_2, h_3$ be 3 $k$-DNF formulas. Then ${\mbox{MAJ}}(h_1,h_2,h_3) = (h_1 \wedge h_2) \vee (h_2 \wedge h_3) \vee
(h_3 \wedge h_1)$ which can be expanded into a $2k$-DNF formula. This is not good enough for Theorem \[maj\], but it allows us to conclude that pac-learnability of $k$-DNF implies compression of $k$-DNF into $2k$-DNF. | Occam algorithm for $ { \bf R}$.
Let a representation system $ { \bf R}$ be polynomially closed under either exception or majority - of - three, or both. Then $ { \bf R}$ is deterministically polynomially pac - learnable iff there be a polynomial time Occam algorithm for $ { \bf R}$.
[ * Example. * ] { } Consider doorsill circuits, acyclic circuits whose node calculate threshold functions of the form $ a_1x_1 + a_2x_2 + \cdots + a_nx_n \geq \delta$, $ x_i \in \{0,1\ }, a_i,\delta \in N$ (note that no expressive power is profit by allowing rational weights and threshold). A simple way of represent circuits over the binary alphabet is to number each lymph node and use [ * prefix - free encoding * ] { } of these numbers. For example, encode $ i$ as $ 1^{|\mbox{bin}(i)|}0\mbox{bin}(i)$, the binary representation of $ i$ precede by its length in unary. A complete node encoding then consist of the encoded index, encoded weights, threshold, encoded degree, and encoded indices of the nodes corresponding to its inputs. A complete circuit can be encoded with a node - count followed by a sequence of lymph node - encoding. For this representation, a majority - of - three algorithm is easily constructed that renumbers two of its three stimulation representations, and combines the three by adding a 3 - stimulation node computing the majority function $ x_1+x_2+x_3 \geq 2$. It is clear that under this theatrical performance, the system of threshold circuits are polynomially closed under majority - of - three. On the other hand they are not closed under exceptions, or under the exception lists of [ @board ].
[ * Example. * ] { } Let $ h_1, h_2, h_3 $ be 3 $ k$-DNF formulas. Then $ { \mbox{MAJ}}(h_1,h_2,h_3) = (h_1 \wedge h_2) \vee (h_2 \wedge h_3) \vee
(h_3 \wedge h_1)$ which can be extend into a $ 2k$-DNF formula. This is not dear enough for Theorem \[maj\ ], but it allows us to conclude that pac - learnability of $ k$-DNF implies compression of $ k$-DNF into $ 2k$-DNF. | Ocfam algorithm for ${\bf R}$.
Leu a representatiou systek ${\bf R}$ be polyvomially closed under either ezceptuons or majority-of-thred, or both. Then ${\bf R}$ iw deterministically polynklialnb pac-learnable lff there iv a polynomial tkmz Occam algorithm for ${\bf R}$.
[*Example.*]{} Cogsider yhgeshold circuijs, acjcjic dprguits whose nodes compute threshkld funbtions of the fork $a_1x_1 + a_2x_2 + \cdots +a_nx_n \geq \dflta$, $x_i \in \{0,1\}, a_i,\delta \in N$ (note thaj no wxpressive puwer is gapued by allosing rational weights and thresfold). C simple wai kf tepresenting circlits over the binary dlphabey is to number earh nide and use [*prefix-frex encodings*]{} of these numbers. Xox instance, encode $i$ aw $1^{|\nbox{bhn}(i)|}0\mtox{bkb}(i)$, ghe bmnady repgesxntation of $i$ preceded by its length in umawj. A complete hode egcjding then consists of the encoded indeq, endoded weights, threshold, encoded degree, and ejcoded inqices of the nodes corresponding to its inputs. A womplxtd cnvguit xaj be encoded with a node-count followed by a sqsutncv of node-encodingf. For this teorrfentation, a mxjoritv-kf-fhree algorithm is easily consrructed trat tenumbers two of its three unput represvntarions, and combines the three yy addong a 3-input node computing tke majkrity functlon $x_1+x_2+x_3 \gsd 2$. It is clear tfat utder this representation, tre system of chreshola citcuits wre polynolialln closed under majogity-oy-threa. On the ohher hand they are not closed uiver exceptionx, mr lnder the exceition lists of [@foard].
[*Example.*]{} Lgt $h_1, h_2, h_3$ ye 3 $k$-DVF formulaa. Then ${\kbox{MAJ}}(h_1,h_2,h_3) = (h_1 \wedge h_2) \vea (h_2 \wedge h_3) \vxe
(h_3 \wedge h_1)$ wyich can be dxpanded into s $2k$-DNF fogmbla. This us not good enough fot Fheorem \[maj\], but nu aolows us to convluae ehwt pww-learnabilitf of $k$-DVG impuies comprewrion of $k$-DNF into $2k$-DNF. | Occam algorithm for ${\bf R}$. Let a ${\bf be polynomially under either exceptions ${\bf is deterministically polynomially iff there is polynomial time Occam algorithm for ${\bf [*Example.*]{} Consider threshold circuits, acyclic circuits whose nodes compute threshold functions of the $a_1x_1 + a_2x_2 + \cdots +a_nx_n \geq \delta$, $x_i \in \{0,1\}, a_i,\delta \in (note no power gained by allowing rational weights and threshold). A simple way of representing circuits over the binary is to number each node and use [*prefix-free of these numbers. For encode $i$ as $1^{|\mbox{bin}(i)|}0\mbox{bin}(i)$, the representation $i$ preceded its in A complete node then consists of the encoded index, encoded weights, threshold, encoded degree, and encoded indices of the nodes to its complete circuit be with node-count followed by of node-encodings. For this representation, a easily constructed that renumbers two of its three representations, and the three by adding a 3-input computing the majority function $x_1+x_2+x_3 \geq 2$. It clear that under this representation, the system of threshold circuits are polynomially closed under majority-of-three. other hand they are closed under exceptions, under exception of [*Example.*]{} Let h_2, h_3$ be 3 $k$-DNF formulas. Then ${\mbox{MAJ}}(h_1,h_2,h_3) = (h_1 \wedge \vee (h_2 \wedge h_3) \vee (h_3 \wedge h_1)$ which can into $2k$-DNF formula. This not good enough for \[maj\], it allows us to pac-learnability $k$-DNF $k$-DNF $2k$-DNF. | Occam algorithm for ${\bf R}$.
Let a rEpresentatIon sySteM ${\bf r}$ bE polYnomIally closed undER eitHer exceptions or majoritY-of-thReE, Or boTH. THen ${\bf r}$ is deteRMiNIStiCaLlY poLyNOmIally Pac-LearnabLe iff there Is a PoLynomial time oCcAm algorithM foR ${\bf R}$.
[*Example.*]{} COnsIder thReShoLD circUitS, acycLic cirCUits whOse nodes cOmPUte thrEShold fuNCTiOns oF the form $a_1x_1 + a_2x_2 + \cdots +A_Nx_N \Geq \delta$, $x_i \in \{0,1\}, a_i,\Delta \iN N$ (NOtE THat No eXpressive pOwEr is gAIned by aLLoWING raTIonal weights aNd threshold). a SimPle way Of RepREsentiNg cirCuITs oVer the binarY alpHabet is to Number EAch node ANd use [*prEfix-frEe eNcoDingS*]{} Of ThEse NuMBerS. foR inSTanCe, encode $I$ aS $1^{|\mBox{biN}(i)|}0\mbOX{BIN}(i)$, thE biNary RepreSentation of $i$ pRecEded BY itS lengTh in uNary. a cOmpleTe node EncodInG then consists of The eNcoded indEx, eNcOdeD wEightS, ThreshOld, EncOded degRee, and eNCodEd INDIcEs of the nodes corresPoNDInG to its inPuts. A cOMpLeTE circuit CaN be EncoDED with A nodE-CoUnt folloWed by a SEqUeNce of noDe-EncodiNgS. FoR thIs repREsenTation, A majoritY-of-thREe algorithm is eASily constructED tHAT rENumbErs Two of its thrEe inPUt rePresENtAtiONs, and CombiNeS ThE Three by adding a 3-input NoDe compUting The majority fuNction $x_1+x_2+x_3 \gEQ 2$. iT is clear That UNdER this representAtion, The system oF ThresholD circUits are pOlynomialLY Closed unDer MajOriTy-oF-THrEe. On the other hAND theY aRe not clOseD under eXcePtiOns, Or uNdEr the excePtion lisTs Of [@BoArD].
[*ExAmple.*]{} lEt $h_1, h_2, h_3$ be 3 $k$-dNf foRmUlaS. Then ${\MBox{MAJ}}(H_1,h_2,h_3) = (h_1 \wEdge H_2) \vEe (H_2 \WedGe h_3) \vee
(h_3 \WEdGE H_1)$ whiCh CaN be eXpaNdEd intO a $2k$-Dnf foRmula. ThIs is not goOd eNOugh FoR THeorem \[mAj\], but it allows Us To conclude ThAt pAc-learNABility of $K$-DNF implies compression oF $K$-DNF intO $2k$-DnF. | Occam algorithm for ${\bf R}$.
Let a re pre sen ta tion sys tem ${\bf R}$b e po lynomially closed unde r eit he r exc e pt ionsor majo r it y - of- th re e,or bo th. T hen ${\bfR}$ is det erm in istically po l yn omially pa c-l earnable iff th ere is a po l ynomi altimeOccama lgorit hm for ${ \b f R}$.[*Examp l e .* ]{}Consider threshol d c i rcuits, acycli c circ ui t sw h ose no des comput ethres h old fun c ti o n s of the form $a_1 x_1 + a_2x_ 2 +\cdots + a_n x _n \ge q \de lt a $,$x_i \in \{ 0,1\ }, a_i,\d elta \ i n N$ (n o te that no ex pre ssi ve p o we risga i ned by al l owi ng ratio na lweigh ts a n d t hres hol d).A sim ple way of re pre sent i ngcircu its o verth e bin ary al phabe tis to number ea ch n ode and u se[* pre fi x-fre e encod ing s*] {} of t hese nu m ber s. F o rinstance, encode $ i$ a s$1^{|\mb ox{bin } (i )| } 0\mbox{b in }(i )$,t h e bin aryr ep resentat ion of $i $precede dby its l eng thin un a ry.A comp lete nod e enc o ding then cons i sts of the en c od e d i n dex, en coded weigh ts,t hres hold , e nco d ed de gree, a n de ncoded indices of t he nodes corr esponding toits inputs . A complet e ci r cu i t can be encod ed wi th a node- c ount fol lowed by a se quence of n ode-enco din gs. Fo r t h i srepresentatio n , a m aj ority-o f-t hree al gor ith m i s e as ily const ructed t ha tre nu mbe rs tw o of itsth ree i npu t rep r esenta tions , an dco m bin es thet hr e e byad di ng a 3- in put n odec omp uting t he majori tyf unct io n$x_1+x_ 2+x_3 \geq 2$ .It is clea rtha t unde r this rep resentation, the system of thre sho ld ci rcui ts are po lyn omiall y c l osed u nder m ajori ty -of - t hree. O nthe o ther handt h eyare n ot clo sed und er exceptions, oru nde r the excepti onlist s of [@ b oa r d].
[ *Ex a m ple.*]{} Let $h _1, h_2, h _3 $ b e 3 $k$-DN F fo rm ulas. T hen ${\ mbox{ M AJ}}(h_ 1,h_2,h_3 ) = (h_1\w edge h _2) \vee (h_2 \wedgeh_3) \vee (h_3\ we dge h _1) $ whic hcan be e xpande d in to a$2k$-D NF formu la. T hi s is not good enough for Theore m \[ma j\],but it allow s u s to conclude tha t pac-lear nab ili ty of $k $ -DNFimpl i es co m press iono f $k$-DNF in to$ 2 k$ -DNF. | Occam_algorithm for_${\bf R}$.
Let a representation_system ${\bf_R}$_be polynomially_closed_under either exceptions_or majority-of-three, or_both. Then ${\bf R}$_is deterministically polynomially_pac-learnable_iff there is a polynomial time Occam algorithm for ${\bf R}$.
[*Example.*]{} Consider threshold circuits,_acyclic_circuits whose_nodes_compute_threshold functions of the form_$a_1x_1 + a_2x_2 + \cdots_+a_nx_n \geq_\delta$, $x_i \in \{0,1\}, a_i,\delta \in N$ (note_that_no expressive power_is gained by allowing rational weights and threshold). A_simple way of representing circuits over_the binary alphabet_is_to_number each node and_use [*prefix-free encodings*]{} of these numbers._For instance, encode $i$ as $1^{|\mbox{bin}(i)|}0\mbox{bin}(i)$,_the binary representation of $i$ preceded by_its length in unary. A complete_node encoding then consists of_the encoded_index, encoded weights, threshold, encoded_degree, and encoded_indices of_the nodes corresponding_to its inputs. A complete circuit_can be encoded_with a node-count followed by a_sequence_of node-encodings. For_this_representation,_a majority-of-three_algorithm is easily_constructed_that renumbers_two_of its three input representations, and_combines_the three by adding a 3-input node_computing the majority function_$x_1+x_2+x_3_\geq 2$. It is_clear that under this representation,_the system of threshold circuits are_polynomially closed_under majority-of-three._On the other hand they are not closed under exceptions, or_under the exception lists of [@board].
[*Example.*]{}_Let $h_1, h_2, h_3$_be 3_$k$-DNF_formulas. Then ${\mbox{MAJ}}(h_1,h_2,h_3)_=_(h_1 \wedge_h_2) \vee (h_2 \wedge h_3) \vee
(h_3 \wedge_h_1)$ which_can be expanded into a $2k$-DNF_formula. This is not_good_enough for Theorem \[maj\], but it allows_us to conclude that pac-learnability of_$k$-DNF implies compression of $k$-DNF_into_$2k$-DNF. |
above).
Let us prove that if $X^{**}$ has the metric approximation property then $D((X^{**},k_1,\cdots,k_n)\geq D(X,k_1,\cdots,k_n)$. Given $\varepsilon >0$, let $\{P_j\}_{j=1}^n$ be a set of polynomials on $X^{**}$ with $deg(P_j)=k_j$ such that $$D(X^{**},k_1,\cdots,k_n)\left \Vert \prod_{j=1}^{n} P_{j} \right \Vert \leq (1+\varepsilon)\prod_{j=1}^{n} \left \Vert P_{j} \right \Vert.\nonumber$$
Analogous to the proof of Theorem \[main thm\], since $X^{**}$ has the metric approximation, we can construct finite type polynomials $Q_1,\cdots,Q_n$ on $X^{**}$ with $deg(Q_j)=k_j$, $\Vert Q_j \Vert_K \geq \Vert P_j \Vert (1-\varepsilon)^2$ for some finite set $K\subseteq B_{X^{**}}$ and that $$D(X^{**},k_1,\cdots,k_n)\left \Vert \prod_{j=1}^{n} Q_{j} \right \Vert < (1+\varepsilon)\prod_{j=1}^{n} \left \Vert P_{j} \right \Vert. \nonumber$$
Suppose that $Q_j=\sum_{t=1}^{m_j}(\psi_{j,t})^{r_{j,t}}$ and consider the spaces $N=\rm{span} \{K\}$ and $M=\rm{span} \{\psi_{j,t} \}$. By the principle of local reflexivity (see [@D]), applied to $X^*$ (thinking $N$ as a subspaces of $(X^*)^*$ and $M$ as a subspaces of $(X^*)^{**}$), there is an $(1+\varepsilon)-$isomorphism $T:M\rightarrow X^*$ such that $$JT(\psi)(x)=\psi(x) \mbox{ } \forall x\in N, \mbox{ } \forall \psi\in M\cap X^*=M,$$ where $J:X^*\rightarrow X^{ | above).
Let us prove that if $ X^{**}$ has the metric estimate place then $ D((X^{**},k_1,\cdots, k_n)\geq D(X, k_1,\cdots, k_n)$. feed $ \varepsilon > 0 $, let $ \{P_j\}_{j=1}^n$ be a hardening of polynomial on $ X^{**}$ with $ deg(P_j)=k_j$ such that $ $ D(X^{**},k_1,\cdots, k_n)\left \Vert \prod_{j=1}^{n } P_{j } \right \Vert \leq (1+\varepsilon)\prod_{j=1}^{n } \left \Vert P_{j } \right \Vert.\nonumber$$
Analogous to the proof of Theorem \[main thm\ ], since $ X^{**}$ has the metric estimate, we can reconstruct finite type polynomial $ Q_1,\cdots, Q_n$ on $ X^{**}$ with $ deg(Q_j)=k_j$, $ \Vert Q_j \Vert_K \geq \Vert P_j \Vert (1-\varepsilon)^2 $ for some finite rig $ K\subseteq B_{X^{**}}$ and that $ $ D(X^{**},k_1,\cdots, k_n)\left \Vert \prod_{j=1}^{n } Q_{j } \right \Vert < (1+\varepsilon)\prod_{j=1}^{n } \left \Vert P_{j } \right \Vert. \nonumber$$
Suppose that $ Q_j=\sum_{t=1}^{m_j}(\psi_{j, t})^{r_{j, t}}$ and regard the spaces $ N=\rm{span } \{K\}$ and $ M=\rm{span } \{\psi_{j, t } \}$. By the principle of local reflexivity (visualize [ @D ]), applied to $ X^*$ (thinking $ N$ as a subspaces of $ (X^*)^*$ and $ M$ as a subspaces of $ (X^*)^{**}$), there is an $ (1+\varepsilon)-$isomorphism $ T: M\rightarrow X^*$ such that $ $ JT(\psi)(x)=\psi(x) \mbox { } \forall x\in N, \mbox { } \forall \psi\in M\cap X^*=M,$$ where $ J: X^*\rightarrow X^ { | ablve).
Let us prove that if $W^{**}$ has the metric approxmmation properth then $D((X^{**},k_1,\cdots,k_n)\geq D(X,k_1,\cdotd,k_b)$. Givtu $\varepsilon >0$, let $\{P_j\}_{g=1}^n$ be a svt of poltnommals on $X^{**}$ with $dxf(P_j)=k_j$ smeh thzb $$D(X^{**},k_1,\edits,k_n)\left \Vert \prod_{j=1}^{n} P_{j} \right \Vert \lex (1+\xaxepsilon)\prod_{j=1}^{n} \left \Vert P_{j} \right \Vewt.\nonumneg$$
Analogous to jhe pgojf or Theorem \[main thm\], since $X^{**}$ has the metric approximatiom, we can construct finite hype polynomials $Q_1,\cdotd,Q_n$ on $X^{**}$ wijg $dqt(Q_j)=k_j$, $\Vert Q_g \Vert_K \geq \Vert P_j \Vett (1-\varepsilon)^2$ for some finite seg $K\suyseteq B_{X^{**}}$ abd thwj $$D(X^{**},k_1,\cdots,k_n)\oeft \Nert \prod_{j=1}^{n} Q_{m} \right \Vert < (1+\farepsilon)\prod_{m=1}^{n} \leht \Vwrt P_{j} \right \Vert. \noiumber$$
Suppose that $Q_t=\sum_{t=1}^{m_j}(\psh_{j,c})^{r_{j,t}}$ and consider the spaces $N=\rm{vpan} \{J\}$ avd $J=\rk{slan} \{\psl_{j,t} \}$. By the pdinciple of local reflexivity (xeq [@D]), applied to $X^*$ (thigkyng $N$ as a subspaces of $(X^*)^*$ and $M$ as a sutspzces of $(X^*)^{**}$), there is an $(1+\vqrepsilon)-$isomorphism $J:M\rightarrjw X^*$ such that $$JT(\psi)(x)=\psi(x) \mbox{ } \forall x\in N, \mbox{ } \forelu \pwi\ln M\zqp X^*=M,$$ where $J:X^*\rightarrow X^{ | above). Let us prove that if $X^{**}$ metric property then D(X,k_1,\cdots,k_n)$. Given $\varepsilon set polynomials on $X^{**}$ $deg(P_j)=k_j$ such that \Vert \prod_{j=1}^{n} P_{j} \right \Vert \leq \left \Vert P_{j} \right \Vert.\nonumber$$ Analogous to the proof of Theorem \[main thm\], $X^{**}$ has the metric approximation, we can construct finite type polynomials $Q_1,\cdots,Q_n$ on with $\Vert \Vert_K \Vert P_j \Vert (1-\varepsilon)^2$ for some finite set $K\subseteq B_{X^{**}}$ and that $$D(X^{**},k_1,\cdots,k_n)\left \Vert \prod_{j=1}^{n} Q_{j} \Vert < (1+\varepsilon)\prod_{j=1}^{n} \left \Vert P_{j} \right \Vert. Suppose that $Q_j=\sum_{t=1}^{m_j}(\psi_{j,t})^{r_{j,t}}$ and the spaces $N=\rm{span} \{K\}$ and \{\psi_{j,t} By the of reflexivity [@D]), applied to (thinking $N$ as a subspaces of $(X^*)^*$ and $M$ as a subspaces of $(X^*)^{**}$), there is an $T:M\rightarrow X^*$ $$JT(\psi)(x)=\psi(x) \mbox{ \forall N, } \forall \psi\in where $J:X^*\rightarrow X^{ | above).
Let us prove that if $X^{**}$ has The metric aPproxImaTioN pRopeRty tHen $D((X^{**},k_1,\cdots,k_n)\gEQ D(X,k_1,\Cdots,k_n)$. Given $\varepsilon >0$, Let $\{P_j\}_{J=1}^n$ BE a seT Of PolynOmials oN $x^{**}$ wITH $deG(P_J)=k_J$ suCh THaT $$D(X^{**},k_1,\cDotS,k_n)\left \vert \prod_{j=1}^{n} p_{j} \rIgHt \Vert \leq (1+\varEPsIlon)\prod_{j=1}^{n} \LefT \Vert P_{j} \right \verT.\nonumBeR$$
AnALogouS to The prOof of THEorem \[mAin thm\], sinCe $x^{**}$ Has the MEtric apPROxImatIon, we can construct FInITe type polynomiAls $Q_1,\cdOtS,q_n$ ON $x^{**}$ wiTh $dEg(Q_j)=k_j$, $\Vert q_j \vert_K \GEq \Vert P_J \veRT (1-\VArePSilon)^2$ for some fInite set $K\suBSetEq B_{X^{**}}$ anD tHat $$d(x^{**},k_1,\cdotS,k_n)\leFt \vErt \Prod_{j=1}^{n} Q_{j} \rigHt \VeRt < (1+\varepsiLon)\proD_{J=1}^{n} \left \VERt P_{j} \rigHt \Vert. \NonUmbEr$$
SuPPoSe ThaT $Q_J=\Sum_{T=1}^{M_j}(\Psi_{J,T})^{r_{j,T}}$ and consIdEr The spAces $n=\RM{SPan} \{K\}$ And $m=\rm{sPan} \{\psI_{j,t} \}$. By the princIplE of lOCal RefleXivitY (see [@d]), aPplieD to $X^*$ (thInkinG $N$ As a subspaces of $(X^*)^*$ And $M$ As a subspaCes Of $(x^*)^{**}$), thErE is an $(1+\VArepsiLon)-$IsoMorphisM $T:M\righTArrOw x^*$ SUCh That $$JT(\psi)(x)=\psi(x) \mbox{ } \FoRALl X\in N, \mbox{ } \Forall \PSi\In m\Cap X^*=M,$$ wheRe $j:X^*\rIghtARRow X^{ | above).
Let us prove tha t if $X^{* *}$ h asthe m etri c ap proximation pr o pert y then $D((X^{**},k_1, \cdot s, k _n)\ g eq D(X, k_1,\cd o ts , k _n) $. G ive n$ \v areps ilo n >0$,let $\{P_j \}_ {j =1}^n$ be as et of polyno mia ls on $X^{** }$with $ de g(P _ j)=k_ j$suchthat $ $ D(X^{* *},k_1,\c do t s,k_n) \ left \V e r t\pro d_{j=1}^{n} P_{j} \r i ght \Vert \leq (1+\v ar e ps i l on) \pr od_{j=1}^{ n} \lef t \VertP _{ j } \ri g ht \Vert.\non umber$$
An a log ous to t hep roof o f The or e m \ [main thm\] , si nce $X^{* *}$ ha s the me t ric app roxima tio n,we c a nco nst ru c t f i ni tet ype polynom ia ls $Q_1 ,\cd o t s , Q_n$ on $X^ {**}$ with $deg(Q_ j)= k_j$ , $\ VertQ_j \ Vert _K \geq \Vert P_j\V ert (1-\varepsi lon) ^2$ for s ome f ini te set$ K\subs ete q B _{X^{** }}$ and tha t$ $ D (X ^{**},k_1,\cdots,k _n ) \ le ft \Vert \prod _ {j =1 } ^{n} Q_{ j} \r ight \ Vert< (1 + \v arepsilo n)\pro d _{ j= 1}^{n}\l eft \V er t P _{j } \ri g ht \Vert. \nonumb er$$Suppose that $ Q _j=\sum_{t=1} ^ {m _ j }( \ psi_ {j, t})^{r_{j,t }}$a nd c onsi d er th e spac es $N =\ r m{ s pan} \{K\}$ and $M= \r m{span } \{\ psi_{j,t} \}$ . By the p r i n ciple of loc a lr eflexivity (se e [@D ]), applie d to $X^* $ (th inking $ N$ as a s u b spaces o f $ (X^ *)^ *$a n d$M$ as a subs p a cesof $(X^*) ^{* *}$), t her e i s a n $ (1 +\varepsi lon)-$is om or ph is m $ T:M\r i ghtarrow X ^*$ s uch that $$JT(\ psi)( x)=\ ps i( x ) \ mbox{ } \f o r allx\ in N,\mb ox { } \ fora l l \ psi\inM\cap X^* =M, $ $ wh er e$J:X^*\ rightarrow X^ { | above).
Let_us prove_that if $X^{**}$ has_the metric_approximation_property then_$D((X^{**},k_1,\cdots,k_n)\geq_D(X,k_1,\cdots,k_n)$. Given $\varepsilon_>0$, let $\{P_j\}_{j=1}^n$_be a set of_polynomials on $X^{**}$_with_$deg(P_j)=k_j$ such that $$D(X^{**},k_1,\cdots,k_n)\left \Vert \prod_{j=1}^{n} P_{j} \right \Vert \leq (1+\varepsilon)\prod_{j=1}^{n} \left \Vert P_{j}_\right_\Vert.\nonumber$$
Analogous to_the_proof_of Theorem \[main thm\], since_$X^{**}$ has the metric approximation,_we can_construct finite type polynomials $Q_1,\cdots,Q_n$ on $X^{**}$ with_$deg(Q_j)=k_j$,_$\Vert Q_j \Vert_K_\geq \Vert P_j \Vert (1-\varepsilon)^2$ for some finite set_$K\subseteq B_{X^{**}}$ and that $$D(X^{**},k_1,\cdots,k_n)\left \Vert_\prod_{j=1}^{n} Q_{j} \right_\Vert_<_(1+\varepsilon)\prod_{j=1}^{n} \left \Vert P_{j}_\right \Vert. \nonumber$$
Suppose that $Q_j=\sum_{t=1}^{m_j}(\psi_{j,t})^{r_{j,t}}$_and consider the spaces $N=\rm{span} \{K\}$_and $M=\rm{span} \{\psi_{j,t} \}$. By the principle_of local reflexivity (see [@D]), applied_to $X^*$ (thinking $N$ as_a subspaces_of $(X^*)^*$ and $M$ as_a subspaces of_$(X^*)^{**}$), there_is an $(1+\varepsilon)-$isomorphism_$T:M\rightarrow X^*$ such that $$JT(\psi)(x)=\psi(x) \mbox{_} \forall x\in_N, \mbox{ } \forall \psi\in M\cap_X^*=M,$$_where $J:X^*\rightarrow X^{ |
are, however, not suitable for the description of single-particle dynamics in nuclei. Following the field theoretical approach by A. Bohr and B. R. Mottelson to describe nuclear interactions between single-particle and vibrational degrees of freedom, we have undertaken important steps to build a microscopic dynamic nuclear model. In connection to that, one important issue that needs to be better understood is the renormalization of the effective interaction in the particle-vibration approach. One possible way to renormalize the interaction is the so called [*subtraction method*]{}. In this contribution we will implement the [*subtraction method*]{} for the first time in our model and study its consequences.'
address: |
$^{1}$Dipartimento di Fisica, Università degli Studi di Milano, Via Celoria 16, I-20133 Milano, Italy\
$^2$INFN, Sezione di Milano, Via Celoria 16, I-20133 Milano, Italy\
$^3$ELI-NP, Horia Hulubei National Institute for Physics and Nuclear Engineering, 30 Reactorului Street, RO-077125, Bucharest-Magurele, Romania
author:
- 'X. Roca-Maza$^{1,2}$, Y.F. Niu$^{2,3}$, G. Colò$^{1,2}$, P. F. Bortignon$^{1,2}$'
bibliography:
- 'bibliography.bib'
title: 'Towards a self-consistent dynamical nuclear model'
---
Introduction {#intro}
============
In our research group, we are interested in developing a microscopic nuclear structure model that can be eventually applied also to nuclear reactions. Such an achievement, would solve a long standing problem in theoretical nuclear physics [@nazarewicz2016].
We have recently developed two different energy density functionals (EDFs) [@ddmed; @sami] based on the Density Functional Theory [@HK; @KS2; @engel2011]. In nuclear physics EDFs are commonly derived in an approximate way from a phenomenological interaction solved via the Hartree (H) or Hartree-Fock (HF) approximations [@bender2003; @vretenar2005]. Such type of models are very helpful in understanding different nuclear properties such as nuclear masses and sizes as well as the excitation energy of nuclear collective states such as giant resonances. In the latter case a minimal extension of the theory is | are, however, not suitable for the description of single - particle dynamics in core. take after the field theoretical approach by A. Bohr and B. R. Mottelson to describe nuclear interaction between single - particle and vibrational degree of exemption, we have undertaken important steps to build a microscopic dynamic nuclear mannequin. In connection to that, one important exit that need to be better understood is the renormalization of the effective interaction in the particle - vibration approach. One potential way to renormalize the interaction is the so call [ * subtraction method * ] { }. In this contribution we will implement the [ * subtraction method acting * ] { } for the first fourth dimension in our model and learn its consequences.'
address: |
$ ^{1}$Dipartimento di Fisica, Università degli Studi di Milano, Via Celoria 16, I-20133 Milano, Italy\
$ ^2$INFN, Sezione di Milano, Via Celoria 16, I-20133 Milano, Italy\
$ ^3$ELI - NP, Horia Hulubei National Institute for Physics and Nuclear Engineering, 30 Reactorului Street, RO-077125, Bucharest - Magurele, Romania
writer:
-' X. Roca - Maza$^{1,2}$, Y.F. Niu$^{2,3}$, G. Colò$^{1,2}$, P. F. Bortignon$^{1,2}$'
bibliography:
-' bibliography.bib'
title:' Towards a self - consistent dynamical nuclear model'
---
Introduction { # intro }
= = = = = = = = = = = =
In our research group, we are interested in developing a microscopic nuclear structure model that can be eventually applied also to nuclear reactions. Such an achievement, would solve a retentive standing problem in theoretical nuclear physics [ @nazarewicz2016 ].
We have recently modernize two unlike energy density functionals (EDFs) [ @ddmed; @sami ] based on the Density Functional Theory [ @HK; @KS2; @engel2011 ]. In nuclear physics EDFs are commonly derived in an approximate way from a phenomenological interaction solved via the Hartree (H) or Hartree - Fock (HF) approximation [ @bender2003; @vretenar2005 ]. Such type of models are very helpful in understanding different nuclear properties such as nuclear masses and size as well as the excitation energy of nuclear collective states such as giant resonance. In the latter case a minimal extension of the theory is | arf, however, not suitable fur the descriptnin of vingle-larticle dynamics in nuclei. Followinj thw fieod theoretical approacf by A. Bojr and B. R. Mirtelson to describc nucmcar iutxractions betwegn single-pardicle and vibrdtkoual degrees of freedom, we have undereaken ikplrtant steps tj buplq a jpcvoscopic dynamic nuclear model. Ih connebtion to that, one important issue that needd to be better understlod is the tsnownalization ow the effebcive interadtion in the particle-vibration xpprocch. One poswivle fay to renocmalizv the interacbpon is dhe so valled [*subtracbion ketyod*]{}. In this contributmon we will implemenj the [*subtsaetion method*]{} for the diest thme hn oje mudem end study ita consequehces.'
address: |
$^{1}$Dipartimento do Spxica, Universjtà degji Studi di Milano, Via Celoria 16, I-20133 Milano, Ptalg\
$^2$INFN, Sezione di Mioano, Via Celoria 16, I-20133 Mllano, Itajy\
$^3$ELI-NP, Horia Hulubei National Institute for [hysirs anb Nucldqr Engineering, 30 Reactorului Street, RO-077125, Bucharest-Izgirvle, Romania
author:
- 'X. Roca-Maza$^{1,2}$, Y.V. Myu$^{2,3}$, G. Colò$^{1,2}$, P. F. Cortigukn$^{1,2}$'
gibliography:
- 'biblilgraphy.fib'
tirle: 'Towarqs a self-consistent dynamical nyclear model'
---
Pntriduction {#intro}
============
In obr research yroup, ee arr interested in developnng a jicroscopic nuclear agructure model tfat cdn be evtvtually applied ajso to nurlear reactiuns. Xuch ag achievemfnt, would solve a long shandiug prmblem in tjeoretical nuclear physics [@nazacxwicz2016].
We have tecantny develjped bwo different egergy density yunctioncls (EDWs) [@ddmed; @szmi] basxd on the Degsity Functiotwl Theory [@HK; @KS2; @engej2011]. In nucoear phhrics EDFs are vommonly btrived in qn approximate way fruj a phenomenoloyncql interaction xolxed vpa vhe Hwstree (H) or Hdrtrde-Fuvk (HF) approximatlonr [@bemder2003; @vretenar2005]. Such tfpe kf models are very hclpful in understwnding differrnt nuclear properuies snch as nuclrar masses and sizes as well as tge excitahiok energy of ntcleqr collectivz states such as giant resonances. In the latter case a minumal extension of tke theory is | are, however, not suitable for the description dynamics nuclei. Following field theoretical approach R. to describe nuclear between single-particle and degrees of freedom, we have undertaken steps to build a microscopic dynamic nuclear model. In connection to that, one issue that needs to be better understood is the renormalization of the effective in particle-vibration One way to renormalize the interaction is the so called [*subtraction method*]{}. In this contribution we will the [*subtraction method*]{} for the first time in model and study its address: | $^{1}$Dipartimento di Fisica, degli di Milano, Celoria I-20133 Italy\ $^2$INFN, Sezione Milano, Via Celoria 16, I-20133 Milano, Italy\ $^3$ELI-NP, Horia Hulubei National Institute for Physics and Nuclear Engineering, Reactorului Street, Romania author: 'X. Y.F. G. Colò$^{1,2}$, P. bibliography: - 'bibliography.bib' title: 'Towards a model' --- Introduction {#intro} ============ In our research we are in developing a microscopic nuclear structure that can be eventually applied also to nuclear Such an achievement, would solve a long standing problem in theoretical nuclear physics [@nazarewicz2016]. We developed two different energy functionals (EDFs) [@ddmed; based the Functional [@HK; @KS2; In nuclear physics EDFs are commonly derived in an approximate way a phenomenological interaction solved via the Hartree (H) or Hartree-Fock [@bender2003; Such type of are very helpful in different properties such as nuclear sizes well energy nuclear states such as giant In the latter case a extension of the theory | are, however, not suitable for tHe descriptIon of SinGle-PaRticLe dyNamics in nuclei. fOlloWing the field theoreticaL apprOaCH by A. bOhR and B. r. MottelSOn TO DesCrIbE nuClEAr InterActIons betWeen single-ParTiCle and vibratIOnAl degrees oF frEedom, we have uNdeRtaken ImPorTAnt stEps To buiLd a micROscopiC dynamic nUcLEar modEL. In connECTiOn to That, one important iSSuE That needs to be bEtter uNdERsTOOd iS thE renormaliZaTion oF The effeCTiVE INteRAction in the paRticle-vibraTIon ApproaCh. one POssiblE way tO rENorMalize the inTeraCtion is thE so calLEd [*subtrACtion meThod*]{}. In ThiS coNtriBUtIoN we WiLL imPLeMenT The [*SubtractIoN mEthod*]{} For tHE FIRst tIme In ouR modeL and study its cOnsEqueNCes.'
AddreSs: |
$^{1}$DipArtiMeNto di fisica, univeRsItà degli Studi di milaNo, Via CeloRia 16, i-20133 MIlaNo, italy\
$^2$inFN, SezIonE di milano, VIa CelorIA 16, I-20133 MIlANO, itAly\
$^3$ELI-NP, Horia HulubEi nATiOnal InstItute fOR PHySIcs and NuClEar engiNEEring, 30 reacTOrUlui StreEt, RO-077125, BuCHaReSt-MagurElE, RomanIa
AutHor:
- 'x. Roca-mAza$^{1,2}$, Y.f. Niu$^{2,3}$, G. COlò$^{1,2}$, P. F. BorTignoN$^{1,2}$'
Bibliography:
- 'biBLiography.bib'
tITlE: 'tOwARds a SelF-consistent DynaMIcal NuclEAr ModEL'
---
IntrOductIoN {#InTRo}
============
In our research grouP, wE are inTeresTed in developiNg a microscOPIC nuclear StruCTuRE model that can bE evenTually applIEd also to NucleAr reactiOns. Such an ACHievemenT, woUld SolVe a LONg Standing problEM In thEoRetical NucLear phySicS [@naZarEwiCz2016].
we have recEntly devElOpEd TwO diFfereNT energy dEnSitY fUncTionaLS (EDFs) [@dDmed; @sAmi] bAsEd ON thE DensitY fuNCTionAl thEory [@hK; @Ks2; @eNgel2011]. IN nucLEar Physics eDFs are coMmoNLy deRiVeD in an apProximate way fRoM a phenomenOlOgiCal intERAction soLved via the Hartree (H) or HarTRee-Fock (hF) aPproxImatIons [@bendeR2003; @vrEtenar2005]. sucH Type of Models Are veRy HelPFUl in uNDErStaNdIng differeNT NucLear pRoPertIes such As nuclear masses and SIzeS as well as the eXciTatiON EnErgY Of NUclEaR ColLECtive states such As giant resOnANcEs. In the latTEr cAsE a minimAl extenSion oF The theoRy is | are, however, not suitabl e for thedescr ipt ion o f si ngle -particle dyna m icsin nuclei. Following t he fi el d the o re tical approa c hb y A. B oh r a nd B. R. M ott elson t o describe nu cl ear interact i on s betweensin gle-particle an d vibr at ion a l deg ree s offreedo m , we h ave under ta k en imp o rtant s t e ps tobuild a microscop i cd ynamic nuclear model .I nc o nne cti on to that ,one i m portant is s u e th a t needs to be better und e rst ood is t her enorma lizat io n of the effect iveinteracti on int he part i cle-vib ration ap pro ach. On epos si b lew ay to ren ormalize t he inte ract i o n is t heso c alled [*subtractio n m etho d *]{ }. In this con tr ibuti on wewillim plement the [*s ubtr action me tho d* ]{} f or th e first ti mein ourmodel a n d s tu d y it s consequences.'
a dd r e ss : |
$^{1}$ D ip ar t imento d iFis ica, U niver sità de gli Stud i di M i la no , Via C el oria 1 6, I- 201 33 Mi l ano, Italy \
$^ 2$INF N , Sezione di M i lano, Via Cel o ri a 16 , I-2 013 3 Milano, I taly \
$^3 $ EL I-N P , Hor ia Hu lu b ei National Institutefo r Phys ics a nd Nuclear En gineering, 3 0 Reactor ului St r eet, RO-077125 , Buc harest-Mag u rele, Ro mania
author:
- 'X. Ro c a -Maza$^{ 1,2 }$, Y. F.N i u$ ^{2,3}$, G. C o l ò$^{ 1, 2}$, P. F. Bortig non $^{ 1,2 }$'
b ibliograp hy:
- 'b ib li og ra phy .bib' title: ' To war ds aself- c onsist ent d ynam ic al nuc lear mo d el ' ---
I nt rodu cti on {#in tro} === ======= ==
In ou r r e sear ch g roup, w e are interes te d in devel op ing a mic r o scopic n uclear structure modelt hat can be even tual ly applie d a lso to nu c lear r eactio ns. S uc h a n achie v e me nt, w ould solve a lo ng st an ding proble m in theoretical n u cle ar physics [@ naz arew i c z2 016 ] .Weha v e r e c ently developed two diffe re n tenergy den s ity f unction als (ED Fs) [ @ ddmed;@sami] ba sed on th eDens i t y F unctionalTheory [ @HK; @KS2 ; @eng e l2 011]. In nucle ar ph ysics EDFsa recommo nly de ri ved in an a pp roximate way from a phenomenolo gicalinter act ion solve d v i a t he Hartre e (H ) or Hartr ee- Foc k (HF ) a p proxi mati o ns [@ b ender 2003 ; @vretena r 20 05] . Su ch type ofm o d els arever y helpf ul i n understanding d i fferent nuclea r pr o p ert ies such a s nuclear mass esan d sizes as w ell as theexcitati on energ y of n uclear collec t i ve states suc h a s giant r eso na n ces. In t he latter cas ea mini mal ex t ensi o n of the theory i s | are,_however, not_suitable for the description_of single-particle_dynamics_in nuclei._Following_the field theoretical_approach by A._Bohr and B. R._Mottelson to describe_nuclear_interactions between single-particle and vibrational degrees of freedom, we have undertaken important steps to_build_a microscopic_dynamic_nuclear_model. In connection to that,_one important issue that needs_to be_better understood is the renormalization of the effective_interaction_in the particle-vibration_approach. One possible way to renormalize the interaction is_the so called [*subtraction method*]{}. In_this contribution we_will_implement_the [*subtraction method*]{} for_the first time in our model_and study its consequences.'
address: |
_ $^{1}$Dipartimento di Fisica, Università degli_Studi di Milano, Via Celoria 16,_I-20133 Milano, Italy\
_ $^2$INFN,_Sezione di Milano, Via Celoria_16, I-20133 Milano,_Italy\
_ $^3$ELI-NP,_Horia Hulubei National Institute for Physics_and Nuclear Engineering,_30 Reactorului Street, RO-077125, Bucharest-Magurele, Romania
author:
-_'X._Roca-Maza$^{1,2}$, Y.F. Niu$^{2,3}$,_G._Colò$^{1,2}$,_P. F._Bortignon$^{1,2}$'
bibliography:
- 'bibliography.bib'
title: 'Towards_a_self-consistent dynamical_nuclear_model'
---
Introduction {#intro}
============
In our research group, we_are_interested in developing a microscopic nuclear structure_model that can be_eventually_applied also to nuclear_reactions. Such an achievement, would_solve a long standing problem in_theoretical nuclear_physics [@nazarewicz2016].
We_have recently developed two different energy density functionals (EDFs) [@ddmed; @sami]_based on the Density Functional Theory_[@HK; @KS2; @engel2011]. In_nuclear physics_EDFs_are commonly derived_in_an approximate_way from a phenomenological interaction solved via_the Hartree_(H) or Hartree-Fock (HF) approximations [@bender2003;_@vretenar2005]. Such type of_models_are very helpful in understanding different_nuclear properties such as nuclear masses_and sizes as well as_the_excitation_energy of nuclear collective states_such as giant resonances. In the_latter case a_minimal extension of the theory is |
magnitude* $|M_\ast|$ of a finitely presented graded persistence module $M_\ast$ is defined by $$|M_\ast|
=
\sum_{k=0}^{\infty}
\sum_{i=1}^{m_k}
(-1)^k (e^{-a_{k,i}} - e^{-b_{k,i}})$$ where $M_\ast$ has bars $[a_{k,1},b_{k,1}),\ldots,[a_{k,m_k},b_{k,m_k})$ in degree $k\geq 0$. Note that a bar $[a,b)$ makes a contribution of $\pm(e^{-a} - e^{-b})$ to the magnitude, so that longer bars make a greater contribution, in line with the general philosophy of persistent homology.
Persistent magnitude has good formal properties: we show that it is additive with respect to exact sequences, and that the magnitude of a tensor product of persistence modules is the product of the magnitudes of the factors, so long as the tensor product is understood in an appropriate derived sense.
Now suppose that we have a persistent homology theory defined for some class of mathematical objects, for example the Rips homology of metric spaces. By applying persistent magnitude to the persistent homology, we obtain a new numerical invariant of the mathematical objects in question. Our first example of this process is the case of the sublevel set persistent homology of Morse functions, where the resulting magnitude invariant is a (signed and weighted) count of the critical points of the original function.
Our most detailed example of persistent magnitude in action is the *Rips magnitude*. This is the numerical invariant of finite metric spaces obtained by taking the persistent magnitude of the Rips homology, and is given by the weighted simplex-count $$|X|_\mathrm{Rips}
=
\sum_{\emptyset\neq A\subseteq X}(-1)^{\#A-1}e^{-\operatorname{diam}(A)}.$$ We compute the Rips magnitude of cycle graphs with their path, Euclidean and geodesic metrics. In each case they are determined by a number-theoretical formula reminiscent of the sum of divisors function.
In the original setting of magnitude, defining the magnitude of infinite metric spaces is not straightforward: the simplest method is to take the supremum of the magnitude of all finite subspaces of the given infinite metric space, but there are alternatives, and currently the theory only works well in the | magnitude * $ |M_\ast|$ of a finitely presented graded persistence faculty $ M_\ast$ is define by $ $ |M_\ast|
=
\sum_{k=0}^{\infty }
\sum_{i=1}^{m_k }
(-1)^k (e^{-a_{k, i } } - e^{-b_{k, i}})$$ where $ M_\ast$ has bar $ [ a_{k,1},b_{k,1}),\ldots,[a_{k, m_k},b_{k, m_k})$ in academic degree $ k\geq 0$. Note that a bar $ [ a, b)$ makes a contribution of $ \pm(e^{-a } - e^{-b})$ to the order of magnitude, so that long bar make a greater contribution, in tune with the general philosophy of dogged homology.
haunting magnitude has good formal properties: we prove that it is additive with respect to exact succession, and that the magnitude of a tensor product of persistence modules is the merchandise of the magnitudes of the factors, so long as the tensor product is understood in an appropriate derive sense.
Now suppose that we have a persistent homology hypothesis specify for some class of mathematical objects, for example the Rips homology of metric spaces. By applying persistent magnitude to the haunting homology, we obtain a new numeral invariant of the numerical object in question. Our first example of this process is the case of the sublevel set persistent homology of Morse routine, where the resulting magnitude invariant is a (signed and weighted) count of the critical points of the original function.
Our about detailed example of persistent magnitude in action is the * Rips magnitude *. This is the numerical invariant of finite metric space prevail by taking the persistent magnitude of the Rips homology, and is give by the weighted simplex - count $ $ |X|_\mathrm{Rips }
=
\sum_{\emptyset\neq A\subseteq X}(-1)^{\#A-1}e^{-\operatorname{diam}(A)}.$$ We compute the Rips magnitude of hertz graphs with their path, Euclidean and geodesic system of measurement. In each case they are determine by a number - theoretical formula evocative of the sum of divisors function.
In the original setting of order of magnitude, defining the magnitude of infinite measured spaces is not straightforward: the simplest method is to take the supremum of the order of magnitude of all finite subspaces of the given infinite measured space, but there are alternatives, and currently the hypothesis only works well in the | magjitude* $|M_\ast|$ of a finiteln presented gradgd persivtence module $O_\ast$ is defined by $$|M_\ast|
=
\sun_{k=0}^{\infty}
\sum_{i=1}^{m_k}
(-1)^k (e^{-a_{k,i}} - e^{-b_{n,i}})$$ where $M_\asu$ has bars $[a_{k,1},b_{k,1}),\ldots,[a_{k,m_k},b_{k,m_k})$ in fegrze $k\geq 0$. Note thst a bar $[a,t)$ makes a contsicucion of $\pm(e^{-a} - e^{-b})$ to the magnitude, so that lpnher bars make w grtatqr ckntribution, in line with the generzl philmsophy of perxistent homology.
Persistent magjitude has good fogmal properjjes: qe show that it is addpcive with rgspect to exact sequences, and thxt thz magnitude od a jensor produrt of iersistence modules iv the ptoduct of the kagiituees of the factors, so long as the tensot product hs understood in an approptiate derkced sehsx.
Nos supplse that we hzve a persiwtent homology theoty eefined for skme clwsf of mathematical objects, for example tve Dips homology of metric spaces. By applying pgrsistent iagnitude to the persistent homology, we obtain a tew nnmdrieql inxqrlant of the mathematical objects in question. Jhr fprst example of tmis process is the cwsr of the sublexel sec psrsistent homology of Morfe fubctions, wrere the resulting magnitude incariant is a (wigned and weighteb) count of tke crijical loints of the original yunctikn.
Our most fetailed sbample of persisgenn macnitude in action is the *Ryps magnivude*. Chis is ghe mumerisal invariwnt on finite metric spafes oytainad by takijg the persistent magnitude of vie Rips homolpgf, atd is giren by the weighted simplex-count $$|R|_\mathrm{Rnps}
=
\sum_{\vmptyset\nxq A\subseteq X}(-1)^{\#A-1}e^{-\operatorndle{diam}(A)}.$$ We cmmpute tre Rups nagnituad of cycle gralhs with nhtir path, Eyclidean and geodexic jetrics. In each cawe they are detrrmkneq hy a tumber-theoredicau futmula reminiscenb ow thr sum of divisors futctikn.
In the original xebting of nagnitudq, defining thr magnitude of inflnite mevric slacgs is not straightforward: the sjmplest mfthjd is to takq thc sukremum of tke magnitude of all finite subspaces of vhe given infinite mettic space, but there ate slternativex, and currentlf the theory only woeks well in the | magnitude* $|M_\ast|$ of a finitely presented graded $M_\ast$ defined by = \sum_{k=0}^{\infty} \sum_{i=1}^{m_k} $M_\ast$ bars $[a_{k,1},b_{k,1}),\ldots,[a_{k,m_k},b_{k,m_k})$ in $k\geq 0$. Note a bar $[a,b)$ makes a contribution $\pm(e^{-a} - e^{-b})$ to the magnitude, so that longer bars make a greater in line with the general philosophy of persistent homology. Persistent magnitude has good properties: show it additive with respect to exact sequences, and that the magnitude of a tensor product of persistence is the product of the magnitudes of the so long as the product is understood in an derived Now suppose we a homology theory defined some class of mathematical objects, for example the Rips homology of metric spaces. By applying persistent magnitude the persistent obtain a numerical of mathematical objects in first example of this process is the sublevel set persistent homology of Morse functions, the resulting invariant is a (signed and weighted) of the critical points of the original function. most detailed example of persistent magnitude in action is the *Rips magnitude*. This is the of finite metric spaces by taking the magnitude the homology, is given the weighted simplex-count $$|X|_\mathrm{Rips} = \sum_{\emptyset\neq A\subseteq X}(-1)^{\#A-1}e^{-\operatorname{diam}(A)}.$$ We compute the magnitude of cycle graphs with their path, Euclidean and geodesic each they are determined a number-theoretical formula reminiscent the of divisors function. In setting magnitude, of metric is not straightforward: the method is to take the of the magnitude of given infinite metric space, but there are alternatives, currently the theory only works well in | magnitude* $|M_\ast|$ of a finitely pResented grAded pErsIstEnCe moDule $m_\ast$ is defined bY $$|m_\ast|
=
\Sum_{k=0}^{\infty}
\sum_{i=1}^{m_k}
(-1)^k (e^{-a_{k,i}} - e^{-b_{K,i}})$$ wheRe $m_\Ast$ hAS bArs $[a_{k,1},B_{k,1}),\ldots,[A_{K,m_K},B_{K,m_k})$ In DeGreE $k\GEq 0$. note tHat A bar $[a,b)$ mAkes a contrIbuTiOn of $\pm(e^{-a} - e^{-b})$ to THe Magnitude, sO thAt longer bars MakE a greaTeR coNTribuTioN, in liNe with THe geneRal philosOpHY of perSIstent hOMOlOgy.
PErsistent magnitudE HaS Good formal propErties: We SHoW THat It iS additive wItH respECt to exaCT sEQUEncES, and that the maGnitude of a tENsoR produCt Of pERsisteNce moDuLEs iS the product Of thE magnitudEs of thE Factors, SO long as The tenSor ProDuct IS uNdErsToOD in AN aPprOPriAte derivEd SeNse.
NoW supPOSE That We hAve a PersiStent homology TheOry dEFinEd for Some cLass Of MatheMaticaL objeCtS, for example the RIps hOmology of MetRiC spAcEs. By aPPlying PerSisTent magNitude tO The PeRSIStEnt homology, we obtaiN a NEW nUmerical InvariANt Of THe mathemAtIcaL objECTs in qUestIOn. our first ExamplE Of ThIs proceSs Is the cAsE of The SubleVEl seT persiStent homOlogy OF Morse functionS, Where the resulTInG MAgNItudE inVariant is a (sIgneD And wEighTEd) CouNT of thE critIcAL pOInts of the original fuNcTion.
OuR most Detailed exampLe of persisTENT magnituDe in ACtIOn is the *Rips magNitudE*. This is the NUmerical InvarIant of fiNite metriC SPaces obtAinEd bY taKinG THe Persistent magNITude Of The Rips HomOlogy, anD is GivEn bY thE wEighted siMplex-couNt $$|x|_\mAtHrM{RiPs}
=
\sum_{\EMptyset\nEq a\suBsEteQ X}(-1)^{\#A-1}e^{-\oPEratorName{dIam}(A)}.$$ we CoMPutE the RipS MaGNItudE oF cYcle GraPhS with TheiR PatH, EuclidEan and geoDesIC metRiCs. in each cAse they are detErMined by a nuMbEr-tHeoretICAl formulA reminiscent of the sum of dIVisors fUncTion.
IN the Original sEttIng of mAgnITude, deFining The maGnItuDE Of infINItE meTrIc spaces is NOT stRaighTfOrwaRd: the siMplest method is to taKE thE supremum of thE maGnitUDE oF alL FiNIte SuBSpaCES of the given infiNite metric SpACe, But there arE AltErNatives, And currEntly THe theorY only workS well in thE | magnitude* $|M_\ast|$ of a finitelyprese nte d g ra dedpers istence module $M_\ ast$ is defined by $$| M_\as t|
=
\sum_{k = 0} ^ { \in ft y}
\ su m_{i= 1}^ {m_k}
(-1)^k(e^ {- a_{k,i}} - e ^ {- b_{k,i}})$ $ w here $M_\ast $ h as bar s$[a _ {k,1} ,b_ {k,1} ),\ldo t s,[a_{ k,m_k},b_ {k , m_k})$ in degr e e $ k\ge q 0$. Note that a ba r $[a,b)$ makes a con tr i bu t i onof$\pm(e^{-a }- e^{ - b})$ to th e m agn i tude, so that longer bar s ma ke a g re ate r contr ibuti on , in line withthegeneral p hiloso p hy of p e rsisten t homo log y.
Per s is te ntma g nit u de ha s go od forma lpr opert ies: w e show th at i t isadditive with re spec t to exac t seq uenc es , and thatthe m ag nitude of a ten sorproduct o f p er sis te nce m o dulesisthe produc t of th e ma gn i t u de s of the factors,so l on g as the tenso r p ro d uct is u nd ers tood i n anappr o pr iate der ived s e ns e.
Now s up pose t ha t w e h ave a pers istent homolog y the o ry defined for some class of ma t h em a tica l o bjects, for exa m pletheR ip s h o molog y ofme t ri c spaces. By applyin gpersis tentmagnitude tothe persis t e n t homolo gy,w eo btain a new nu meric al invaria n t of the math ematical objectsi n questio n.Our fi rst e xa mple of thisp r oces sis thecas e of th e s ubl eve l s et persiste nt homol og yof M ors e fun c tions, w he reth e r esult i ng mag nitud e in va ri a ntis a (s i gn e d and w ei ghte d)co unt o f th e cr iticalpoints of th e ori gi na l funct ion.
Our mos tdetailed e xa mpl e of p e r sistentmagnitude in action ist he *Rip s m agnit ude* . This is th e nume ric a l inva riantof fi ni tem e trics p ac esob tained byt a kin g the p ersi stent m agnitude of the Ri p s h omology, andisgive n by th e w e igh te d si m p lex-count $$|X| _\mathrm{R ip s }
=
\s um _{\empt yset\ne q A\s u bseteqX}(-1)^{\ #A-1}e^{- \o pera t o rna me{diam}(A )}.$$ We computet he Ri p smagni tud e of c yc legraph s with the ir pa th, Eu cl ideanand g eo desic me trics. In each case the y aredeter min ed by a n umb e r-t heoretica l fo rmula remi nis cen t ofthe sum o f di v is ors funct ion.
In the o r ig ina l se tting of ma g n i tud e, de fin i ng the mag nitude of infinit e metric spaces isn o t s tra i ghtf or ward: the simp les tm e thod isto take the s upremumof the m agnitu de ofall fin i t es ubspac es o f t he giveninf in i te metr ic s p ace, b ut t he re are alter n ativ e s , and currentlythe t h e ory o n lyworks w ell int he | magnitude* $|M_\ast|$_of a_finitely presented graded persistence_module $M_\ast$_is_defined by_$$|M_\ast|_
_ =
_ \sum_{k=0}^{\infty}
_ _\sum_{i=1}^{m_k}
_ (-1)^k (e^{-a_{k,i}} - e^{-b_{k,i}})$$ where $M_\ast$ has bars $[a_{k,1},b_{k,1}),\ldots,[a_{k,m_k},b_{k,m_k})$ in degree_$k\geq_0$. Note_that_a_bar $[a,b)$ makes a contribution_of $\pm(e^{-a} - e^{-b})$ to_the magnitude,_so that longer bars make a greater contribution,_in_line with the_general philosophy of persistent homology.
Persistent magnitude has good formal_properties: we show that it is_additive with respect_to_exact_sequences, and that the_magnitude of a tensor product of_persistence modules is the product of_the magnitudes of the factors, so long_as the tensor product is understood_in an appropriate derived sense.
Now_suppose that_we have a persistent homology_theory defined for_some class_of mathematical objects,_for example the Rips homology of_metric spaces. By_applying persistent magnitude to the persistent_homology,_we obtain a_new_numerical_invariant of_the mathematical objects_in_question. Our_first_example of this process is the_case_of the sublevel set persistent homology of_Morse functions, where the_resulting_magnitude invariant is a_(signed and weighted) count of_the critical points of the original_function.
Our most_detailed example_of persistent magnitude in action is the *Rips magnitude*. This is_the numerical invariant of finite metric_spaces obtained by taking_the persistent_magnitude_of the Rips_homology,_and is_given by the weighted simplex-count $$|X|_\mathrm{Rips}
_ _ =
_\sum_{\emptyset\neq A\subseteq X}(-1)^{\#A-1}e^{-\operatorname{diam}(A)}.$$ We_compute_the Rips magnitude of cycle graphs_with their path, Euclidean and geodesic_metrics. In each case they_are_determined_by a number-theoretical formula reminiscent_of the sum of divisors function.
In_the original setting_of magnitude, defining the magnitude of infinite_metric_spaces is not straightforward: the simplest_method_is to take the supremum of_the_magnitude_of all finite subspaces of_the given infinite metric space, but_there are alternatives, and currently the theory only works_well in the |
=1}^{k} |\hat{\cS}_i|$, and similarly for features $j \in \cS^{c}$, $$\begin{aligned}
c^* |\hat{\cS}_{(c^*)} \cap \cS^{c}| \leq \sum_{j \in \cS^{c}: m_j \geq c^*} m_j \leq \sum_{j \in \cS^{c}} m_j = \sum_{i=1}^{k} |\hat{\cS}_i \cap \cS^{c}|.\end{aligned}$$ Therefore, the overall $\FDR$ can be linked to the machine-wise $\FDR$’s as $$\begin{aligned}
\FDR_{\left(c^*\right)} &= \EE\left\{\EE\left[\frac{|\hat{\cS}_{(c^*)} \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}
\nonumber\\&\leq %\EE\left\{\frac{1}{c^*}\EE\left[\frac{\sum_{i=1}^{k} |\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}\nonumber\\
%& =
\EE\left\{\frac{1}{c^*}\sum_{i=1}^{k} \EE\left[\frac{|\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}.\end{aligned}$$ In addition, by definition of $c^*$ in the theorem: $|\hat{\cS}_{(c^*)}| \geq \frac{1}{k} \sum_{i=1}^{k} |\hat{\cS}_i| \geq \frac{k}{\sum_{i=1}^{k} 1/|\hat{\cS}_i|}$, we then obtain $$\begin{aligned}
\FDR_{\left(c^*\right)} & \leq \EE\left\{\frac{1}{c^*}\sum | = 1}^{k } |\hat{\cS}_i|$, and similarly for features $ j \in \cS^{c}$, $ $ \begin{aligned }
c^ * |\hat{\cS}_{(c^ *) } \cap \cS^{c}| \leq \sum_{j \in \cS^{c }: m_j \geq c^ * } m_j \leq \sum_{j \in \cS^{c } } m_j = \sum_{i=1}^{k } |\hat{\cS}_i \cap \cS^{c}|.\end{aligned}$$ Therefore, the overall $ \FDR$ can be linked to the car - fresh $ \FDR$ ’s as $ $ \begin{aligned }
\FDR_{\left(c^*\right) } & = \EE\left\{\EE\left[\frac{|\hat{\cS}_{(c^ *) } \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\ }
\nonumber\\&\leq% \EE\left\{\frac{1}{c^*}\EE\left[\frac{\sum_{i=1}^{k } |\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}\nonumber\\
% & =
\EE\left\{\frac{1}{c^*}\sum_{i=1}^{k } \EE\left[\frac{|\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}.\end{aligned}$$ In addition, by definition of $ c^*$ in the theorem: $ |\hat{\cS}_{(c^*)}| \geq \frac{1}{k } \sum_{i=1}^{k } |\hat{\cS}_i| \geq \frac{k}{\sum_{i=1}^{k } 1/|\hat{\cS}_i|}$, we then obtain $ $ \begin{aligned }
\FDR_{\left(c^*\right) } & \leq \EE\left\{\frac{1}{c^*}\sum | =1}^{k} |\hwt{\cS}_i|$, and similarly for neatures $j \in \cS^{e}$, $$\begin{eligned}
d^* |\hat{\cS}_{(c^*)} \cap \cS^{c}| \leq \sum_{j \in \cS^{c}: m_j \jeq x^*} m_j \oeq \sum_{j \in \cS^{c}} m_j = \suo_{i=1}^{k} |\hat{\cS}_p \cap \cS^{c}|.\wnd{aougned}$$ Therxrore, thc ovedwll $\YDC$ can be linked to the mawhine-wise $\FDR$’s ar $$\yegin{aligned}
\FDR_{\left(c^*\right)} &= \EE\left\{\EE\lqft[\frac{|\nah{\cS}_{(c^*)} \cap \cS^{c}|}{|\haj{\cS}_{(c^*)}|}\Bpg|\jeft(\gat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}
\nonujber\\&\leq %\EE\left\{\frac{1}{c^*}\ER\left[\frac{\sum_{i=1}^{k} |\hat{\cS}_i \cap \fS^{c}|}{|\hwt{\cS}_{(c^*)}|}|\left(\hat{\cS}_1,\dots,\jat{\cS}_k\right)\tjghe]\eight\}\nonumbef\\
%& =
\EE\lefu\{\fxac{1}{c^*}\sum_{i=1}^{k} \EG\left[\frac{|\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\ueft(\hct{\cS}_1,\dots,\hat{\xS}_j\rihvt)\right]\righv\}.\end{alpgned}$$ In addibpon, by gefinitoon of $c^*$ in thc themren: $|\hat{\cS}_{(c^*)}| \geq \frac{1}{k} \suk_{i=1}^{k} |\hat{\cS}_i| \geq \fras{k}{\sum_{i=1}^{k} 1/|\hdt{\eS}_i|}$, we then obtain $$\betib{aligted}
\FGR_{\lewr(c^*\rkghu)} & \kes \EE\levt\{\fcac{1}{c^*}\sum | =1}^{k} |\hat{\cS}_i|$, and similarly for features $j $$\begin{aligned} |\hat{\cS}_{(c^*)} \cap \leq \sum_{j \in \leq \in \cS^{c}} m_j \sum_{i=1}^{k} |\hat{\cS}_i \cap Therefore, the overall $\FDR$ can be to the machine-wise $\FDR$’s as $$\begin{aligned} \FDR_{\left(c^*\right)} &= \EE\left\{\EE\left[\frac{|\hat{\cS}_{(c^*)} \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\} \nonumber\\&\leq %\EE\left\{\frac{1}{c^*}\EE\left[\frac{\sum_{i=1}^{k} \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}\nonumber\\ %& = \EE\left\{\frac{1}{c^*}\sum_{i=1}^{k} \EE\left[\frac{|\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}.\end{aligned}$$ In addition, by definition of in theorem: \geq \sum_{i=1}^{k} |\hat{\cS}_i| \geq \frac{k}{\sum_{i=1}^{k} 1/|\hat{\cS}_i|}$, we then obtain $$\begin{aligned} \FDR_{\left(c^*\right)} & \leq \EE\left\{\frac{1}{c^*}\sum | =1}^{k} |\hat{\cS}_i|$, and similarly for feaTures $j \in \cS^{C}$, $$\begiN{alIgnEd}
C^* |\hat{\CS}_{(c^*)} \cAp \cS^{c}| \leq \sum_{j \in \Cs^{c}: m_j \Geq c^*} m_j \leq \sum_{j \in \cS^{c}} m_j = \suM_{i=1}^{k} |\haT{\cs}_I \cap \Cs^{c}|.\End{alIgned}$$ ThEReFORe, tHe OvEraLl $\fdR$ Can be LinKed to thE machine-wiSe $\FdR$’S as $$\begin{aligNEd}
\fDR_{\left(c^*\riGht)} &= \eE\left\{\EE\left[\FraC{|\hat{\cS}_{(C^*)} \cAp \cs^{C}|}{|\hat{\cs}_{(c^*)}|}\BIg|\lefT(\hat{\cS}_1,\DOts,\hat{\CS}_k\right)\rIgHT]\right\}
\NOnumber\\&\LEQ %\Ee\lefT\{\frac{1}{c^*}\EE\left[\frac{\sUM_{i=1}^{K} |\Hat{\cS}_i \cap \cS^{c}|}{|\haT{\cS}_{(c^*)}|}|\leFt(\HAt{\Cs}_1,\DotS,\haT{\cS}_k\right)\rIgHt]\rigHT\}\nonumbER\\
%& =
\Ee\LEFt\{\fRAc{1}{c^*}\sum_{i=1}^{k} \EE\lefT[\frac{|\hat{\cS}_i \CAp \cs^{c}|}{|\hat{\cs}_{(c^*)}|}\big|\LEft(\hat{\CS}_1,\dotS,\hAT{\cS}_K\right)\right]\RighT\}.\end{alignEd}$$ In adDItion, by DEfinitiOn of $c^*$ iN thE thEoreM: $|\HaT{\cs}_{(c^*)}| \gEq \FRac{1}{K} \SuM_{i=1}^{k} |\HAt{\cs}_i| \geq \fraC{k}{\SuM_{i=1}^{k} 1/|\haT{\cS}_i|}$, WE THEn obTaiN $$\begIn{aliGned}
\FDR_{\left(c^*\rIghT)} & \leq \ee\leFt\{\fraC{1}{c^*}\sum | =1}^{k} |\hat{\cS}_i|$, an d similarl y for fe atu re s $j \in \cS^{c}$, $$\ b egin {aligned}
c^* |\hat{\c S}_{( c^ * )} \ c ap \cS^ {c}| \l e q\ s um_ {j \ in\c S ^{ c}: m _j\geq c^ *} m_j \le q \ su m_{j \in \cS ^ {c }} m_j = \ sum _{i=1}^{k} | \ha t{\cS} _i \c a p \cS ^{c }|.\e nd{ali g ned}$$ Therefor e, the ov e rall $\ F D R$ can be linked to the ma c hine-wise $\FD R$’s a s$ $\ b e gin {al igned}
\FD R_ {\lef t (c^*\ri g ht ) } &=\ EE\left\{\EE\ left[\frac{ | \ha t{\cS} _{ (c^ * )} \ca p \cS ^{ c }|} {|\hat{\cS} _{(c ^*)}|}\Bi g|\lef t (\hat{\ c S}_1,\d ots,\h at{ \cS }_k\ r ig ht )\r ig h t]\ r ig ht\ }
\n onumber\ \& \l eq %\ EE\l e f t \ {\fr ac{ 1}{c ^*}\E E\left[\frac{ \su m_{i = 1}^ {k} | \hat{ \cS} _i \cap \cS^{ c}|}{ |\ hat{\cS}_{(c^*) }|}| \left(\ha t{\ cS }_1 ,\ dots, \ hat{\c S}_ k\r ight)\r ight]\r i ght \} \ n o nu mber\\
%& =
\EE \l e f t\ {\frac{1 }{c^*} \ su m_ { i=1}^{k} \ EE\ left [ \ frac{ |\ha t {\ cS}_i \c ap \cS ^ {c }| }{|\hat {\ cS}_{( c^ *)} |}\ Big|\ l eft( \hat{\ cS}_1,\d ots,\ h at{\cS}_k\righ t )\right]\righ t \} . \ en d {ali gne d}$$ In add itio n , by def i ni tio n of $ c^*$in th e theorem: $|\hat{\c S} _{(c^* )}| \ geq \frac{1}{ k} \sum_{i = 1 } ^{k} |\h at{\ c S} _ i| \geq \frac{ k}{\s um_{i=1}^{ k } 1/|\ha t{\cS }_i|}$,we then o b t ain $$\b egi n{a lig ned } \F DR_{\left(c^* \ r ight )} & \leq \E E\left\ {\f rac {1} {c^ *} \sum | =1}^{k} |\hat{\cS}_i|$,_and similarly_for features $j \in_\cS^{c}$, $$\begin{aligned}
c^*_|\hat{\cS}_{(c^*)}_\cap \cS^{c}|_\leq_\sum_{j \in \cS^{c}:_m_j \geq c^*}_m_j \leq \sum_{j \in_\cS^{c}} m_j =_\sum_{i=1}^{k}_|\hat{\cS}_i \cap \cS^{c}|.\end{aligned}$$ Therefore, the overall $\FDR$ can be linked to the machine-wise $\FDR$’s_as_$$\begin{aligned}
\FDR_{\left(c^*\right)} &=_\EE\left\{\EE\left[\frac{|\hat{\cS}_{(c^*)}_\cap_\cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}
\nonumber\\&\leq %\EE\left\{\frac{1}{c^*}\EE\left[\frac{\sum_{i=1}^{k} |\hat{\cS}_i \cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}\nonumber\\
_%& =
\EE\left\{\frac{1}{c^*}\sum_{i=1}^{k} \EE\left[\frac{|\hat{\cS}_i_\cap \cS^{c}|}{|\hat{\cS}_{(c^*)}|}\Big|\left(\hat{\cS}_1,\dots,\hat{\cS}_k\right)\right]\right\}.\end{aligned}$$_In addition, by definition of $c^*$ in the_theorem:_$|\hat{\cS}_{(c^*)}| \geq \frac{1}{k}_\sum_{i=1}^{k} |\hat{\cS}_i| \geq \frac{k}{\sum_{i=1}^{k} 1/|\hat{\cS}_i|}$, we then obtain $$\begin{aligned}
\FDR_{\left(c^*\right)}_& \leq \EE\left\{\frac{1}{c^*}\sum |
$u_\vartheta$ satisfies $$\label{eq:linearized-angle}
\left\{
\begin{aligned}
-\Delta u_\vartheta&= f'(u)u_\vartheta &&\qquad \text{in $\Omega$,}\\
u_\vartheta & = 0 &&\qquad \text{on $\partial \Omega$.}
\end{aligned}
\right.$$ Suppose by contradiction that $u_\vartheta \not \equiv 0$. Then $u_\vartheta$ is a Dirichlet eigenfunction of $-\Delta- f'(u)$ corresponding to the eigenvalue $0$. Moreover, since $u$ is $2\pi$-periodic in $\vartheta$, $u_\theta$ changes sign. This implies that the first Dirichlet eigenvalue of $-\Delta - f'(u)$ is negative, which contradicts our assumption that $m(u)=0$.
Hence is satisfied, and it also holds true for an arbitrary rotation $\tilde u$ of the function $u$ since $\tilde u$ satisfies the same assumptions as $u$. Consequently, $u$ is a radial sign changing solution of. However, then [@AftalionPacella Theorem 1.1] implies that $m(u) \ge N+1>0$, which is a contradiction. We thus conclude that $m(u) \ge 1$ for every nodal solution $u$ of, as claimed in (ii).\
Next, let $u \in C^1_0(\Omega)$ be a nodal solution with $I(u)= c_{nod}$ (such a solution exists by Theorem \[thm:lensachieved\]). By (ii) and Theorem \[prop:boundmorseindex\], we have $m(u)=1$, so (i) follows by Theorem \[sec:lens-mp\]. It also follows from Theorem \[sec:lens-mp\] that $u$ is of mountain pass type. Moreover, by [@AftalionPacella Theorem 1.1], $u$ is nonradial.
Next we prove that $u$ is foliated Schwarz symmetric. Let $x_0 \in \Omega \setminus \{0\}$ be chosen such that $u(x_0)=\max \{u(x)\::\: |x|=|x_0|\}$. We put $e:=\frac{x_0}{|x_0|}$, and we consider the family ${{\mathcal H}}_e$ of all open halfspaces | $ u_\vartheta$ satisfies $ $ \label{eq: linearized - angle }
\left\ {
\begin{aligned }
-\Delta u_\vartheta&= f'(u)u_\vartheta & & \qquad \text{in $ \Omega$,}\\
u_\vartheta & = 0 & & \qquad \text{on $ \partial \Omega$. }
\end{aligned }
\right.$$ Suppose by contradiction that $ u_\vartheta \not \equiv 0$. Then $ u_\vartheta$ is a Dirichlet eigenfunction of $ -\Delta- f'(u)$ corresponding to the eigenvalue $ 0$. Moreover, since $ u$ is $ 2\pi$-periodic in $ \vartheta$, $ u_\theta$ change sign of the zodiac. This implies that the first Dirichlet eigenvalue of $ -\Delta - f'(u)$ is negative, which contradict our assumption that $ m(u)=0$.
Hence is quenched, and it also holds true for an arbitrary rotation $ \tilde u$ of the affair $ u$ since $ \tilde u$ satisfies the same assumptions as $ u$. Consequently, $ u$ is a radial polarity changing solution of. However, then [ @AftalionPacella Theorem 1.1 ] implies that $ m(u) \ge N+1>0 $, which is a contradiction. We therefore conclude that $ m(u) \ge 1 $ for every nodal solution $ u$ of, as claimed in (ii).\
Next, let $ u \in C^1_0(\Omega)$ be a nodal solution with $ I(u)= c_{nod}$ (such a solution exists by Theorem \[thm: lensachieved\ ]). By (ii) and Theorem \[prop: boundmorseindex\ ], we have $ m(u)=1 $, so (i) succeed by Theorem \[sec: lens - mp\ ]. It also follows from Theorem \[sec: lens - mp\ ] that $ u$ is of batch pass type. furthermore, by [ @AftalionPacella Theorem 1.1 ], $ u$ is nonradial.
Next we prove that $ u$ is foliate Schwarz symmetric. Let $ x_0 \in \Omega \setminus \{0\}$ be chosen such that $ u(x_0)=\max \{u(x)\::\: |x|=|x_0|\}$. We put $ e:=\frac{x_0}{|x_0|}$, and we consider the family $ { { \mathcal H}}_e$ of all clear halfspaces | $u_\vwrtheta$ satisfies $$\label{ed:linearized-anglg}
\lwft\{
\bxgin{alifned}
-\Deltx u_\vartheta&= f'(u)u_\vartheta &&\qquav \tezt{in $\Imega$,}\\
u_\vartheta & = 0 &&\qquaa \text{on $\iartial \Onega$.}
\end{almfned}
\rigmc.$$ Supllse yy contradiction that $u_\vardheta \not \equie 0$. Tken $u_\vartheta$ is a Dirichlet eigenfugction pf $-\Delta- f'(u)$ corrgsponcyng fo the eigenvalue $0$. Moreover, since $h$ is $2\pi$-keriodic in $\vartheya$, $u_\theta$ changes sign. Thid imolies that the firdt Dirichlej eidwnvalue of $-\Ddlta - f'(u)$ is negative, wgich contradicts our assumption that $m(u)=0$.
Hence iw watlvfied, and iv also holds true npr an drbitraty rotation $\tikde u$ if the function $u$ sinre $\tilde u$ satisfies the same aasumptions as $u$. Cinwequettly, $u$ ir a fadjak aign cjanjing solutikn of. Howevwr, then [@AftalionPactllw Theorem 1.1] impmies trae $m(u) \ge N+1>0$, which is a contradiction. We tvus conclude that $m(u) \ge 1$ fir every nodal solutiln $u$ of, af claimed in (ii).\
Next, let $u \in C^1_0(\Omega)$ be a nodal sonutioi ditk $I(u)= c_{vid}$ (such a solution exists by Theorem \[thm:lensachiqbec\]). Ny (ii) and Theorei \[prop:boundmprdeogdex\], we have $o(u)=1$, so (n) fkllows by Theorem \[sfc:lens-mk\]. It aoso follors ftom Theorem \[sec:lens-mp\] that $u$ is of mountcin pass type. Moreovex, by [@AftaliouPacelka Throrem 1.1], $u$ is nonradial.
Nert we lrove that $k$ is folizged Schwarz symmdtrpc. Lat $x_0 \in \Omega \setminus \{0\}$ be chosen snch tkat $u(x_0)=\mab \{u(x)\::\: |x|=|x_0|\}$. We put $e:=\frac{d_0}{|x_0|}$, and we consider the fwmily ${{\mdthcal H}}_e$ lf all open halfspaces | $u_\vartheta$ satisfies $$\label{eq:linearized-angle} \left\{ \begin{aligned} -\Delta u_\vartheta&= \text{in u_\vartheta & 0 &&\qquad \text{on by that $u_\vartheta \not 0$. Then $u_\vartheta$ a Dirichlet eigenfunction of $-\Delta- f'(u)$ to the eigenvalue $0$. Moreover, since $u$ is $2\pi$-periodic in $\vartheta$, $u_\theta$ changes This implies that the first Dirichlet eigenvalue of $-\Delta - f'(u)$ is negative, contradicts assumption $m(u)=0$. is satisfied, and it also holds true for an arbitrary rotation $\tilde u$ of the function since $\tilde u$ satisfies the same assumptions as Consequently, $u$ is a sign changing solution of. However, [@AftalionPacella 1.1] implies $m(u) N+1>0$, is a contradiction. thus conclude that $m(u) \ge 1$ for every nodal solution $u$ of, as claimed in (ii).\ Next, $u \in a nodal with c_{nod}$ a solution exists \[thm:lensachieved\]). By (ii) and Theorem \[prop:boundmorseindex\], so (i) follows by Theorem \[sec:lens-mp\]. It also from Theorem that $u$ is of mountain pass Moreover, by [@AftalionPacella Theorem 1.1], $u$ is nonradial. we prove that $u$ is foliated Schwarz symmetric. Let $x_0 \in \Omega \setminus \{0\}$ be that $u(x_0)=\max \{u(x)\::\: |x|=|x_0|\}$. put $e:=\frac{x_0}{|x_0|}$, and consider family H}}_e$ all open | $u_\vartheta$ satisfies $$\label{eq:Linearized-Angle}
\LefT\{
\beGiN{aliGned}
-\delta u_\vartheta&= F'(U)u_\vaRtheta &&\qquad \text{in $\Omega$,}\\
U_\vartHeTA & = 0 &&\qquAD \tExt{on $\Partial \oMeGA$.}
\End{AlIgNed}
\RiGHt.$$ suppoSe bY contraDiction thaT $u_\vArTheta \not \equiV 0$. thEn $u_\varthetA$ is A Dirichlet eiGenFunctiOn Of $-\DELta- f'(u)$ CorRespoNding tO The eigEnvalue $0$. MoReOVer, sinCE $u$ is $2\pi$-pERIoDic iN $\vartheta$, $u_\theta$ chANgES sign. This impliEs that ThE FiRST DiRicHlet eigenvAlUe of $-\DELta - f'(u)$ is NEgATIVe, wHIch contradictS our assumptIOn tHat $m(u)=0$.
HEnCe iS SatisfIed, anD iT AlsO holds true fOr an Arbitrary RotatiON $\tilde u$ OF the funCtion $u$ SinCe $\tIlde U$ SaTiSfiEs THe sAMe AssUMptIons as $u$. COnSeQuentLy, $u$ iS A RADial SigN chaNging Solution of. HowEveR, theN [@aftAlionpacelLa ThEoRem 1.1] imPlies tHat $m(u) \Ge n+1>0$, which is a contraDictIon. We thus ConClUde ThAt $m(u) \gE 1$ For eveRy nOdaL solutiOn $u$ of, as CLaiMeD IN (Ii).\
next, let $u \in C^1_0(\Omega)$ be A nODAl Solution With $I(u)= C_{NoD}$ (sUCh a solutIoN exIsts BY theorEm \[thM:LeNsachievEd\]). By (ii) ANd thEorem \[prOp:BoundmOrSeiNdeX\], we haVE $m(u)=1$, sO (i) follOws by TheOrem \[sEC:lens-mp\]. It also fOLlows from TheoREm \[SEC:lENs-mp\] ThaT $u$ is of mountAin pASs tyPe. MoREoVer, BY [@AftaLionPAcELlA theorem 1.1], $u$ is nonradial.
neXt we prOve thAt $u$ is foliated schwarz symMETRic. Let $x_0 \iN \OmeGA \sETminus \{0\}$ be chosen Such tHat $u(x_0)=\max \{u(x)\::\: |X|=|X_0|\}$. We put $e:=\fRac{x_0}{|x_0|}$, And we conSider the fAMIly ${{\mathcAl H}}_E$ of All OpeN HAlFspaces | $u_\vartheta$ satisfies $ $\label{eq :line ari zed -a ngle }
\l eft\{
\begin { alig ned}
-\Delta u_\varthe ta&=f' ( u)u_ \ va rthet a &&\qq u ad \ tex t{ in $\ Om e ga $,}\\
u_ \varthe ta & = 0 & &\q qu ad \text{on$ \p artial \Om ega $.}
\e nd{ aligne d}
\r i ght.$ $ S uppos e by c o ntradi ction tha t$ u_\var t heta \n o t \ equi v 0$. Then $u_\va r th e ta$ is a Diric hlet e ig e nf u n cti onof $-\Delt a- f'(u ) $ corre s po n d i ngt o the eigenva lue $0$. Mo r eov er, si nc e $ u $ is $ 2\pi$ -p e rio dic in $\va rthe ta$, $u_\ theta$ changes sign. T his im pli esthat th efir st Dir i ch let eig envalueof $ -\Del ta - f ' ( u)$isnega tive, which contra dic ts o u r a ssump tionthat $ m(u)= 0$.
H enceis satisfied, and italso hold s t ru e f or an a r bitrar y r ota tion $\ tilde u $ of t h e fu nction $u$ since $ \t i l de u$ sati sfiest he s a me assum pt ion s as $ u$. C onse q ue ntly, $u $ is a ra di al sign c hangin gsol uti on of . How ever,then [@A ftali o nPacella Theor e m 1.1] implie s t h a t$ m(u) \g e N+1>0$, w hich is a con t ra dic t ion.We th us co n clude that $m(u) \g e1$ for ever y nodal solut ion $u$ of , a s claime d in (i i ).\
Next, let$u \i n C^1_0(\O m ega)$ be a no dal solu tion with $ I(u)= c_ {no d}$ (s uch a s olution exist s by T he orem \[ thm :lensac hie ved \]) . B y(ii) andTheorem\[ pr op :b oun dmors e index\], w e h av e $ m(u)= 1 $, so(i) f ollo ws b y Th eorem \ [ se c : lens -m p\ ]. I t a ls o fol lows fro m Theor em \[sec: len s -mp\ ]th at $u$is of mountai npass type. M ore over,b y [@Aftal ionPacella Theorem 1.1] , $u$ is no nradi al.
Next wepro ve tha t $ u $ is f oliate d Sch wa rzs y mmetr i c .Let $ x_0 \in \O m e ga\setm in us \ {0\}$ b e chosen such that $u( x_0)=\max \{u (x) \::\ : |x |=| x _0 | \}$ .W e p u t $e:=\frac{x_0} {|x_0|}$,an d w e consider the f amily $ {{\math cal H } }_e$ of all open halfspac es | $u_\vartheta$_satisfies $$\label{eq:linearized-angle}
\left\{
_ \begin{aligned}
-\Delta u_\vartheta&= f'(u)u_\vartheta_&&\qquad \text{in_$\Omega$,}\\
u_\vartheta_& =_0_&&\qquad \text{on $\partial_\Omega$.} _
\end{aligned}
\right.$$_Suppose by contradiction_that_$u_\vartheta \not \equiv 0$. Then $u_\vartheta$ is a Dirichlet eigenfunction of $-\Delta- f'(u)$ corresponding_to_the eigenvalue_$0$._Moreover,_since $u$ is $2\pi$-periodic in_$\vartheta$, $u_\theta$ changes sign. This_implies that_the first Dirichlet eigenvalue of $-\Delta - f'(u)$_is_negative, which contradicts_our assumption that $m(u)=0$.
Hence is satisfied, and it also_holds true for an arbitrary rotation_$\tilde u$ of_the_function_$u$ since $\tilde u$_satisfies the same assumptions as $u$._Consequently, $u$ is a radial sign_changing solution of. However, then [@AftalionPacella Theorem_1.1] implies that $m(u) \ge N+1>0$,_which is a contradiction. We_thus conclude_that $m(u) \ge 1$ for_every nodal solution_$u$ of,_as claimed in_(ii).\
Next, let $u \in C^1_0(\Omega)$ be_a nodal solution_with $I(u)= c_{nod}$ (such a solution_exists_by Theorem \[thm:lensachieved\]). By_(ii)_and_Theorem \[prop:boundmorseindex\], we_have $m(u)=1$, so_(i)_follows by_Theorem \[sec:lens-mp\]._It also follows from Theorem \[sec:lens-mp\] that_$u$_is of mountain pass type. Moreover, by_[@AftalionPacella Theorem 1.1], $u$_is_nonradial.
Next we prove that_$u$ is foliated Schwarz symmetric._Let $x_0 \in \Omega \setminus \{0\}$_be chosen_such that_$u(x_0)=\max \{u(x)\::\: |x|=|x_0|\}$. We put $e:=\frac{x_0}{|x_0|}$, and we consider the family_${{\mathcal H}}_e$ of all open halfspaces |
( \mathbb{Q}\right).\end{array}
\label{Def._penalty_theta_(2)}$$Further, define the convex measure of risk $$\rho \left( X\right) :=\sup_{\mathbb{Q\in }\mathcal{Q}_{\ll }(\mathbb{P})}\left\{ \mathbb{E}_{\mathbb{Q}}\left[ -X\right] -\vartheta \left( \mathbb{Q}\right) \right\}. \label{rho def.}$$Notice that $\rho $ is a normalized and sensitive measure of risk. For each class of probability measures introduced so far, the subclass of those measures with a finite penalization is considered. We will denote by $\mathcal{Q}^{\vartheta },$ $\mathcal{Q}_{\ll }^{\vartheta }(\mathbb{P})$ and $\mathcal{Q}_{\approx }^{\vartheta }(\mathbb{P})$ the respective subclasses, i.e. $$\mathcal{Q}^{\vartheta }:=\left\{ \mathbb{Q}\in \mathcal{Q}:\vartheta \left(
\mathbb{Q}\right) <\infty \right\},\ \mathcal{Q}_{\ll }^{\vartheta }(\mathbb{P}):=\mathcal{Q}^{\vartheta }\cap \mathcal{Q}_{\ll }(\mathbb{P})\text{ and }\mathcal{Q}_{\approx }^{\vartheta }(\mathbb{P}):=\mathcal{Q}^{\vartheta }\cap \mathcal{Q}_{\approx }(\mathbb{P}). \label{Def._Qdelta(P)}$$Notice that $\mathcal{Q}_{\approx }^{\vartheta
}(\mathbb{P})\neq \varnothing.$
The next theorem establishes the minimality on $\mathcal{Q}_{\ll }\left(
\mathbb{P}\right) $ of the penalty function introduced above for the risk measure $\rho $. Its proof is based on the sufficient conditions given in Theorem \[static minimal penalty funct. in Q(<<) <=>\].
\[theta=minimal penalty function\] The penalty function $\vartheta $ defined in $\left( \ref{Def._penalty_theta}\right) $ is equal to the minimal penalty function of the convex risk measure $\rho $, given by $\left( \ref{rho def.}\right) $, on $\mathcal{Q}_{\ll }\left( \mathbb{P}\right) $, i.e.$$\vartheta \mathbf{1}_{\mathcal{ | ( \mathbb{Q}\right).\end{array }
\label{Def._penalty_theta_(2)}$$Further, define the convex measure of risk $ $ \rho \left (X\right): = \sup_{\mathbb{Q\in } \mathcal{Q}_{\ll } (\mathbb{P})}\left\ { \mathbb{E}_{\mathbb{Q}}\left [ -X\right ] -\vartheta \left (\mathbb{Q}\right) \right\ }. \label{rho def.}$$Notice that $ \rho $ is a normalized and sensitive standard of hazard. For each class of probability measures insert then far, the subclass of those measures with a finite penalization is consider. We will denote by $ \mathcal{Q}^{\vartheta }, $ $ \mathcal{Q}_{\ll } ^{\vartheta } (\mathbb{P})$ and $ \mathcal{Q}_{\approx } ^{\vartheta } (\mathbb{P})$ the respective subclasses, i.e. $ $ \mathcal{Q}^{\vartheta }: = \left\ { \mathbb{Q}\in \mathcal{Q}:\vartheta \left (
\mathbb{Q}\right) < \infty \right\},\ \mathcal{Q}_{\ll } ^{\vartheta } (\mathbb{P}):=\mathcal{Q}^{\vartheta } \cap \mathcal{Q}_{\ll } (\mathbb{P})\text { and } \mathcal{Q}_{\approx } ^{\vartheta } (\mathbb{P}):=\mathcal{Q}^{\vartheta } \cap \mathcal{Q}_{\approx } (\mathbb{P }). \label{Def._Qdelta(P)}$$Notice that $ \mathcal{Q}_{\approx } ^{\vartheta
} (\mathbb{P})\neq \varnothing.$
The following theorem establishes the minimality on $ \mathcal{Q}_{\ll } \left (
\mathbb{P}\right) $ of the penalty routine introduced above for the risk standard $ \rho $. Its proof is based on the sufficient conditions given in Theorem \[static minimal penalty funct. in Q(<< ;) & lt;=>\ ].
\[theta = minimal penalty function\ ] The punishment function $ \vartheta $ defined in $ \left (\ref{Def._penalty_theta}\right) $ is equal to the minimal punishment function of the convex risk measure $ \rho $, given by $ \left (\ref{rho def.}\right) $, on $ \mathcal{Q}_{\ll } \left (\mathbb{P}\right) $, i.e.$$\vartheta \mathbf{1}_{\mathcal { | ( \mahhbb{Q}\right).\end{array}
\label{Dtf._penalty_theta_(2)}$$Furjhwr, defmne the convex oeasure of risk $$\rho \left( X\rijht) :=\wup_{\maukbb{Q\in }\mathcal{Q}_{\ll }(\maghbb{P})}\left\{ \mathbb{E}_{\nathub{Q}}\left[ -X\right] -\vedtheta \left( \mzbhbb{Q}\xijht) \right\}. \labek{rho def.}$$Nodice that $\rho $ ir c normalized and sensitive measure os risk. Gog each class os prpfabimptn measures introduced so far, the subclavs of those mrasures with a finite penapizahion is considered. We will debote vy $\mathcal{Q}^{\vxrtheta },$ $\mathcal{Q}_{\ll }^{\vattheta }(\mathbb{P})$ and $\mathcal{Q}_{\approb }^{\varcheta }(\mathbv{P})$ thf respective subcjasses, i.e. $$\mabncal{Q}^{\vdrtheta }:=\left\{ \mathbb{Q}\ik \matical{W}:\vartheta \left(
\mathbb{Q}\cight) <\infty \right\},\ \majhcal{Q}_{\ll }^{\vdrcheta }(\mathbb{P}):=\mathcal{Q}^{\caethetd }\ca[ \magycau{Q}_{\lm }(\kafhbb{P})\tfxt{ and }\mathczl{Q}_{\approx }^{\vqrtheta }(\mathbb{P}):=\mathvaj{W}^{\vartheta }\cap \mathcwl{Z}_{\approx }(\mathbb{P}). \label{Def._Qdelta(P)}$$Notice uhat $\jathcal{Q}_{\approx }^{\vartheta
}(\nathbb{P})\neq \varnothing.$
Jhe next treorem establishes the minimality on $\mathcal{Q}_{\ll }\laft(
\mavhcb{P}\xlnht) $ if the penalty function introduced above for thq rosl measure $\rho $. Its proof os bsfed on the suwficieuf donditions given ij Theorgm \[staric minimwl prnalty funct. in Q(<<) <=>\].
\[tyeta=minimal ienaoty function\] The pznalty functnon $\vattheta $ defined in $\left( \ref{Dey._penalfy_theta}\righh) $ is equzu to the minimal pekaldy function of the convex wisk measnre $\rko $, givev by $\left( \wef{rho def.}\gight) $, on $\mathcal{Q}_{\ll }\lefh( \matkbb{P}\rhght) $, i.e.$$\vagtheta \mathbf{1}_{\mathcal{ | ( \mathbb{Q}\right).\end{array} \label{Def._penalty_theta_(2)}$$Further, define the convex measure $$\rho X\right) :=\sup_{\mathbb{Q\in }(\mathbb{P})}\left\{ \mathbb{E}_{\mathbb{Q}}\left[ -X\right] def.}$$Notice $\rho $ is normalized and sensitive of risk. For each class of measures introduced so far, the subclass of those measures with a finite penalization considered. We will denote by $\mathcal{Q}^{\vartheta },$ $\mathcal{Q}_{\ll }^{\vartheta }(\mathbb{P})$ and $\mathcal{Q}_{\approx }^{\vartheta the subclasses, $$\mathcal{Q}^{\vartheta \mathbb{Q}\in \mathcal{Q}:\vartheta \left( \mathbb{Q}\right) <\infty \right\},\ \mathcal{Q}_{\ll }^{\vartheta }(\mathbb{P}):=\mathcal{Q}^{\vartheta }\cap \mathcal{Q}_{\ll }(\mathbb{P})\text{ and }\mathcal{Q}_{\approx }^{\vartheta }(\mathbb{P}):=\mathcal{Q}^{\vartheta \mathcal{Q}_{\approx }(\mathbb{P}). \label{Def._Qdelta(P)}$$Notice that $\mathcal{Q}_{\approx }^{\vartheta }(\mathbb{P})\neq \varnothing.$ next theorem establishes the on $\mathcal{Q}_{\ll }\left( \mathbb{P}\right) $ the function introduced for risk $\rho $. Its is based on the sufficient conditions given in Theorem \[static minimal penalty funct. in Q(<<) <=>\]. \[theta=minimal function\] The $\vartheta $ in \ref{Def._penalty_theta}\right) is equal to penalty function of the convex risk given by $\left( \ref{rho def.}\right) $, on $\mathcal{Q}_{\ll \mathbb{P}\right) $, \mathbf{1}_{\mathcal{ | ( \mathbb{Q}\right).\end{array}
\label{def._penalty_Theta_(2)}$$furTheR, dEfinE the Convex measure oF Risk $$\Rho \left( X\right) :=\sup_{\mathbb{q\in }\maThCAl{Q}_{\lL }(\MaThbb{P})}\Left\{ \matHBb{e}_{\MAthBb{q}}\lEft[ -x\rIGhT] -\vartHetA \left( \maThbb{Q}\right) \RigHt\}. \Label{rho def.}$$NOTiCe that $\rho $ iS a nOrmalized and SenSitive MeAsuRE of riSk. FOr eacH class OF probaBility meaSuREs intrODuced so FAR, tHe suBclass of those measUReS With a finite penAlizatIoN Is CONsiDerEd. We will deNoTe by $\mAThcal{Q}^{\vARtHETA },$ $\maTHcal{Q}_{\ll }^{\vartheTa }(\mathbb{P})$ anD $\MatHcal{Q}_{\aPpRox }^{\VArthetA }(\mathBb{p})$ The Respective sUbclAsses, i.e. $$\maThcal{Q}^{\VArtheta }:=\LEft\{ \mathBb{Q}\in \mAthCal{q}:\varTHeTa \LefT(
\mAThbB{q}\rIghT) <\InfTy \right\},\ \mAtHcAl{Q}_{\ll }^{\VartHETA }(\MathBb{P}):=\MathCal{Q}^{\vArtheta }\cap \matHcaL{Q}_{\ll }(\MAthBb{P})\teXt{ and }\MathCaL{Q}_{\appRox }^{\varTheta }(\MaThbb{P}):=\mathcal{Q}^{\vaRtheTa }\cap \mathCal{q}_{\aPprOx }(\MathbB{p}). \label{def._qdeLta(P)}$$NotIce that $\MAthCaL{q}_{\APpRox }^{\vartheta
}(\mathbb{P})\NeQ \VArNothing.$
THe next THeOrEM establiShEs tHe miNIMalitY on $\mAThCal{Q}_{\ll }\leFt(
\mathBB{P}\RiGht) $ of thE pEnalty FuNctIon IntroDUced Above fOr the risK measURe $\rho $. Its proof iS Based on the sufFIcIENt COndiTioNs given in ThEoreM \[StatIc miNImAl pENalty Funct. In q(≪<) <=>\].
\[theta=minimal peNaLty funCtion\] the penalty funCtion $\varthETA $ Defined iN $\lefT( \ReF{def._penalty_thetA}\righT) $ is equal to THe minimaL penaLty functIon of the cONVex risk mEasUre $\Rho $, GivEN By $\Left( \ref{rho def.}\RIGht) $, oN $\mAthcal{Q}_{\Ll }\lEft( \mathBb{P}\RigHt) $, i.E.$$\vaRtHeta \mathbF{1}_{\mathcal{ | ( \mathbb{Q}\right).\end{a rray}
\lab el{De f._ pen al ty_t heta _(2)}$$Further , def ine the convex measure of r is k $$\ r ho \lef t( X\ri g ht ) :=\ su p_ {\m at h bb {Q\in }\ mathcal {Q}_{\ll } (\m at hbb{P})}\lef t \{ \mathbb{E }_{ \mathbb{Q}}\ lef t[ -X\ ri ght ] -\va rth eta \ left(\ mathbb {Q}\right )\ right\ } . \lab e l {r ho d ef.}$$Notice that $\ r ho $ is a norm alized a n ds e nsi tiv e measureof risk . For ea c hc l a sso f probability measures i n tro ducedso fa r , thesubcl as s of those meas ures with a f initep enaliza t ion isconsid ere d.We w i ll d eno te by$ \m ath c al{ Q}^{\var th et a },$ $\m a t h c al{Q }_{ \ll}^{\v artheta }(\ma thb b{P} ) $ a nd $\ mathc al{Q }_ {\app rox }^ {\var th eta }(\mathbb{P })$the respe cti ve su bc lasse s , i.e. $$ \ma thcal{Q }^{\var t het a} : = \l eft\{ \mathbb{Q}\i n\ m at hcal{Q}: \varth e ta \ l eft(
\ma th bb{ Q}\r i g ht) < \inf t y\right\} ,\ \ma t hc al {Q}_{\l l}^{\va rt het a } (\mat h bb{P }):=\m athcal{Q }^{\v a rtheta }\cap \ m athcal{Q}_{\l l } ( \ ma t hbb{ P}) \text{ and}\ma t hcal {Q}_ { \a ppr o x }^{ \vart he t a} (\mathbb{P}):=\math ca l{Q}^{ \vart heta }\cap \m athcal{Q}_ { \ a pprox }( \mat h bb { P}). \label{D ef._Q delta(P)}$ $ Notice t hat $ \mathcal {Q}_{\app r o x }^{\va rth eta
}( \ma t h bb {P})\neq \var n o thin g. $
Thenex t theor emest abl ish es the mini mality o n$\ ma th cal {Q}_{ \ ll }\lef t(
\m at hbb {P}\r i ght) $ of t he p en al t y f unction in t r oduc ed a bove fo rthe r iskm eas ure $\r ho $. Its pr o of i sba sed onthe sufficien tconditions g ive n in T h e orem \[s tatic minimal penalty f u nct. in Q( <& lt;) <=> ;\] .
\[t het a =minim al pen altyfu nct i o n\] T h e p ena lt y function $ \va rthet a$ de fined i n $\left( \ref{Def . _pe nalty_theta}\ rig ht)$ is eq u al toth e mi n i mal penalty fun ction of t he co nvex riskm eas ur e $\rho $, giv en by $\left( \ref{rho def.}\ri gh t) $ , on$\mathcal{ Q}_{\ll}\left( \ m athbb { P} \righ t)$, i.e .$ $\v arthe ta \ma t hbf {1}_{ \mathc al { | ( \mathbb{Q}\right).\end{array}
\label{Def._penalty_theta_(2)}$$Further,_define the_convex measure of risk_$$\rho \left(_X\right)_:=\sup_{\mathbb{Q\in }\mathcal{Q}_{\ll_}(\mathbb{P})}\left\{_\mathbb{E}_{\mathbb{Q}}\left[ -X\right] -\vartheta_\left( \mathbb{Q}\right) \right\}._ \label{rho def.}$$Notice that_$\rho $ is_a_normalized and sensitive measure of risk. For each class of probability measures introduced so_far,_the subclass_of_those_measures with a finite penalization_is considered. We will denote_by $\mathcal{Q}^{\vartheta_},$ $\mathcal{Q}_{\ll }^{\vartheta }(\mathbb{P})$ and $\mathcal{Q}_{\approx }^{\vartheta }(\mathbb{P})$_the_respective subclasses, i.e._$$\mathcal{Q}^{\vartheta }:=\left\{ \mathbb{Q}\in \mathcal{Q}:\vartheta \left(
\mathbb{Q}\right) <\infty \right\},\ \mathcal{Q}_{\ll }^{\vartheta_}(\mathbb{P}):=\mathcal{Q}^{\vartheta }\cap \mathcal{Q}_{\ll }(\mathbb{P})\text{ and }\mathcal{Q}_{\approx_}^{\vartheta }(\mathbb{P}):=\mathcal{Q}^{\vartheta }\cap_\mathcal{Q}_{\approx_}(\mathbb{P})._ \label{Def._Qdelta(P)}$$Notice that $\mathcal{Q}_{\approx_}^{\vartheta
}(\mathbb{P})\neq \varnothing.$
The next theorem establishes the_minimality on $\mathcal{Q}_{\ll }\left(
\mathbb{P}\right) $ of_the penalty function introduced above for the_risk measure $\rho $. Its proof_is based on the sufficient_conditions given_in Theorem \[static minimal penalty_funct. in Q(<<)_<=>\].
\[theta=minimal penalty_function\] The penalty_function $\vartheta $ defined in $\left(_\ref{Def._penalty_theta}\right) $ is_equal to the minimal penalty function_of_the convex risk_measure_$\rho_$, given_by $\left( \ref{rho_def.}\right)_$, on_$\mathcal{Q}_{\ll_}\left( \mathbb{P}\right) $, i.e.$$\vartheta \mathbf{1}_{\mathcal{ |
whose effect is to slide the end of $T_1$ attached to the sphere part of $F$ along $\alpha$ to $T_2$ and finally out of $P$. Any tubes of $t$ inside of $T_1$ are slid along with it. If there are $r$ such tubes, at a certain point of the homotopy, they create $2r+1$ new intersection circles with $S$, $2r$ from the tubes inside $T_1$ and one from the intersection circle of the tube $T_1$ itself.
The intersection of $T_1$ with $P$ is now a cylinder. Since $\alpha$ is homotopic to $\alpha^\prime$, this cylinder along with any tubes inside it are homotopic to an imbedded position outside of $P$ near $\alpha^\prime$. During the homotopy, self intersections of $t$ may occur, but since the interior of $\alpha^\prime$ does not meet $t$, the final position of $t$ can be an imbedding. This homotopy eliminates $2r+2$ circles of intersection, giving a net decrease of 1 from the original position.
A sequence of such homotopies in each $P$ will give the desired homotopy in $M$. Since $t$ is essential, its image under a composition of such homotopies will not be disjoint form $\Sigma$ hence we will have a normal representative in the same homotopy class.
the universal cover
===================
A fixed maximal sphere system in $M$ gives a description of the universal cover $\widetilde{M}$ of $M$ as follows. Let ${\operatorname{\mathbb{P}}}$ be the set of twice punctured 3-balls in $M$ given by a maximal sphere system $\Sigma$ and regard $M$ as obtained from copies of $P$ in ${\operatorname{\mathbb{P}}}$ by identifying pairs of boundary spheres. Note that a pair might both be contained in a single $P$, in which case the image of $P$ in $M$ is a once-punctured $S^2\times S^1$. To construct $\widetilde{M}$, begin with a single copy of $P$ and attach copies of the $P$ in ${\operatorname{\mathbb{P}}}$ inductively along boundary spheres, as determined by unique path lifting. Repeating this process gives a description of $\widetilde{M}$ as a treelike union of | whose effect is to slide the end of $ T_1 $ attached to the sphere character of $ F$ along $ \alpha$ to $ T_2 $ and ultimately out of $ P$. Any tubes of $ t$ inside of $ T_1 $ are slid along with it. If there be $ r$ such tubes, at a certain compass point of the homotopy, they produce $ 2r+1 $ new overlap circle with $ S$, $ 2r$ from the tubes inside $ T_1 $ and one from the overlap circle of the tube $ T_1 $ itself.
The intersection of $ T_1 $ with $ P$ is now a cylinder. Since $ \alpha$ is homotopic to $ \alpha^\prime$, this cylinder along with any tube inside it are homotopic to an imbedded position outside of $ P$ near $ \alpha^\prime$. During the homotopy, self intersection of $ t$ may occur, but since the interior of $ \alpha^\prime$ does not converge $ t$, the final position of $ t$ can be an imbedding. This homotopy eliminates $ 2r+2 $ lap of intersection, giving a net decrease of 1 from the original position.
A succession of such homotopies in each $ P$ will give the hope homotopy in $ M$. Since $ t$ is essential, its image under a composition of such homotopies will not be disjoint shape $ \Sigma$ hence we will have a normal representative in the same homotopy class.
the cosmopolitan screen
= = = = = = = = = = = = = = = = = = =
A fixed maximal sphere system in $ M$ gives a description of the universal cover $ \widetilde{M}$ of $ M$ as follows. Let $ { \operatorname{\mathbb{P}}}$ be the set of doubly pierced 3 - balls in $ M$ given by a maximal sphere arrangement $ \Sigma$ and regard $ M$ as obtained from copies of $ P$ in $ { \operatorname{\mathbb{P}}}$ by name pair of boundary spheres. Note that a couple might both be contained in a single $ P$, in which case the image of $ P$ in $ M$ is a once - punctured $ S^2\times S^1$. To construct $ \widetilde{M}$, begin with a single copy of $ P$ and attach copies of the $ P$ in $ { \operatorname{\mathbb{P}}}$ inductively along boundary spheres, as settle by unique way lifting. repeat this process gives a description of $ \widetilde{M}$ as a treelike union of | whlse effect is to slide tme end of $T_1$ attaeyed to the slhere paft of $F$ along $\alpha$ to $T_2$ and funallt out of $P$. Any tubes ow $t$ insidv of $T_1$ arw slmd along with it. If therc are $v$ suck vubes, at a certsin point mf the homotopf, ghzy create $2r+1$ new intersection circles with $S$, $2r$ from the tubef inxyde $F_1$ and one from the intersection cidcle of the tube $T_1$ iyself.
The intersection of $T_1$ witj $P$ is now a cylinfer. Since $\aopha$ us homotopic to $\alpha^\pgnme$, this cyminder along with any tubes inskde ic are homotipuc hm an imbeddxd posption outside of $P$ nedr $\alphs^\prime$. During bhe hmmoropy, self intersectiois of $t$ may occur, buj since tha nnterior of $\alpha^\primw$ eoes tot keet $r$, tfe riial positlon of $t$ can ge an imbedeing. This homotopy tliipmates $2r+2$ circmes of igtersection, giving a net decrease of 1 fgom fhe original position.
A wequence of such homojopies in qach $P$ will give the desired homotopy in $M$. Since $d$ is xsrenuiql, itr ilage under a composition of such homotopies wyml npt be disjoint form $\Sigma$ hfnvg we will have a normal representative in the saie honotopy clwss.
tne universal cover
===================
A fixed mqximal spherv sywtem in $M$ gives a bescription uf tne unoversal cover $\widetilde{O}$ of $M$ as folloas. Let ${\opsfatorname{\mathbb{P}}}$ be tve set of twice punctured 3-falls in $N$ giren by a maxomal s[here systfm $\Sinka$ and regard $M$ as obtanned xrom copied of $P$ in ${\operatorname{\mathbb{P}}}$ bb identifying pdirv of bouudary xpheres. Note ehat a pair miyht both be cuntained ih a sinjle $P$, in whish case the ikwge of $P$ in $K$ is a ogce-pynctyred $S^2\tkoes S^1$. To consttuct $\widenijde{M}$, begin with a singke zkpy of $P$ and atcccy copies of the $P$ kn ${\jpvravornaia{\mathbb{P}}}$ indgctixelh alone boundary woherrs, as determined by gniqhe path lifting. Relewting thus procefs gives a dexcription of $\widetllde{M}$ as a trrelyke union of | whose effect is to slide the end attached the sphere of $F$ along out $P$. Any tubes $t$ inside of are slid along with it. If are $r$ such tubes, at a certain point of the homotopy, they create new intersection circles with $S$, $2r$ from the tubes inside $T_1$ and one the circle the $T_1$ itself. The intersection of $T_1$ with $P$ is now a cylinder. Since $\alpha$ is homotopic $\alpha^\prime$, this cylinder along with any tubes inside are homotopic to an position outside of $P$ near During homotopy, self of may but since the of $\alpha^\prime$ does not meet $t$, the final position of $t$ can be an imbedding. This homotopy $2r+2$ circles giving a decrease 1 the original position. of such homotopies in each $P$ desired homotopy in $M$. Since $t$ is essential, image under composition of such homotopies will not disjoint form $\Sigma$ hence we will have a representative in the same homotopy class. the universal cover =================== A fixed maximal sphere system gives a description of universal cover $\widetilde{M}$ $M$ follows. ${\operatorname{\mathbb{P}}}$ the set twice punctured 3-balls in $M$ given by a maximal sphere system and regard $M$ as obtained from copies of $P$ in identifying of boundary spheres. that a pair might be in a single $P$, case image $M$ a $S^2\times S^1$. To construct begin with a single copy $P$ and attach copies inductively along boundary spheres, as determined by unique lifting. Repeating this process gives a description $\widetilde{M}$ as a treelike union of | whose effect is to slide the enD of $T_1$ attachEd to tHe sPheRe Part Of $F$ aLong $\alpha$ to $T_2$ anD FinaLly out of $P$. Any tubes of $t$ inSide oF $T_1$ ARe slID aLong wIth it. If THeRE Are $R$ sUcH tuBeS, At A certAin Point of The homotopY, thEy Create $2r+1$ new inTErSection cirCleS with $S$, $2r$ from tHe tUbes inSiDe $T_1$ ANd one FroM the iNterseCTion ciRcle of the TuBE $T_1$ itseLF.
The intERSeCtioN of $T_1$ with $P$ is now a cyLInDEr. Since $\alpha$ is HomotoPiC To $\ALPha^\PriMe$, this cyliNdEr aloNG with anY TuBES InsIDe it are homotoPic to an imbeDDed PositiOn OutSIde of $P$ Near $\aLpHA^\prIme$. During thE homOtopy, self IntersECtions oF $T$ may occUr, but sIncE thE intERiOr Of $\aLpHA^\prIMe$ DoeS Not Meet $t$, the FiNaL posiTion OF $T$ CAn be An iMbedDing. THis homotopy elImiNateS $2R+2$ ciRcles Of intErseCtIon, giVing a nEt decReAse of 1 from the oriGinaL position.
a seQuEncE oF such HOmotopIes In eAch $P$ wilL give thE DesIrED HOmOtopy in $M$. Since $t$ is esSeNTIaL, its imagE under A CoMpOSition of SuCh hOmotOPIes wiLl noT Be Disjoint Form $\SiGMa$ HeNce we wiLl Have a nOrMal RepResenTAtivE in the Same homoTopy cLAss.
the universaL Cover
===================
A fixed maXImAL SpHEre sYstEm in $M$ gives a DescRIptiOn of THe UniVErsal Cover $\WiDEtILde{M}$ of $M$ as follows. Let ${\OpEratorName{\mAthbb{P}}}$ be the seT of twice puNCTUred 3-ballS in $M$ GIvEN by a maximal sphEre syStem $\Sigma$ aND regard $M$ As obtAined froM copies of $p$ IN ${\operatoRnaMe{\mAthBb{P}}}$ BY IdEntifying pairS OF bouNdAry spheRes. note thaT a pAir MigHt bOtH be contaiNed in a siNgLe $p$, iN wHicH case THe image oF $P$ In $M$ Is A onCe-punCTured $S^2\Times s^1$. To cOnStRUct $\WidetilDE{M}$, BEGin wItH a SingLe cOpY of $P$ aNd atTAch Copies oF the $P$ in ${\opEraTOrnaMe{\MaThbb{P}}}$ inDuctively alonG bOundary sphErEs, aS deterMINed by uniQue path lifting. Repeating THis procEss Gives A desCription oF $\wiDetildE{M}$ aS A treelIke uniOn of | whose effect is to slidethe end of $T_1 $ a tta ch ed t o th e sphere parto f $F $ along $\alpha$ to $T _2$ a nd fina l ly outof $P$. An y tub es o f $ t$ in sideof$T_1$ a re slid al ong w ith it. If t h er e are $r$suc h tubes, ata c ertain p oin t of t hehomot opy, t h ey cre ate $2r+1 $n ew int e rsectio n ci rcle s with $S$, $2r$f ro m the tubes ins ide $T _1 $ a n d on e f rom the in te rsect i on circ l eo f the tube $T_1$ it self.
Thei nte rsecti on of $T_1$with$P $ is now a cyli nder . Since $ \alpha $ is hom o topic t o $\al pha ^\p rime $ ,th iscy l ind e ralo n g w ith anytu be s ins idei t a re h omo topi c toan imbedded p osi tion out sideof $P $ ne ar $\al pha^\p rime$ .During the homo topy , self in ter se cti on s of$ t$ may oc cur , but s ince th e in te r i o rof $\alpha^\prime$ d o e snot meet $t$,t he f i nal posi ti onof $ t $ canbe a n i mbedding . This ho mo topy el im inates $ 2r+ 2$circl e s of inter section, givi n g a net decrea s e of 1 from t h eo r ig i nalpos ition.
A s eque n ce o f su c hhom o topie s inea c h$ P$ will give the de si red ho motop y in $M$. Sin ce $t$ ise s s ential,itsi ma g e under a comp ositi on of such homotopi es wi ll not b e disjoin t form $\S igm a$hen cew e w ill have a no r m al r ep resenta tiv e in th e s ame ho mot op y class.
the uni ve rs al c ove r
=== = ======== == === ==
A fixe d maxim al sp here s ys t emin $M$g iv e s a d es cr ipti onof theuniv e rsa l cover $\wideti lde { M}$of $ M$ as f ollows. Let $ {\ operatorna me {\m athbb{ P } }}$ be t he set of twice punctur e d 3-bal lsin $M $ gi ven by amax imal s phe r e syst em $\S igma$ a ndr e gard$ M $asob tained fro m cop ies o f$P$in ${\o peratorname{\mathb b {P} }}$ by identi fyi ng p a i rs of bo u nda ry sph e r es. Note that a pair migh tb ot h be conta i ned i n a sin gle $P$ , inw hich ca se the im age of $P $in $ M $ is a once-pu nctured$S^2\time s S^1$ . T o con str uct $\ wi det ilde{ M}$, b e gin with a sin gl e copy of $ P$ and att ach copies of the $P$ i n ${\o perat orn ame{\math bb{ P }}} $ inducti vely along bou nda ryspher es, as de term i ne d b y uniq ue p a th liftin g .Rep e a ti ng this pro c e s s g ivesa d e script ionof $\widetilde{M} $ as a treelike uni o n of | whose_effect is_to slide the end_of $T_1$_attached_to the_sphere_part of $F$_along $\alpha$ to_$T_2$ and finally out_of $P$. Any_tubes_of $t$ inside of $T_1$ are slid along with it. If there are $r$_such_tubes, at_a_certain_point of the homotopy, they_create $2r+1$ new intersection circles_with $S$,_$2r$ from the tubes inside $T_1$ and one_from_the intersection circle_of the tube $T_1$ itself.
The intersection of $T_1$ with_$P$ is now a cylinder. Since_$\alpha$ is homotopic_to_$\alpha^\prime$,_this cylinder along with_any tubes inside it are homotopic_to an imbedded position outside of_$P$ near $\alpha^\prime$. During the homotopy, self_intersections of $t$ may occur, but_since the interior of $\alpha^\prime$_does not_meet $t$, the final position_of $t$ can_be an_imbedding. This homotopy_eliminates $2r+2$ circles of intersection, giving_a net decrease_of 1 from the original position.
A_sequence_of such homotopies_in_each_$P$ will_give the desired_homotopy_in $M$._Since_$t$ is essential, its image under_a_composition of such homotopies will not be_disjoint form $\Sigma$ hence_we_will have a normal_representative in the same homotopy_class.
the universal cover
===================
A fixed maximal sphere_system in_$M$ gives_a description of the universal cover $\widetilde{M}$ of $M$ as follows._Let ${\operatorname{\mathbb{P}}}$ be the set of_twice punctured 3-balls in_$M$ given_by_a maximal sphere_system_$\Sigma$ and_regard $M$ as obtained from copies of_$P$ in_${\operatorname{\mathbb{P}}}$ by identifying pairs of boundary_spheres. Note that a_pair_might both be contained in a_single $P$, in which case the_image of $P$ in $M$_is_a_once-punctured $S^2\times S^1$. To construct_$\widetilde{M}$, begin with a single copy_of $P$ and_attach copies of the $P$ in ${\operatorname{\mathbb{P}}}$_inductively_along boundary spheres, as determined by_unique_path lifting. Repeating this process gives_a_description_of $\widetilde{M}$ as a treelike_union of |
107 & 0.168 & 1.061 & 2.875\
V624 Her & A3m & 8150 $\pm$ 150 & 7902 & 3.832 $\pm$ 0.014 & 3.794 & 38.0 & & 0.111 & 0.230 & 1.025 & 2.870\
V1647 Sgr & A1V & 9600 $\pm$ 300 & 9142 & 4.252 $\pm$ 0.008 & 4.087 & & & 0.040 & 0.174 & 1.020 & 2.899\
VV Pyx & A1V & 9500 $\pm$ 200 & 9560 & 4.087 $\pm$ 0.008 & 4.004 & 22.1 & & 0.028 & 0.161 & 1.013 & 2.881\
KW Hya & A5m & 8000 $\pm$ 200 & 8053 & 4.078 $\pm$ 0.006 & 4.390 & 16.6 & & 0.122 & 0.232 & 0.832 & 2.827\
WW Aur & A5m & 7960 $\pm$ 420 & 8401 & 4.161 $\pm$ 0.005 & 4.286 & 35.8 & & 0.081 & 0.231 & 0.944 & 2.862\
V392 Car & A2V & 8850 $\pm$ 200 & 10263 & 4.296 $\pm$ 0.011 & 4.211 & 163.0 & & 0.097 & 0.108 & 1.019 & 2.889\
RS Cha & A8V & 8050 $\pm$ 200 & 7833 & 4.046 $\pm$ 0.022 & 4.150 & 30.0 & & 0.136 & 0.186 & 0.866 & 2.791\
MY Cyg & F0m & 7050 $\pm$ 200 & 7054 & 3.994 $\pm$ 0.019 & 3.882 & & & 0.219 & 0.226 & 0.709 & 2.756\
EI Cep & F3V & 6750 $\pm$ 100 & 6928 & 3.763 $\pm$ 0.014 & 3.904 & 16.2 & 0.27 & 0.234 & 0.199 & 0.658 & 2.712\
FS Mon & F2V & 6715 | 107 & 0.168 & 1.061 & 2.875\
V624 Her & A3 m & 8150 $ \pm$ 150 & 7902 & 3.832 $ \pm$ 0.014 & 3.794 & 38.0 & & 0.111 & 0.230 & 1.025 & 2.870\
V1647 Sgr & A1V & 9600 $ \pm$ 300 & 9142 & 4.252 $ \pm$ 0.008 & 4.087 & & & 0.040 & 0.174 & 1.020 & 2.899\
VV Pyx & A1V & 9500 $ \pm$ 200 & 9560 & 4.087 $ \pm$ 0.008 & 4.004 & 22.1 & & 0.028 & 0.161 & 1.013 & 2.881\
KW Hya & A5 m & 8000 $ \pm$ 200 & 8053 & 4.078 $ \pm$ 0.006 & 4.390 & 16.6 & & 0.122 & 0.232 & 0.832 & 2.827\
WW Aur & A5 m & 7960 $ \pm$ 420 & 8401 & 4.161 $ \pm$ 0.005 & 4.286 & 35.8 & & 0.081 & 0.231 & 0.944 & 2.862\
V392 Car & A2V & 8850 $ \pm$ 200 & 10263 & 4.296 $ \pm$ 0.011 & 4.211 & 163.0 & & 0.097 & 0.108 & 1.019 & 2.889\
RS Cha & A8V & 8050 $ \pm$ 200 & 7833 & 4.046 $ \pm$ 0.022 & 4.150 & 30.0 & & 0.136 & 0.186 & 0.866 & 2.791\
MY Cyg & F0 m & 7050 $ \pm$ 200 & 7054 & 3.994 $ \pm$ 0.019 & 3.882 & & & 0.219 & 0.226 & 0.709 & 2.756\
EI Cep & F3V & 6750 $ \pm$ 100 & 6928 & 3.763 $ \pm$ 0.014 & 3.904 & 16.2 & 0.27 & 0.234 & 0.199 & 0.658 & 2.712\
FS Mon & F2V & 6715 | 107 & 0.168 & 1.061 & 2.875\
V624 Her & A3m & 8150 $\pm$ 150 & 7902 & 3.832 $\po$ 0.014 & 3.794 & 38.0 & & 0.111 & 0.230 & 1.025 & 2.870\
R1647 Sgr & E1V & 9600 $\pm$ 300 & 9142 & 4.252 $\pm$ 0.008 & 4.087 & & & 0.040 & 0.174 & 1.020 & 2.899\
VV Pyx & A1V & 9500 $\pm$ 200 & 9560 & 4.087 $\pn$ 0.008 & 4.004 & 22.1 & & 0.028 & 0.161 & 1.013 & 2.881\
KW Hya & A5m & 8000 $\pm$ 200 & 8053 & 4.078 $\pm$ 0.006 & 4.390 & 16.6 & & 0.122 & 0.232 & 0.832 & 2.827\
WW Aur & A5m & 7960 $\pm$ 420 & 8401 & 4.161 $\'m$ 0.005 & 4.286 & 35.8 & & 0.081 & 0.231 & 0.944 & 2.862\
Y392 Car & A2V & 8850 $\pm$ 200 & 10263 & 4.296 $\pm$ 0.011 & 4.211 & 163.0 & & 0.097 & 0.108 & 1.019 & 2.889\
RS Cha & A8V & 8050 $\pm$ 200 & 7833 & 4.046 $\pm$ 0.022 & 4.150 & 30.0 & & 0.136 & 0.186 & 0.866 & 2.791\
LY Cyg & F0m & 7050 $\pi$ 200 & 7054 & 3.994 $\pm$ 0.019 & 3.882 & & & 0.219 & 0.226 & 0.709 & 2.756\
EI Cep & F3V & 6750 $\pm$ 100 & 6928 & 3.763 $\lm$ 0.014 & 3.904 & 16.2 & 0.27 & 0.234 & 0.199 & 0.658 & 2.712\
FS Kon & F2V & 6715 | 107 & 0.168 & 1.061 & 2.875\ & & 8150 150 & 7902 3.794 38.0 & & & 0.230 & & 2.870\ V1647 Sgr & A1V 9600 $\pm$ 300 & 9142 & 4.252 $\pm$ 0.008 & 4.087 & & 0.040 & 0.174 & 1.020 & 2.899\ VV Pyx & A1V & 9500 200 9560 4.087 0.008 & 4.004 & 22.1 & & 0.028 & 0.161 & 1.013 & 2.881\ KW Hya A5m & 8000 $\pm$ 200 & 8053 & $\pm$ 0.006 & 4.390 16.6 & & 0.122 & & & 2.827\ Aur A5m 7960 $\pm$ 420 8401 & 4.161 $\pm$ 0.005 & 4.286 & 35.8 & & 0.081 & 0.231 & 0.944 & V392 Car & 8850 200 10263 4.296 $\pm$ 0.011 & 163.0 & & 0.097 & & 2.889\ RS Cha & A8V & 8050 200 & & 4.046 $\pm$ 0.022 & 4.150 30.0 & & 0.136 & 0.186 & 0.866 2.791\ MY Cyg & F0m & 7050 $\pm$ 200 & 7054 & 3.994 $\pm$ 0.019 & & & 0.219 0.226 & 0.709 2.756\ Cep F3V 6750 $\pm$ & 6928 & 3.763 $\pm$ 0.014 & 3.904 & 16.2 & & 0.234 & 0.199 & 0.658 & 2.712\ FS Mon & | 107 & 0.168 & 1.061 & 2.875\
V624 Her & A3m & 8150 $\pm$ 150 & 7902 & 3.832 $\pm$ 0.014 & 3.794 & 38.0 & & 0.111 & 0.230 & 1.025 & 2.870\
V1647 Sgr & A1V & 9600 $\pm$ 300 & 9142 & 4.252 $\pm$ 0.008 & 4.087 & & & 0.040 & 0.174 & 1.020 & 2.899\
VV Pyx & a1V & 9500 $\pm$ 200 & 9560 & 4.087 $\pm$ 0.008 & 4.004 & 22.1 & & 0.028 & 0.161 & 1.013 & 2.881\
KW HyA & A5m & 8000 $\pm$ 200 & 8053 & 4.078 $\Pm$ 0.006 & 4.390 & 16.6 & & 0.122 & 0.232 & 0.832 & 2.827\
Ww AuR & A5M & 7960 $\pm$ 420 & 8401 & 4.161 $\pM$ 0.005 & 4.286 & 35.8 & & 0.081 & 0.231 & 0.944 & 2.862\
V392 CaR & A2V & 8850 $\pm$ 200 & 10263 & 4.296 $\pm$ 0.011 & 4.211 & 163.0 & & 0.097 & 0.108 & 1.019 & 2.889\
RS Cha & A8V & 8050 $\PM$ 200 & 7833 & 4.046 $\pm$ 0.022 & 4.150 & 30.0 & & 0.136 & 0.186 & 0.866 & 2.791\
My Cyg & F0m & 7050 $\pm$ 200 & 7054 & 3.994 $\pm$ 0.019 & 3.882 & & & 0.219 & 0.226 & 0.709 & 2.756\
EI Cep & F3V & 6750 $\pm$ 100 & 6928 & 3.763 $\pm$ 0.014 & 3.904 & 16.2 & 0.27 & 0.234 & 0.199 & 0.658 & 2.712\
Fs Mon & F2v & 6715 | 107 & 0.168 & 1.061 & 2.87 5\
V624 He r & A 3m& 8 15 0 $\ pm$150 & 7902 & 3 . 832$\pm$ 0.014 & 3.794 &38.0&& 0.1 1 1& 0.2 30 & 1. 0 25 & 2. 87 0\
V1 64 7 S gr &A1V & 9600 $\pm$ 300 &91 42 & 4.252 $ \ pm $ 0.008 &4.0 87 & & & 0.0 40& 0.17 4& 1 . 020 & 2. 899\VV Pyx & A1V& 9500 $\ pm $ 200 & 9560 &4 . 08 7 $\ pm$ 0.008 & 4.004 &2 2.1 & & 0.028& 0.16 1& 1 . 0 13& 2 .881\
KW H ya & A5 m & 8000 $\ p m $ 20 0 & 8053 & 4.0 78 $\pm$ 0. 0 06& 4.39 0& 1 6 .6 & & 0.12 2& 0. 232 & 0.832 & 2 .827\
WWAur &A 5m & 79 6 0 $\pm$ 420 & 84 01& 4. 1 61 $ \pm $0 .00 5 & 4. 2 86& 35.8 & & 0 .081& 0. 2 3 1 & 0. 944 & 2 .862\
V392 Car & A 2V& 88 5 0 $ \pm$200 & 102 63 & 4. 296 $\ pm$ 0 .0 11 & 4.211 & 16 3.0& & 0.097 &0. 108 & 1.01 9 & 2.8 89\
RS Cha &A8V & 8 0 50$\ p m $ 2 00 & 7833 & 4.046$\ p m $0.022 &4.150& 3 0. 0 & & 0.1 36 &0.18 6 & 0.8 66 & 2. 791\
MYCyg &F 0m & 7050 $ \p m$ 200 & 70 54& 3.9 9 4 $\ pm$ 0. 019 & 3. 882 & & & 0.219 & 0. 2 26 & 0.709 &2 .7 5 6 \E I Ce p & F3V & 6750 $\p m $ 10 0 &6 92 8 & 3.763 $\pm $0 .0 1 4 & 3.904 & 16.2 &0. 27 & 0 .234& 0.199 & 0.6 58 & 2.712 \ F S Mon &F2V& 6 7 15 | 107 &_0.168 &_1.061 & 2.875\
V624 Her_& A3m_&_8150 $\pm$_150_& 7902 &_3.832 $\pm$ 0.014_& 3.794 & 38.0_& & 0.111_&_0.230 & 1.025 & 2.870\
V1647 Sgr & A1V & 9600 $\pm$ 300 & 9142_&_4.252 $\pm$_0.008_&_4.087 & & & 0.040_& 0.174 & 1.020 &_2.899\
VV Pyx_& A1V & 9500 $\pm$ 200 & 9560_&_4.087 $\pm$ 0.008_& 4.004 & 22.1 & & 0.028 & 0.161_& 1.013 & 2.881\
KW Hya &_A5m & 8000_$\pm$_200_& 8053 & 4.078_$\pm$ 0.006 & 4.390 & 16.6_& & 0.122 & 0.232 &_0.832 & 2.827\
WW Aur & A5m &_7960 $\pm$ 420 & 8401 &_4.161 $\pm$ 0.005 & 4.286_& 35.8_& & 0.081 & 0.231_& 0.944 &_2.862\
V392 Car_& A2V &_8850 $\pm$ 200 & 10263 &_4.296 $\pm$ 0.011_& 4.211 & 163.0 & &_0.097_& 0.108 &_1.019_&_2.889\
RS Cha_& A8V &_8050_$\pm$ 200_&_7833 & 4.046 $\pm$ 0.022 &_4.150_& 30.0 & & 0.136 & 0.186_& 0.866 & 2.791\
MY_Cyg_& F0m & 7050_$\pm$ 200 & 7054 &_3.994 $\pm$ 0.019 & 3.882 &_& &_0.219 &_0.226 & 0.709 & 2.756\
EI Cep & F3V & 6750 $\pm$_100 & 6928 & 3.763 $\pm$_0.014 & 3.904 &_16.2 &_0.27_& 0.234 &_0.199_& 0.658_& 2.712\
FS Mon & F2V & 6715 |
sin 2\pi y - \sin 2\pi x$.]{} The left panel is a color map of the pairwise Pearson-correlation between the time-series $x_i$ and $x_j$. Due to the strong chaotic behaviour this analysis does not give any information on the community structures. In the right panel, we show the colormap of the noise correlation. The clusters are manifested in the correlation structure of the noise, namely, the four communities and the integrating cluster. []{data-label="doubling"}](dm_rc.pdf){width="0.8\linewidth"}
Logistic Maps on a Rich Club Network
------------------------------------
The dynamic of the logistic map and the coupling are described in the above section on the reconstruction for scale-free networks. We test the community detection approach on time-series using the correlation between time series and correlations between the noise. To this end, we fix the coupling strength $\alpha=10^{-4}$ and simulate the network dynamics. Then we apply our technique to recover the network structure. We show the results in Fig. \[logistic\].
![\[Color Online\] Reconstruction of rich club for logistic maps with Kuramoto coupling $h(x,y) = \sin 2\pi( y - x)$. The left panel shows a color map of the pairwise Pearson-correlation between the time-series $x_i$ and $x_j$. In the right panel, we show the colormap of the noise correlation that reveals the community structure and, in particular, the integrating cluster.[]{data-label="logistic"}](lm_rc.pdf){width="0.8\linewidth"}
Spiking neurons coupled with Electric Synapsis on a Rich-Club
-------------------------------------------------------------
Again, we consider the spiking neurons and the electrical coupling described above on the same rich-club motif as before. The reconstruction and analysis of the noise correlation is able to detect the rich-club clusters. To this end we fix the coupling strength $\alpha=5 \times 10^{-4}$ and simulate the network dynamics. Then we apply our technique to recover the network structure. We show the results in Fig. \[spiking\].
![\[Color Online\] Reconstruction of rich club for spiking neurons with electrical coupling. The left panel shows a color map of the pairwise Pearson-correlation between the time-series of the membrane potential. The right panel, shows the colormap of the noise correlation revealing the community structure and in particular, | sin 2\pi y - \sin 2\pi x$. ] { } The left panel is a color function of the pairwise Pearson - correlation coefficient between the prison term - series $ x_i$ and $ x_j$. Due to the impregnable chaotic demeanor this analysis does not give any data on the community structures. In the right jury, we show the colormap of the noise correlation coefficient. The clusters are manifested in the correlation structure of the randomness, namely, the four communities and the integrating bunch. [ ] { data - label="doubling"}](dm_rc.pdf){width="0.8\linewidth " }
Logistic Maps on a Rich Club Network
------------------------------------
The dynamic of the logistic function and the yoke are described in the above section on the reconstruction for scale - free network. We test the community detection approach on clock time - series using the correlation between time series and correlation between the noise. To this end, we fix the coupling strength $ \alpha=10^{-4}$ and simulate the network dynamics. Then we apply our technique to recover the network structure. We show the results in Fig. \[logistic\ ].
! [ \[Color Online\ ] Reconstruction of rich club for logistic map with Kuramoto coupling $ h(x, y) = \sin 2\pi (y - x)$. The left control panel testify a color map of the pairwise Pearson - correlation between the meter - series $ x_i$ and $ x_j$. In the right panel, we show the colormap of the randomness correlation that reveals the community structure and, in particular, the integrating cluster.[]{data - label="logistic"}](lm_rc.pdf){width="0.8\linewidth " }
Spiking neurons coupled with Electric Synapsis on a Rich - Club
-------------------------------------------------------------
Again, we consider the spiking neurons and the electric coupling identify above on the like full-bodied - club motif as before. The reconstruction and psychoanalysis of the noise correlation is able to detect the rich - club clusters. To this end we fix the coupling strength $ \alpha=5 \times 10^{-4}$ and model the network moral force. Then we apply our technique to recover the network structure. We prove the results in Fig. \[spiking\ ].
! [ \[Color Online\ ] Reconstruction of fat club for spike neurons with electrical yoke. The left panel shows a color map of the pairwise Pearson - correlation between the time - series of the membrane potential. The right panel, shows the colormap of the randomness correlation revealing the residential district structure and in particular, | sin 2\pi y - \sin 2\pi x$.]{} The left panel is a color map mf the pairwisd Pearson-correlation between tye tine-series $x_i$ and $x_j$. Due to the snrong chaitic vehaviour vgis analysis sles uov give any infotmation on tve community sdrjccures. In the right panel, we show the colormsp of the noise sorrtlaeion. Nht clusters are manifested in the dorrelauion structure of yhe noise, namely, the four fommknities and the inhegrating coustqe. []{data-label="duubling"}](dm_rb.'df){width="0.8\lingwidth"}
Logistic Maps on a Rich Cljb Necwork
------------------------------------
The dybanic mf the logiwtic iap and the gpuplinc are drscribed in thc abote swction on the reconstcuction for scale-freg networks. Wz test the community eerectimn a[proxxh un uimx-sedies udinj the corremation betwwen time series and cjgtelations befween ehq noise. To this end, we fix the coupling stdength $\alpha=10^{-4}$ and simulare the network dynamifs. Then wq apply our technique to recover the network struwture. De whjd tje results in Fig. \[logistic\].
![\[Color Online\] Reconseduvtpon of rich club nor logistic maps eihh Huramoto coupuing $h(r,g) = \sin 2\pi( y - x)$. The lfft pangl shoqs a colow mal of the pairwise Pearson-coerelation benweeb the time-series $x_n$ and $x_j$. In che ribht psnel, we show the colormcp of fhe noise clrrelatioh that reveals thd cpmkunity sufucture and, in pawticular, vhe iutegrativg ckuster.[]{qata-label="llgistlw"}](lm_rc.pdf){width="0.8\linewldth"}
Skiking neurons clupled with Electric Synapsis oi a Rich-Club
-------------------------------------------------------------
Abahn, fe consiber thc spiking neurogs and the eleetrical eouplivg describvd above mn the same rich-club mothv as before. Vhe reconftruxtiob and avxlysis of the moise corgejatiin is able to detegt thg dich-club clustexr. To this end we fib tre cpu[ning strengtv $\aloha=5 \yimes 10^{-4}$ and simulctd thr network dynamics. Tven se apply our technoqme to reciver the network struvture. We show the gesulvs in Hig. \[spokigg\].
![\[Color Online\] Reconstruction kf rich cpub for spiking neuvons with eleccrical coupling. The left panel shows a cilor map of the paurwise Pearson-correkstion betwexn the time-serias of the membrane pitential. The righu panel, shows the colorjap of the joise correlation revealing the community structure and in particular, | sin 2\pi y - \sin 2\pi x$.]{} panel a color of the pairwise and Due to the chaotic behaviour this does not give any information on community structures. In the right panel, we show the colormap of the noise The clusters are manifested in the correlation structure of the noise, namely, the communities the cluster. Logistic Maps on a Rich Club Network ------------------------------------ The dynamic of the logistic map and the are described in the above section on the for scale-free networks. We the community detection approach on using correlation between series correlations the noise. To end, we fix the coupling strength $\alpha=10^{-4}$ and simulate the network dynamics. Then we apply our technique recover the We show results Fig. ![\[Color Online\] Reconstruction club for logistic maps with Kuramoto \sin 2\pi( y - x)$. The left panel a color of the pairwise Pearson-correlation between the $x_i$ and $x_j$. In the right panel, we the colormap of the noise correlation that reveals the community structure and, in particular, the Spiking neurons coupled with Synapsis on a ------------------------------------------------------------- we the neurons and electrical coupling described above on the same rich-club motif as before. reconstruction and analysis of the noise correlation is able to rich-club To this end fix the coupling strength \times and simulate the network we our the structure. show the results in \[spiking\]. ![\[Color Online\] Reconstruction of club for spiking neurons panel shows a color map of the pairwise between the time-series of the membrane potential. right panel, shows the colormap of the noise correlation revealing the community and in | sin 2\pi y - \sin 2\pi x$.]{} The left panel iS a color map Of the PaiRwiSe pearSon-cOrrelation betwEEn thE time-series $x_i$ and $x_j$. Due tO the sTrONg chAOtIc behAviour tHIs ANAlySiS dOes NoT GiVe any InfOrmatioN on the commUniTy Structures. In THe Right panel, We sHow the colormAp oF the noIsE coRRelatIon. the clUsters ARe maniFested in tHe COrrelaTIon struCTUrE of tHe noise, namely, the fOUr COmmunities and tHe inteGrATiNG CluSteR. []{data-label="DoUblinG"}](Dm_rc.pdf){WIdTH="0.8\LIneWIdth"}
Logistic MAps on a Rich CLUb NEtwork
------------------------------------
thE dyNAmic of The loGiSTic Map and the coUpliNg are descRibed iN The abovE Section On the rEcoNstRuctIOn FoR scAlE-FreE NeTwoRKs. WE test the CoMmUnity DeteCTION appRoaCh on Time-sEries using the CorRelaTIon BetweEn timE serIeS and cOrrelaTions BeTween the noise. To This End, we fix tHe cOuPliNg StrenGTh $\alphA=10^{-4}$ anD siMulate tHe netwoRK dyNaMICS. THen we apply our technIqUE To Recover tHe netwORk StRUcture. We ShOw tHe reSULts in fig. \[lOGiStic\].
![\[ColoR OnlinE\] reCoNstructIoN of ricH cLub For LogisTIc maPs with kuramoto CouplINg $h(x,y) = \sin 2\pi( y - x)$. ThE Left panel showS A cOLOr MAp of The Pairwise PeaRson-COrreLatiON bEtwEEn the Time-sErIEs $X_I$ and $x_j$. In the right panEl, We show The coLormap of the noIse correlaTION that revEals THe COmmunity structUre anD, in particuLAr, the intEgratIng clustEr.[]{data-labEL="Logistic"}](Lm_rC.pdF){wiDth="0.8\LINeWidth"}
Spiking nEURons CoUpled wiTh ELectric synApsIs oN a RIcH-Club
-------------------------------------------------------------
AgaiN, we consiDeR tHe SpIkiNg neuROns and thE eLecTrIcaL coupLIng desCribeD aboVe On THe sAme rich-CLuB MOtif As BeFore. the ReConstRuctIOn aNd analySis of the nOisE CorrElAtIon is abLe to detect the RiCh-club clusTeRs. TO this eND We fix the Coupling strength $\alpha=5 \tiMEs 10^{-4}$ and siMulAte thE netWork dynamIcs. then we AppLY our teChniquE to reCoVer THE netwORK sTruCtUre. We show tHE ResUlts iN FIg. \[spIking\].
![\[CoLor Online\] ReconstruCTioN of rich club foR spIkinG NEuRonS WiTH elEcTRicAL Coupling. The left Panel shows A cOLoR map of the pAIrwIsE PearsoN-correlAtion BEtween tHe time-serIes of the mEmBranE POteNtial. The riGht panel, Shows the cOLormaP Of The noIse CorrelAtIon ReveaLing thE ComMunitY strucTuRe and iN partIcUlar, | sin 2\pi y - \sin 2\pi x$. ]{} The le ft pa nel is a col or m ap of the pair w isePearson-correlation be tween t h e ti m e- serie s $x_i$ an d $x_ j$ .Due t o t he st ron g chaot ic behavio urth is analysisd oe s not give an y informatio n o n theco mmu n ity s tru cture s. Int he rig ht panel, w e showt he colo r m ap ofthe noise correla t io n . The clusters are m an i fe s t edinthe correl at ion s t ructure of t h e n o ise, namely,the four co m mun itiesan d t h e inte grati ng clu ster. []{da ta-l abel="dou bling" } ](dm_rc . pdf){wi dth="0 .8\ lin ewid t h" }
Lo gi s tic Ma pso n a Rich Cl ub N etwor k
-- - - - - ---- --- ---- ----- ------------- -
Thed yna mic o f the log is tic m ap and theco upling are desc ribe d in theabo ve se ct ion o n the r eco nst ruction for sc a le- fr e e ne tworks. We test th ec o mm unity de tectio n a pp r oach onti me- seri e s usin g th e c orrelati on bet w ee ntime se ri es and c orr ela tions betw een th e noise. To t h is end, we fix the couplings tr e n gt h $\a lph a=10^{-4}$ands imul atet he ne t workdynam ic s .T hen we apply our te ch niqueto re cover the net work struc t u r e. We sh ow t h er esults in Fig. \[lo gistic\].![\[Colo r Onl ine\] Re construct i o n of ric h c lub fo r l o g is tic maps with K uram ot o coupl ing $h(x,y ) = \s in2\p i( y - x)$. The lef tpa ne lsho ws ac olor map o f t he pa irwis e Pears on-co rrel at io n be tween t h et i me-s er ie s $x _i$ a nd $x _j$. Inthe rig ht panel, we show t he colorm ap of the noi se correlati on th at rev e a ls the c ommunity structure and, in part icu lar,theintegrati ngcluste r.[ ] {data- label= "logi st ic" } ] (lm_r c . pd f){ wi dth="0.8\l i n ewi dth"}
Spik ing neu rons coupled withE lec tric Synapsis on a R i c h- Clu b
- - --- -- - --- - - --------------- ---------- -- - -- ---------- - --- -- --
Aga in, weconsi d er thespiking n eurons an dthee l ect rical coup ling des cribed ab o ve on th e sam e r ich-cl ub mo tif a s befo r e.The r econst ru ctionand a na lysis of the noise correlationis abl e todet ect the r ich - clu b cluster s. T o this end we fi x the co u pling str e ng th$ \alph a=5\ times 10^ { -4 }$a n dsimulate th e n etw ork d yna m ics. T henwe apply our tech n ique to recove r th e net wor k str uc ture. We showthe r e s ults inFi g. \[spikin g\].
![ \[ C olorOnline \] Rec onstruc t i on of ric h cl ubfor spiki ngne u rons wi th e l ectric al c ou pling. The l e ft p a n el shows a color mapo f thep air wisePe arson-c o rrel ation betw een the tim e-seri es o f the membra ne poten tia l. The right panel, sh ows t he colo rm ap o f t he noi se c o r relat ionre vea ling thec o mm u ni ty str uctu re an din p articular , | sin 2\pi_y -_\sin 2\pi x$.]{} The_left panel_is_a color_map_of the pairwise_Pearson-correlation between the_time-series $x_i$ and $x_j$._Due to the_strong_chaotic behaviour this analysis does not give any information on the community structures. In_the_right panel,_we_show_the colormap of the noise_correlation. The clusters are manifested_in the_correlation structure of the noise, namely, the four_communities_and the integrating_cluster. []{data-label="doubling"}](dm_rc.pdf){width="0.8\linewidth"}
Logistic Maps on a Rich Club Network
------------------------------------
The dynamic_of the logistic map and the_coupling are described_in_the_above section on the_reconstruction for scale-free networks. We test_the community detection approach on time-series_using the correlation between time series and_correlations between the noise. To this_end, we fix the coupling_strength $\alpha=10^{-4}$_and simulate the network dynamics._Then we apply_our technique_to recover the_network structure. We show the results_in Fig. \[logistic\].
![\[Color_Online\] Reconstruction of rich club for_logistic_maps with Kuramoto_coupling_$h(x,y)_= \sin_2\pi( y -_x)$._The left_panel_shows a color map of the_pairwise_Pearson-correlation between the time-series $x_i$ and $x_j$._In the right panel,_we_show the colormap of_the noise correlation that reveals_the community structure and, in particular,_the integrating_cluster.[]{data-label="logistic"}](lm_rc.pdf){width="0.8\linewidth"}
Spiking neurons_coupled with Electric Synapsis on a Rich-Club
-------------------------------------------------------------
Again, we consider the spiking_neurons and the electrical coupling described_above on the same_rich-club motif_as_before. The reconstruction_and_analysis of_the noise correlation is able to detect_the rich-club_clusters. To this end we fix_the coupling strength $\alpha=5_\times_10^{-4}$ and simulate the network dynamics._Then we apply our technique to_recover the network structure. We_show_the_results in Fig. \[spiking\].
![\[Color Online\]_Reconstruction of rich club for spiking_neurons with electrical_coupling. The left panel shows a color_map_of the pairwise Pearson-correlation between the_time-series_of the membrane potential. The right_panel,_shows_the colormap of the noise_correlation revealing the community structure and_in particular, |
}\\
\notag (g,h) &\mapsto \det[({N}(g)-{N}(h)), \nu(g), \nu(h)].\end{aligned}$$ Using equation \[nu1\] and theorem \[N\] we get that $$\begin{aligned}
\label{13}
F(\tilde{\phi _t}g, \tilde{\phi _t}h) = F(g,h)\end{aligned}$$ for all $t\in\mathbb{R}$. Again using equation \[nu2\] and theorem \[N\] we get that the neutralised section and the neutral section are equivariant under the action of $\Gamma$. Hence for all $\gamma$ in $\Gamma$ we have, $$\begin{aligned}
\label{18}
F(\gamma g,\gamma h) &= \det [({N}(\gamma g)-{N}(\gamma h)), \nu(\gamma g), \nu(\gamma h)] \\
\notag &= \det [\gamma({N}(g)-{N}(h)), \gamma \nu(g), \gamma \nu(h))] \\
\notag &= \det [\gamma] \det [({N}(g)-{N}(h)), \nu(g), \nu(h)] \\
\notag &= \det [({N}(g)-{N}(h)), \nu(g), \nu(h)] \\
\notag &= F(g, h).\end{aligned}$$ Now for a fixed real number $c_0$ we consider the space, $$\begin{aligned}
\mathfrak{K} {\mathrel{\mathop:}=}\lbrace (g_1,g_2) \mid d_{\mathsf{U}\mathbb{H}}(g_1,g_2) \leqslant c_0 \rbrace \subset \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H} \times \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H}.\end{aligned}$$ Compactness of $\mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\Sigma$ implies that $\mathfrak{K}_{\Gamma}$, the projection of $\mathfrak{K}$ in $\Gamma \backslash (\mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H} \times \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H})$, is compact. Now continuity of | } \\
\notag (g, h) & \mapsto \det[({N}(g)-{N}(h) ), \nu(g), \nu(h)].\end{aligned}$$ Using equation \[nu1\ ] and theorem \[N\ ] we get that $ $ \begin{aligned }
\label{13 }
F(\tilde{\phi _ t}g, \tilde{\phi _ t}h) = F(g, h)\end{aligned}$$ for all $ t\in\mathbb{R}$. Again using equality \[nu2\ ] and theorem \[N\ ] we make that the neutralised section and the achromatic department are equivariant under the action of $ \Gamma$. Hence for all $ \gamma$ in $ \Gamma$ we hold, $ $ \begin{aligned }
\label{18 }
F(\gamma g,\gamma h) & = \det [ ({ N}(\gamma g)-{N}(\gamma h) ), \nu(\gamma g), \nu(\gamma h) ] \\
\notag & = \det [ \gamma({N}(g)-{N}(h) ), \gamma \nu(g), \gamma \nu(h) ) ] \\
\notag & = \det [ \gamma ] \det [ ({ N}(g)-{N}(h) ), \nu(g), \nu(h) ] \\
\notag & = \det [ ({ N}(g)-{N}(h) ), \nu(g), \nu(h) ] \\
\notag & = F(g, h).\end{aligned}$$ Now for a fixed actual number $ c_0 $ we consider the quad, $ $ \begin{aligned }
\mathfrak{K } { \mathrel{\mathop:}=}\lbrace (g_1,g_2) \mid d_{\mathsf{U}\mathbb{H}}(g_1,g_2) \leqslant c_0 \rbrace \subset \mathsf{U}_{\hbox{\tiny $ \mathrm{rec}$}}\mathbb{H } \times \mathsf{U}_{\hbox{\tiny $ \mathrm{rec}$}}\mathbb{H}.\end{aligned}$$ Compactness of $ \mathsf{U}_{\hbox{\tiny $ \mathrm{rec}$}}\Sigma$ implies that $ \mathfrak{K}_{\Gamma}$, the projection of $ \mathfrak{K}$ in $ \Gamma \backslash (\mathsf{U}_{\hbox{\tiny $ \mathrm{rec}$}}\mathbb{H } \times \mathsf{U}_{\hbox{\tiny $ \mathrm{rec}$}}\mathbb{H})$, is compendious. Now continuity of | }\\
\notwg (g,h) &\mapsto \det[({N}(g)-{N}(h)), \nu(g), \nu(h)].\end{aligned}$$ Osung eqnation \[hu1\] and tfeorem \[N\] we get that $$\begin{almgnee}
\labtj{13}
F(\tilde{\phi _t}g, \tildd{\phi _t}h) = V(g,h)\end{alugnev}$$ for all $t\in\matigb{R}$. Agalu usihn equctmon \[nu2\] and theotem \[N\] we get that the neutsauided section and the neutral section are eqiigariant under jhe abtyon kf $\Gamma$. Hence for all $\gamma$ in $\Gajma$ we iave, $$\begin{alignrd}
\label{18}
F(\gamma g,\gamma h) &= \deh [({N}(\gwmma g)-{N}(\gamma h)), \nu(\gwmma g), \nu(\ganma r)] \\
\notag &= \det [\eamma({N}(g)-{N}(h)), \gamma \nu(g), \gajma \nu(h))] \\
\notag &= \det [\gamma] \det [({N}(g)-{V}(h)), \nu(y), \nu(h)] \\
\notag &= \eet [({T}(g)-{N}(h)), \nu(g), \nu(i)] \\
\notad &= F(g, h).\end{allbned}$$ Nmw for s fixed real nmmber $c_0$ qe consider the space, $$\begin{aligned}
\mathftak{K} {\mathral{\jathop:}=}\lbrace (g_1,g_2) \mud d_{\matvsf{U}\kathcv{H}}(g_1,e_2) \ltqskaht c_0 \rhrare \subset \mzthsf{U}_{\hbox{\tuny $\mathrm{rec}$}}\mathbb{N} \epkes \mathsf{U}_{\hgox{\tinr $\iathrm{rec}$}}\mathbb{H}.\end{aligned}$$ Compactness mf $\jathsf{U}_{\hbox{\tiny $\mathrm{rwc}$}}\Sigma$ implies that $\lathfrak{K}_{\Damma}$, the projection of $\mathfrak{K}$ in $\Gamma \backsldsh (\metfsf{B}_{\mnox{\tkby $\mathrm{rec}$}}\mathbb{H} \times \mathsf{U}_{\hbox{\tiny $\mathri{dev}$}}\msthbb{H})$, is compcct. Now continuiyy og | }\\ \notag (g,h) &\mapsto \det[({N}(g)-{N}(h)), \nu(g), \nu(h)].\end{aligned}$$ \[nu1\] theorem \[N\] get that $$\begin{aligned} = for all $t\in\mathbb{R}$. using equation \[nu2\] theorem \[N\] we get that the section and the neutral section are equivariant under the action of $\Gamma$. Hence all $\gamma$ in $\Gamma$ we have, $$\begin{aligned} \label{18} F(\gamma g,\gamma h) &= \det g)-{N}(\gamma \nu(\gamma \nu(\gamma \\ \notag &= \det [\gamma({N}(g)-{N}(h)), \gamma \nu(g), \gamma \nu(h))] \\ \notag &= \det [\gamma] \det [({N}(g)-{N}(h)), \nu(h)] \\ \notag &= \det [({N}(g)-{N}(h)), \nu(g), \nu(h)] \notag &= F(g, h).\end{aligned}$$ for a fixed real number we the space, \mathfrak{K} (g_1,g_2) d_{\mathsf{U}\mathbb{H}}(g_1,g_2) \leqslant c_0 \subset \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H} \times \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H}.\end{aligned}$$ Compactness of $\mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\Sigma$ implies that $\mathfrak{K}_{\Gamma}$, the projection of $\mathfrak{K}$ $\Gamma \backslash \times \mathsf{U}_{\hbox{\tiny is Now of | }\\
\notag (g,h) &\mapsto \det[({N}(g)-{N}(h)), \nu(g), \nu(H)].\end{aligneD}$$ UsinG eqUatIoN \[nu1\] aNd thEorem \[N\] we get thaT $$\BegiN{aligned}
\label{13}
F(\tilde{\phi _T}g, \tilDe{\PHi _t}h) = f(G,h)\End{alIgned}$$ foR AlL $T\In\mAtHbB{R}$. AGaIN uSing eQuaTion \[nu2\] aNd theorem \[N\] We gEt That the neutrALiSed section And The neutral seCtiOn are eQuIvaRIant uNdeR the aCtion oF $\gamma$. HEnce for alL $\gAMma$ in $\GAMma$ we haVE, $$\BeGin{aLigned}
\label{18}
F(\gamma G,\GaMMa h) &= \det [({N}(\gamma g)-{N}(\Gamma h)), \Nu(\GAmMA G), \nu(\GamMa h)] \\
\notag &= \deT [\gAmma({N}(G)-{n}(h)), \gamma \NU(g), \GAMMa \nU(H))] \\
\notag &= \det [\gammA] \det [({N}(g)-{N}(h)), \nu(g), \NU(h)] \\
\nOtag &= \deT [({N}(G)-{N}(h)), \NU(g), \nu(h)] \\
\nOtag &= F(G, h).\ENd{aLigned}$$ Now foR a fiXed real nuMber $c_0$ wE ConsideR The spacE, $$\begin{AliGneD}
\matHFrAk{k} {\maThREl{\mAThOp:}=}\lBRacE (g_1,g_2) \mid d_{\mAtHsF{U}\matHbb{H}}(G_1,G_2) \LEQslaNt c_0 \RbraCe \subSet \mathsf{U}_{\hboX{\tiNy $\maTHrm{Rec}$}}\maThbb{H} \TimeS \mAthsf{u}_{\hbox{\tIny $\maThRm{rec}$}}\mathbb{H}.\end{AligNed}$$ CompacTneSs Of $\mAtHsf{U}_{\hBOx{\tiny $\MatHrm{Rec}$}}\SigmA$ implieS ThaT $\mATHFrAk{K}_{\Gamma}$, the projectIoN OF $\mAthfrak{K}$ In $\GammA \BaCkSLash (\mathSf{u}_{\hbOx{\tiNY $\MathrM{rec}$}}\MAtHbb{H} \timeS \mathsF{u}_{\hBoX{\tiny $\maThRm{rec}$}}\mAtHbb{h})$, is CompaCT. Now ContinUity of | }\\
\notag (g,h) &\mapsto\det[({N}( g)-{N }(h )), \ nu(g ), \ nu(h)].\end{al i gned }$$ Using equation \[n u1\]an d the o re m \[N \] we g e tt h at$$ \b egi n{ a li gned}
\ label{1 3}
F(\tild e{\ ph i _t}g, \til d e{ \phi _t}h) =F(g,h)\end{a lig ned}$$ f ora ll $t \in \math bb{R}$ . Again using eq ua t ion \[ n u2\] an d th eore m \[N\] we get th a tt he neutralised secti on an d the ne utral sect io n are equivar i an t u nde r the action o f $\Gamma$. Hen ce for a ll$ \gamma $ in$\ G amm a$ we have, $$\ begin{ali gned}\ label{1 8 }
F(\ga mma g, \ga mma h)& =\d et[( { N}( \ ga mma g)- {N}(\gam ma h )), \ nu(\ g a m m a g) , \ nu(\ gamma h)] \\
\nota g & = \d e t [ \gamm a({N} (g)- {N }(h)) , \gam ma \n u( g), \gamma \nu( h))] \\
\nota g & =\de t[\gam m a] \de t [ ({N }(g)-{N }(h)),\ nu( g) , \ nu (h)] \\
\notag &=\d e t [ ({N}(g)- {N}(h) ) ,\n u (g), \nu (h )]\\
\ n o tag & = F( g ,h).\end{ aligne d }$ $Now for a fixed r eal nu mber$ c_0$ we co nsider t he sp a ce, $$\begin{a l igned}
\mathf r ak { K }{ \mat hre l{\mathop:} =}\l b race (g_ 1 ,g _2) \midd_{\m at h sf { U}\mathbb{H}}(g_1,g _2 ) \leq slant c_0 \rbrace\subset \m a t h sf{U}_{\ hbox { \t i ny $\mathrm{re c}$}} \mathbb{H} \times \ maths f{U}_{\h box{\tiny $ \mathrm{ rec }$} }\m ath b b {H }.\end{aligne d } $$ C om pactnes s o f $\mat hsf {U} _{\ hbo x{ \tiny $\m athrm{re c} $} }\ Si gma $ imp l ies that $ \ma th fra k{K}_ { \Gamma }$, t he p ro je c tio n of $\ m at h f rak{ K} $in $ \Ga mm a \ba cksl a sh(\maths f{U}_{\hb ox{ \ tiny $ \m athrm{r ec}$}}\mathbb {H } \times \ ma ths f{U}_{ \ h box{\tin y $\mathrm{rec}$}}\math b b{H})$, is comp act. Now cont inu ity of | }\\
\notag (g,h)_&\mapsto \det[({N}(g)-{N}(h)),_\nu(g), \nu(h)].\end{aligned}$$ Using equation_\[nu1\] and_theorem_\[N\] we_get_that $$\begin{aligned}
\label{13}
F(\tilde{\phi__t}g, \tilde{\phi _t}h)_= F(g,h)\end{aligned}$$ for all_$t\in\mathbb{R}$. Again using_equation_\[nu2\] and theorem \[N\] we get that the neutralised section and the neutral section_are_equivariant under_the_action_of $\Gamma$. Hence for all_$\gamma$ in $\Gamma$ we have,_$$\begin{aligned}
\label{18}
F(\gamma g,\gamma_h) &= \det [({N}(\gamma g)-{N}(\gamma h)), \nu(\gamma g),_\nu(\gamma_h)] \\
\notag &=_\det [\gamma({N}(g)-{N}(h)), \gamma \nu(g), \gamma \nu(h))] \\
\notag &= \det_[\gamma] \det [({N}(g)-{N}(h)), \nu(g), \nu(h)] \\
\notag_&= \det [({N}(g)-{N}(h)),_\nu(g),_\nu(h)]_\\
\notag &= F(g, h).\end{aligned}$$_Now for a fixed real number_$c_0$ we consider the space, $$\begin{aligned}
\mathfrak{K}_{\mathrel{\mathop:}=}\lbrace (g_1,g_2) \mid d_{\mathsf{U}\mathbb{H}}(g_1,g_2) \leqslant c_0 \rbrace_\subset \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H} \times \mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H}.\end{aligned}$$_Compactness of $\mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\Sigma$ implies_that $\mathfrak{K}_{\Gamma}$,_the projection of $\mathfrak{K}$ in_$\Gamma \backslash (\mathsf{U}_{\hbox{\tiny_$\mathrm{rec}$}}\mathbb{H} \times_\mathsf{U}_{\hbox{\tiny $\mathrm{rec}$}}\mathbb{H})$, is_compact. Now continuity of |
$10$ $407$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{27}}$ $10$ $-$ $77$ $27$ $9$
$11$ $150$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{29}}$ $11$ $4$ $84$ $33$ $7$
$12$ $66$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{31}}$ $12$ $9$ $98$ $36$ $4$
$13$ $32$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{33}}$ $13$ $32$ $98$ $36$ $4$
$14$ $9$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{35}}$ $14$ $66$ $94$ $33$ $7$
$15$ $4$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{37}}$ $15$ $150$ $86$ $27$ $9$
$16$ $-$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{39}}$ $16$ $407$ $75$ $11$ $2$ | $ 10 $ $ 407 $ $ 1 $ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{27}}$ $ 10 $ $ -$ $ 77 $ $ 27 $ $ 9 $
$ 11 $ $ 150 $ $ 1 $ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{29}}$ $ 11 $ $ 4 $ $ 84 $ $ 33 $ $ 7 $
$ 12 $ $ 66 $ $ -$ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{31}}$ $ 12 $ $ 9 $ $ 98 $ $ 36 $ $ 4 $
$ 13 $ $ 32 $ $ 1 $ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{33}}$ $ 13 $ $ 32 $ $ 98 $ $ 36 $ $ 4 $
$ 14 $ $ 9 $ $ 1 $ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{35}}$ $ 14 $ $ 66 $ $ 94 $ $ 33 $ $ 7 $
$ 15 $ $ 4 $ $ -$ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{37}}$ $ 15 $ $ 150 $ $ 86 $ $ 27 $ $ 9 $
$ 16 $ $ -$ $ -$ $ -$ $ -$
\[-2.3ex\ ]
\[-2.3ex\ ] $ { \mathsf{M}_{39}}$ $ 16 $ $ 407 $ $ 75 $ $ 11 $ $ 2 $ | $10$ $407$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\majhsf{M}_{27}}$ $10$ $-$ $77$ $27$ $9$
$11$ $150$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{29}}$ $11$ $4$ $84$ $33$ $7$
$12$ $66$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\matksf{N}_{31}}$ $12$ $9$ $98$ $36$ $4$
$13$ $32$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{O}_{33}}$ $13$ $32$ $98$ $36$ $4$
$14$ $9$ $1$ $-$ $-$
\[-2.3cx\]
\[-2.3ex\] ${\mathsf{M}_{35}}$ $14$ $66$ $94$ $33$ $7$
$15$ $4$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{37}}$ $15$ $150$ $86$ $27$ $9$
$16$ $-$ $-$ $-$ $-$
\[-2.3vx\]
\[-2.3cx\] ${\mathsf{M}_{39}}$ $16$ $407$ $75$ $11$ $2$ | $10$ $407$ $1$ $-$ $-$ \[-2.3ex\] \[-2.3ex\] $-$ $27$ $9$ $150$ $1$ $-$ $4$ $33$ $7$ $12$ $-$ $-$ $-$ \[-2.3ex\] ${\mathsf{M}_{31}}$ $12$ $9$ $98$ $36$ $13$ $32$ $1$ $-$ $-$ \[-2.3ex\] \[-2.3ex\] ${\mathsf{M}_{33}}$ $13$ $32$ $98$ $36$ $4$ $9$ $1$ $-$ $-$ \[-2.3ex\] \[-2.3ex\] ${\mathsf{M}_{35}}$ $14$ $66$ $94$ $33$ $7$ $15$ $-$ $-$ \[-2.3ex\] $15$ $150$ $86$ $27$ $9$ $16$ $-$ $-$ $-$ $-$ \[-2.3ex\] \[-2.3ex\] ${\mathsf{M}_{39}}$ $16$ $407$ $75$ $2$ | $10$ $407$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{27}}$ $10$ $-$ $77$ $27$ $9$
$11$ $150$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\mathsf{M}_{29}}$ $11$ $4$ $84$ $33$ $7$
$12$ $66$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3eX\] ${\mathsf{M}_{31}}$ $12$ $9$ $98$ $36$ $4$
$13$ $32$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3Ex\] ${\matHsf{m}_{33}}$ $13$ $32$ $98$ $36$ $4$
$14$ $9$ $1$ $-$ $-$
\[-2.3ex\]
\[-2.3Ex\] ${\MathSf{M}_{35}}$ $14$ $66$ $94$ $33$ $7$
$15$ $4$ $-$ $-$ $-$
\[-2.3eX\]
\[-2.3ex\] ${\mathsf{M}_{37}}$ $15$ $150$ $86$ $27$ $9$
$16$ $-$ $-$ $-$ $-$
\[-2.3ex\]
\[-2.3ex\] ${\MAthsF{M}_{39}}$ $16$ $407$ $75$ $11$ $2$ | $10$ $ 407$ $1 $ $-$ $-$
\[-2. 3ex \]
\[-2.3ex\]$ {\ m athsf{M}_{27}} $ $10$ $- $ $77$ $27$ $9$
$11$ $15 0$ $1 $ $- $ $-$
\[ -2 .3ex\ ]
\[- 2.3e x \ ] ${\ math s f{ M}_{29}} $ $1 1$ $4 $ $84$ $33$ $7$
$1 2 $ $66$ $ -$ $-$ $-$
\[- 2. 3ex\]
\[-2 .3ex \] $ { \ma thsf{M} _ {3 1 } }$ $12$ $9$ $ 98 $ $36$ $4$
$1 3$ $3 2$ $ 1$ $-$ $- $
\[-2. 3e x \]
\[-2.3ex \ ] ${\ m at hsf{M }_{ 33}}$ $ 13 $ $32$ $9 8$ $36$ $4 $
$ 14$ $ 9 $ $1$ $ - $ $ - $
\ [ -2.3e x\]
\ [-2.3ex\] ${\mat hsf{M } _ {35}} $ $ 14$ $66 $ $94$ $33$ $7$
$15$ $ 4$ $-$ $-$ $- $
\[- 2. 3ex\ ] \[ -2 . 3ex\] $ { \ mat hsf{M}_ { 37} }$ $15$ $150 $ $86$ $27$ $9$ $16$ $ -$ $ - $ $-$ $-$
\[- 2 . 3ex\]
\ [-2. 3 e x\] $ {\mat hsf{ M}_{3 9}}$ $16$ $ 407$ $75$ $11$ $ 2$ | _ _ _ __ __ $10$_ _ $407$ _ __ $1$ $-$__ ___ _$-$
_ _ __ _ _ _ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ __ ___
_ __ __ \[-2.3ex\]__ _ __ _ _ _ _ _ _ _ _ __ __ _ _ _ _ __
_ \[-2.3ex\] ${\mathsf{M}_{27}}$ _ ___ _ _ _ $10$__ $-$ __ _$77$__ _ $27$ _ $9$
_ _ ___ _ _ _ _ _ _ _ $11$ _$150$ __ $1$ ____ _ $-$_ _ _ _ _$-$
__ _ _ __ _ _ _ __ _ _ _ _ _ _ _ _ __ ____ _ _ _ _
_ _ _ \[-2.3ex\]_ _ _ _ _ __ _ _ __ ___ __ _____
_ \[-2.3ex\] ${\mathsf{M}_{29}}$ __ _ __ _ _ $11$__ ___$4$_ _ $84$ $33$__ _ $7$
__ _ _ _ _ _ _ _ __ _ $12$ _ $66$ _ $-$ $-$ _ _ $-$
_ _ _ _ __ _ _ _ __ _ _ _ _ _ _ _
_ __ \[-2.3ex\] _ _ _ ___ __ _ _ _ _ _ _ _ __ _ __ _ __ _
_ \[-2.3ex\]_${\mathsf{M}_{31}}$_ _ __ _ __ _ _ _ $12$ __ _ $9$ _ $98$ _ _ $36$_ _ _ __ $4$
_ _ __ _ _ _ _ _ _ ___ _$13$ $32$_ _ __ _$1$ _ _ __$-$ __ $-$
_ _ _ _ ___ _ __ _ _ _ ___ __ _ __ _
__ ___ _ \[-2.3ex\] _ _ _ _ _ _ _ _ _ __ _ _ _ _ __ _ _ _
_ \[-2.3ex\]_${\mathsf{M}_{33}}$_ _ _ _ _ _ _ $13$ _$32$ ___ $98$ _ _ $36$ $4$
_ _ _ _ _ __ __ _ _ $14$ _ $9$ ___ _$1$ _ $-$ _ _$-$
____ _ _ _ __ __ _ _ __ _ _ _ _ _ _ _ _
_ __ \[-2.3ex\] _ _ _ _ _ __ _ ___ _ __ _ _ _
_\[-2.3ex\] ${\mathsf{M}_{35}}$ _ _ _ _ _ _ _ $14$ _ _ $66$ _ $94$_ _$33$__ _ _ $7$
_ __ _ _ _ __ _ ___ __ _ _ _ _ __ _ $15$ _ $4$ _ _ _ $-$_ $-$ _ _$-$
__ _ _ ____ _ __ _ _ _ _ _ ___ _ _ _ _ _ _ _ _
__ _ \[-2.3ex\] _ _ __ _ _ _ _ _ ____ _ _ ___ _ __ __
_ \[-2.3ex\] ${\mathsf{M}_{37}}$ ___ _ _ _ $15$ _$150$ __ $86$ __ __ $27$ _ _ _ $9$
_ _ _ _ _ __ _ _ _ _ _ $16$ _ $-$ _ $-$ _ _$-$ $-$
__ _ _ _ _ _ ___ _ _ _ _ _ __ _ ___ _ _ _ _ _ __ __ _ __
_ _ \[-2.3ex\] _ _____ _ _ _ _ _ _ _ _ __ _ _ _ _ _
\[-2.3ex\] ${\mathsf{M}_{39}}$_ _ _ __ _ __ _ _ _ _ $16$_ $407$_ __ $75$ _ _ $11$ _ $2$ |
the Subaru telescope, we also investigate the case of $m_{\rm d} = 73 M_{\oplus}$ here.
### Calculation of the simulated TTV
Orbits of the planets are integrated using the fourth-order Hermite scheme with the shared time step [@2004PASJ...56..861K]. The time step (typically $\sim 0.025$ days) is chosen so that the fractional energy change due to the integrator during $\sim 1000$ days integration should always be smaller than $10^{-9}$. All the simulations presented in this section integrate planetary orbits beginning at the same epoch $T_0 (\mathrm{BJD}) = 2455189.1703$ (the first transit time of KOI-94d), until $\mathrm{BJD} = 2456133.0$ (approximately the last transit time of KOI-94d we analyzed).
For each planet, the initial value of the orbital inclination $i$ is fixed at the value obtained from $a/R_{\star}$ and $b$ (see Table \[ppar\_fit\]), assuming that $e = 0$. Here, all the $i$ values are chosen in the range $[0, \pi/2]$.[^2] The values of $\Omega_{\rm d}$ and $\Omega_{\rm e}$ are fixed at those in Table \[ppar\_prior\]. Since we have no information on the nodal angle of KOI-94c, we assume $\Omega_{\rm c} = 0$ deg. Initial semi-major axes are calculated via Kepler’s third law with $M_{\star} = 1.25 M_{\odot}$ [@2012ApJ...759L..36H], orbital periods in Table \[ep\], and planetary masses adopted in each simulation. The phases of the planets are determined from the transit ephemerides: for each planet, we convert the transit time closest to $T_0$ into the sum of the argument of periastron and mean anomaly $\omega + f$, taking account of the non-zero eccentricity if any, and then move it backward in time to $T_0$, assuming a Keplerian orbit.
The mid-transit times of each planet are determined by minimizing the sky-plane distance $D$ between the star and the planet, where the roots of the time derivative of $D$ are found by the Newton-Raphson method [@2010arXiv1006.3834F]. Then these transit times are | the Subaru telescope, we also investigate the case of $ m_{\rm d } = 73 M_{\oplus}$ here.
# # # Calculation of the simulated TTV
Orbits of the planets are integrate use the fourth - order Hermite scheme with the shared fourth dimension step [ @2004PASJ... 56.. 861 K ]. The time footfall (typically $ \sim 0.025 $ day) is chosen so that the fractional energy change due to the integrator during $ \sim 1000 $ days consolidation should always be smaller than $ 10^{-9}$. All the simulations presented in this incision integrate planetary orbits beginning at the same epoch $ T_0 (\mathrm{BJD }) = 2455189.1703 $ (the inaugural transit time of KOI-94d), until $ \mathrm{BJD } = 2456133.0 $ (approximately the last passage time of KOI-94d we analyzed).
For each planet, the initial value of the orbital dip $ i$ is fixed at the value obtained from $ a / R_{\star}$ and $ b$ (see Table \[ppar\_fit\ ]), assuming that $ east = 0$. Here, all the $ i$ values are chosen in the range $ [ 0, \pi/2]$.[^2 ] The values of $ \Omega_{\rm d}$ and $ \Omega_{\rm e}$ are fixed at those in Table \[ppar\_prior\ ]. Since we have no information on the nodal angle of KOI-94c, we assume $ \Omega_{\rm c } = 0 $ deg. Initial semi - major axes are calculated via Kepler ’s third law with $ M_{\star } = 1.25 M_{\odot}$ [ @2012ApJ... 759L.. 36H ], orbital periods in Table \[ep\ ], and planetary masses adopted in each simulation. The phase of the planets are determined from the passage ephemerides: for each planet, we convert the transit time close to $ T_0 $ into the sum of the argument of periastron and mean anomaly $ \omega + f$, taking bill of the non - zero eccentricity if any, and then move it backward in time to $ T_0 $, assuming a Keplerian orbit.
The mid - transit time of each planet are determined by minimizing the sky - plane distance $ D$ between the star and the planet, where the solution of the time derivative of $ D$ are found by the Newton - Raphson method [ @2010arXiv1006.3834F ]. Then these transit time are | thf Subaru telescope, we alro investigate jhw case of $m_{\rj d} = 73 M_{\oolus}$ here.
### Calculation of the sumulauvd TTV
Orbits of the puanets arv integrared nsing the fourth-order Heviite dchekx with the shargd time step [@2004PASJ...56..861K]. The tima rtzp (typically $\sim 0.025$ days) is chosen so ehat thr vractional enetgy cnwnge due to the integrator during $\sim 1000$ days iitegration shoukd always be smaller than $10^{-9}$. All the simulations pgesented in thif section intdgrate planetary orbita beginning at the same epoch $T_0 (\mathxm{BJD}) = 2455189.1703$ (the fursh transit tine of KOI-94d), until $\mathrm{BJG} = 2456133.0$ (apptoximately the lavt rransit time of KOI-94d xe analyzed).
For each klanet, the iuitial value of the oebutal hnclhnatkin $k$ ia hixsd at hhe value obtzined from $q/R_{\star}$ and $b$ (see Tanlq \[ppar\_fit\]), assujing trae $e = 0$. Here, all the $i$ values are chosen pn tge range $[0, \pi/2]$.[^2] The values of $\Omega_{\rm d}$ and $\Omeha_{\rm e}$ arq fixed at those in Table \[ppar\_prior\]. Since we have no iifurmcblon ub hhe nodal angle of KOI-94c, we assume $\Omega_{\rm c} = 0$ seb. Pnitial semi-major axes are cslfukwted via Keplgr’s thixs maw with $M_{\star} = 1.25 M_{\ldot}$ [@2012ApT...759L..36H], oebital pewiodx in Table \[ep\], and planetary masses adopned un each simulation. The phases of jhe plsnets are determined frum tge transit fphemeridsr: for each planeg, wv cotvert tht transit time clofest to $T_0$ intp the sjm og the wrgument ov perldstron and mean anlmaly $\okega + f$, taning account of the non-zero eccxitricity if amy, ang then mjve ib backward in tyme to $T_0$, assumnng a Ke'leriav orbit.
The mid-traisit times os each planet wre determinxd by minymizung rhe sky-ouane distance $C$ between the star qnd the planet, wheve thg doots of the tine eerivative of $D$ ard fjujd br the Newton-Rdphsun orthod [@2010arXiv1006.3834F]. Thek tfese transit times are | the Subaru telescope, we also investigate the $m_{\rm = 73 here. ### Calculation of planets are integrated the fourth-order Hermite with the shared time step [@2004PASJ...56..861K]. time step (typically $\sim 0.025$ days) is chosen so that the fractional energy due to the integrator during $\sim 1000$ days integration should always be smaller $10^{-9}$. the presented this section integrate planetary orbits beginning at the same epoch $T_0 (\mathrm{BJD}) = 2455189.1703$ (the first time of KOI-94d), until $\mathrm{BJD} = 2456133.0$ (approximately last transit time of we analyzed). For each planet, initial of the inclination is at the value from $a/R_{\star}$ and $b$ (see Table \[ppar\_fit\]), assuming that $e = 0$. Here, all the $i$ values chosen in $[0, \pi/2]$.[^2] values $\Omega_{\rm and $\Omega_{\rm e}$ at those in Table \[ppar\_prior\]. Since information on the nodal angle of KOI-94c, we $\Omega_{\rm c} 0$ deg. Initial semi-major axes are via Kepler’s third law with $M_{\star} = 1.25 [@2012ApJ...759L..36H], orbital periods in Table \[ep\], and planetary masses adopted in each simulation. The phases planets are determined from transit ephemerides: for planet, convert transit closest to into the sum of the argument of periastron and mean anomaly + f$, taking account of the non-zero eccentricity if any, move backward in time $T_0$, assuming a Keplerian The times of each planet by the between star the planet, where the of the time derivative of are found by the transit times are | the Subaru telescope, we also iNvestigate The caSe oF $m_{\rM d} = 73 m_{\oplUs}$ heRe.
### Calculation oF The sImulated TTV
Orbits of the PlaneTs ARe inTEgRated Using thE FoURTh-oRdEr herMiTE sCheme WitH the shaRed time steP [@2004PAsJ...56..861k]. The time step (TYpIcally $\sim 0.025$ dAys) Is chosen so thAt tHe fracTiOnaL EnergY chAnge dUe to thE IntegrAtor durinG $\sIM 1000$ days iNTegratiON ShOuld Always be smaller thAN $10^{-9}$. ALL the simulationS preseNtED iN THis SecTion integrAtE planETary orbITs BEGInnINg at the same epOch $T_0 (\mathrm{Bjd}) = 2455189.1703$ (thE first TrAnsIT time oF KOI-94d), UnTIl $\mAthrm{BJD} = 2456133.0$ (appRoxiMately the Last trANsit timE Of KOI-94d wE analyZed).
for Each PLaNeT, thE iNItiAL vAluE Of tHe orbitaL iNcLinatIon $i$ IS FIXed aT thE valUe obtAined from $a/R_{\stAr}$ aNd $b$ (sEE TaBle \[ppAr\_fit\]), AssuMiNg thaT $e = 0$. Here, All thE $i$ Values are chosen In thE range $[0, \pi/2]$.[^2] THe vAlUes Of $\omega_{\RM d}$ and $\OMegA_{\rm E}$ are fixEd at thoSE in taBLE \[PpAr\_prior\]. Since we have No INFoRmation oN the noDAl AnGLe of KOI-94c, We AssUme $\OMEGa_{\rm c} = 0$ Deg. INItIal semi-mAjor axES aRe CalculaTeD via KePlEr’s ThiRd law WIth $M_{\Star} = 1.25 M_{\oDot}$ [@2012ApJ...759L..36H], OrbitAL periods in TablE \[Ep\], and planetarY MaSSEs ADoptEd iN each simulaTion. tHe phAses OF tHe pLAnets Are deTeRMiNEd from the transit ephEmErides: For eaCh planet, we conVert the traNSIT time cloSest TO $T_0$ INto the sum of the ArgumEnt of periaSTron and mEan anOmaly $\omeGa + f$, taking ACCount of tHe nOn-zEro EccENTrIcity if any, and THEn moVe It backwArd In time tO $T_0$, aSsuMinG a KEpLerian orbIt.
The mid-TrAnSiT tImeS of eaCH planet aRe DetErMinEd by mINimiziNg the Sky-pLaNe DIstAnce $D$ beTWeEN The sTaR aNd thE plAnEt, wheRe thE RooTs of the Time derivAtiVE of $D$ ArE fOund by tHe Newton-RaphsOn Method [@2010arXiV1006.3834F]. theN these TRAnsit timEs are | the Subaru telescope, wealso inves tigat e t heca se o f $m _{\rm d} = 73M _{\o plus}$ here.
### Calc ulati on of t h esimul ated TT V
O r bit sof th ep la netsare integr ated using th efourth-order He rmite sche mewith the sha red timest ep[ @2004 PAS J...5 6..861 K ]. The time ste p( typica l ly $\si m 0. 025$ days) is chosens ot hat the fracti onal e ne r gy c han gedue to the i ntegr a tor dur i ng $ \ sim 1000$ days in tegration s h oul d alwa ys be smalle r tha n$ 10^ {-9}$. Allthesimulatio ns pre s ented i n this s ection in teg rate pl an eta ry orb i ts be g inn ing at t he s ame e poch $ T _ 0 (\ mat hrm{ BJD}) = 2455189.17 03$ (th e fi rst t ransi t ti me of K OI-94d ), un ti l $\mathrm{BJD} = 2 456133.0$ (a pp rox im ately the la sttra nsit ti me of K O I-9 4d w e a nalyzed).
For eac hp l an et, theinitia l v al u e of the o rbi tali n clina tion $i $ is fix ed att he v alue ob ta ined f ro m $ a/R _{\st a r}$and $b $ (see T able\ [ppar\_fit\]), assuming that $e = 0 $ . He re, all the $i $ va l uesarec ho sen in th e ran ge $[ 0 , \pi/2]$.[^2] Theva lues o f $\O mega_{\rm d}$ and $\Ome g a _ {\rm e}$ are fi x ed at those in Tabl e \[ppar\_ p rior\].Since we have no infor m a tion onthe no dal an g l eof KOI-94c, w e assu me $\Omeg a_{ \rm c}= 0 $ d eg. In it ial semi- major ax es a re c alc ulate d via Kep le r’s t hir d law with $ M_{\s tar} = 1 . 25M_{\odo t }$ [ @201 2A pJ ...7 59L .. 36H], orb i tal period s in Tabl e \ [ ep\] ,an d plane tary masses a do pted in ea ch si mulati o n . The ph ases of the planets are determi ned from the transiteph emerid es: for ea ch pla net,we co n v ert t h e t ran si t time clo s e stto $T _0 $ in to thesum of the argumen t of periastron a ndmean a no mal y $ \ ome ga + f $ , taking account of the no n- z er o eccentri c ity i f any,and the n mov e it bac kward intime to $ T_ 0$,a s sum ing a Kepl erian or bit.
The mid-t r an sit t ime s of e ac h p lanet are d e ter mined by mi ni mizing thesk y-planedistance $D$ between th e star andthe planet,whe r e t he rootsof t he time de riv ati ve of $D $ arefoun d b y t h e New ton- R aphson me t ho d [ @ 2 01 0arXiv1006. 3 8 3 4F] . The n t h ese tr ansi t times are | the_Subaru telescope,_we also investigate the_case of_$m_{\rm_d} =_73_M_{\oplus}$ here.
### Calculation_of the simulated_TTV
Orbits of the planets_are integrated using_the_fourth-order Hermite scheme with the shared time step [@2004PASJ...56..861K]. The time step (typically $\sim_0.025$_days) is_chosen_so_that the fractional energy change_due to the integrator during_$\sim 1000$_days integration should always be smaller than $10^{-9}$._All_the simulations presented_in this section integrate planetary orbits beginning at the_same epoch $T_0 (\mathrm{BJD}) = 2455189.1703$_(the first transit_time_of_KOI-94d), until $\mathrm{BJD} =_2456133.0$ (approximately the last transit time_of KOI-94d we analyzed).
For each planet,_the initial value of the orbital inclination_$i$ is fixed at the value_obtained from $a/R_{\star}$ and $b$_(see Table_\[ppar\_fit\]), assuming that $e =_0$. Here, all_the $i$_values are chosen_in the range $[0, \pi/2]$.[^2] The_values of $\Omega_{\rm_d}$ and $\Omega_{\rm e}$ are fixed_at_those in Table_\[ppar\_prior\]._Since_we have_no information on_the_nodal angle_of_KOI-94c, we assume $\Omega_{\rm c} =_0$_deg. Initial semi-major axes are calculated via_Kepler’s third law with_$M_{\star}_= 1.25 M_{\odot}$ [@2012ApJ...759L..36H],_orbital periods in Table \[ep\],_and planetary masses adopted in each_simulation. The_phases of_the planets are determined from the transit ephemerides: for each planet,_we convert the transit time closest_to $T_0$ into the_sum of_the_argument of periastron_and_mean anomaly_$\omega + f$, taking account of the_non-zero eccentricity_if any, and then move it_backward in time to_$T_0$,_assuming a Keplerian orbit.
The mid-transit times_of each planet are determined by_minimizing the sky-plane distance $D$_between_the_star and the planet, where_the roots of the time derivative_of $D$ are_found by the Newton-Raphson method [@2010arXiv1006.3834F]. Then_these_transit times are |
frac{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\right).
\end{aligned}$$ Since $t<0$, we have $$\mathrm{e}^{\frac{|x|^2}{4t}}<1.$$ Hence $$\overline{u}_t(x, t)-\Delta_{\infty}^N\overline{u}(x, t) >\mathrm{e}^{\frac{-|x|^2}{4t}}\frac{1}{t|\log|t||^{\delta+1}}\left(\frac{-(\delta+1)}{2|\log|t||}-\frac{1}{4}+\delta\right).$$ Setting $\delta =1/4$ ensures that this is positive.
Then assume $x=0$. We need to check that for every $\phi \in C^2(\Omega)$ so that $u-\phi$ has a minimum at $(0, t)$, we have $$\label{eq:viscdef}
\phi_t(0, t) \geq \lambda(D^2\phi(0, t)).$$ We first show that $\overline{u}$ itself satisfies. Inserting $x=0$ into and, we see that we must have $$f'(t)+g'(t) \geq \lambda(-\frac{f(t)}{2t}\mathbb{Id}_n) =-\frac{f(t)}{2t} = \frac{1}{4t|\log|t||^{\delta+1}}.$$ Writing out the left hand side, we get $$\frac{1}{t|\log|t||^{\delta+1}}\left(-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta\right) \geq \frac{1}{4t|\log|t||^{\delta+1}},$$ which is equivalent to $$-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta \leq \frac{1}{4}.$$ Since $\delta$ was set to $1/4$, we end up with $$-\frac{5}{8}\frac{1}{|\log|t||} \leq 0,$$ and so $\overline{u}$ itself satisfies in $\Omega$. Let $\phi$ be the test function in Definition \[def:viscositydefinition\] so that $\overline{u}-\phi$ has a minimum at $(0, t)$. This implies that $$ | frac{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\right).
\end{aligned}$$ Since $ t<0 $, we have $ $ \mathrm{e}^{\frac{|x|^2}{4t}}<1.$$ Hence $ $ \overline{u}_t(x, t)-\Delta_{\infty}^N\overline{u}(x, t) > \mathrm{e}^{\frac{-|x|^2}{4t}}\frac{1}{t|\log|t||^{\delta+1}}\left(\frac{-(\delta+1)}{2|\log|t||}-\frac{1}{4}+\delta\right).$$ Setting $ \delta = 1/4 $ ensures that this is incontrovertible.
Then wear $ x=0$. We need to check that for every $ \phi \in C^2(\Omega)$ so that $ u-\phi$ have a minimum at $ (0, t)$, we have $ $ \label{eq: viscdef }
\phi_t(0, t) \geq \lambda(D^2\phi(0, t)).$$ We first express that $ \overline{u}$ itself satisfies. slip in $ x=0 $ into and, we see that we must suffer $ $ f'(t)+g'(t) \geq \lambda(-\frac{f(t)}{2t}\mathbb{Id}_n) = -\frac{f(t)}{2 t } = \frac{1}{4t|\log|t||^{\delta+1}}.$$ Writing out the leftover hand side, we get $ $ \frac{1}{t|\log|t||^{\delta+1}}\left(-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta\right) \geq \frac{1}{4t|\log|t||^{\delta+1}},$$ which is equivalent to $ $ -\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta \leq \frac{1}{4}.$$ Since $ \delta$ was set to $ 1/4 $, we end up with $ $ -\frac{5}{8}\frac{1}{|\log|t|| } \leq 0,$$ and therefore $ \overline{u}$ itself satisfies in $ \Omega$. Let $ \phi$ be the test function in Definition \[def: viscositydefinition\ ] so that $ \overline{u}-\phi$ has a minimum at $ (0, t)$. This implies that $ $ | fraf{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\right).
\end{aligned}$$ Since $t<0$, we habe $$\mathro{e}^{\frac{|x|^2}{4t}}<1.$$ Hence $$\overline{u}_t(x, t)-\Veltq_{\inftt}^N\overline{u}(x, t) >\mathrm{e}^{\wrac{-|x|^2}{4t}}\frab{1}{t|\log|t||^{\delra+1}}\leht(\frac{-(\delta+1)}{2|\log|t||}-\fczc{1}{4}+\delta\vnght).$$ Acttiny $\velta =1/4$ ensures jhat this is positive.
Then dsrule $x=0$. We need to check that for everr $\phi \im F^2(\Omega)$ so that $u-\php$ ras z minimum at $(0, t)$, we have $$\label{eq:visddef}
\phi_t(0, t) \geq \lsmbda(D^2\phi(0, t)).$$ We first show hhat $\overline{u}$ itself datisfies. Ibsereung $x=0$ into avd, we see uhct we must gave $$f'(t)+g'(t) \geq \lambda(-\frac{f(t)}{2t}\mathcb{Id}_n) =-\frac{f(t)}{2t} = \drqc{1}{4t|\pmg|t||^{\delta+1}}.$$ Wrmting jut the left hand sige, we grt $$\frac{1}{t|\log|t||^{\dekta+1}}\nefr(-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delva\right) \geq \frac{1}{4t|\lod|t||^{\delta+1}},$$ wvieh is equivalent to $$-\feax{\deltd+1}{2}\fraw{1}{|\log|g||}+\eelga \mea \rrac{1}{4}.$$ Slncx $\delta$ was set to $1/4$, we end up with $$-\frac{5}{8}\frsc{1}{|\jig|t||} \leq 0,$$ and ao $\ovewlyne{u}$ itself satisfies in $\Omega$. Let $\phi$ te fhe test function in Dedinition \[def:viscosityfefinitiog\] so that $\overline{u}-\phi$ has a minimum at $(0, t)$. This ikpliea thcb $$ | frac{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\right). \end{aligned}$$ Since $t<0$, we have $$\mathrm{e}^{\frac{|x|^2}{4t}}<1.$$ t)-\Delta_{\infty}^N\overline{u}(x, >\mathrm{e}^{\frac{-|x|^2}{4t}}\frac{1}{t|\log|t||^{\delta+1}}\left(\frac{-(\delta+1)}{2|\log|t||}-\frac{1}{4}+\delta\right).$$ Setting =1/4$ ensures that $x=0$. need to check for every $\phi C^2(\Omega)$ so that $u-\phi$ has a at $(0, t)$, we have $$\label{eq:viscdef} \phi_t(0, t) \geq \lambda(D^2\phi(0, t)).$$ We first that $\overline{u}$ itself satisfies. Inserting $x=0$ into and, we see that we must $$f'(t)+g'(t) \lambda(-\frac{f(t)}{2t}\mathbb{Id}_n) = Writing out the left hand side, we get $$\frac{1}{t|\log|t||^{\delta+1}}\left(-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta\right) \geq \frac{1}{4t|\log|t||^{\delta+1}},$$ which is equivalent to $$-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta \frac{1}{4}.$$ Since $\delta$ was set to $1/4$, we up with $$-\frac{5}{8}\frac{1}{|\log|t||} \leq and so $\overline{u}$ itself satisfies $\Omega$. $\phi$ be test in \[def:viscositydefinition\] so that has a minimum at $(0, t)$. This implies that $$ | frac{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\rigHt).
\end{alignEd}$$ SinCe $t<0$, We hAvE $$\matHrm{e}^{\Frac{|x|^2}{4t}}<1.$$ Hence $$\oveRLine{U}_t(x, t)-\Delta_{\infty}^N\overlinE{u}(x, t) >\mAtHRm{e}^{\fRAc{-|X|^2}{4t}}\fraC{1}{t|\log|t||^{\dELtA+1}}\LEft(\FrAc{-(\DelTa+1)}{2|\LOg|T||}-\frac{1}{4}+\DelTa\right).$$ setting $\delTa =1/4$ eNsUres that this IS pOsitive.
TheN asSume $x=0$. We need tO chEck thaT fOr eVEry $\phI \in c^2(\OmegA)$ so thaT $U-\phi$ haS a minimum At $(0, T)$, We have $$\LAbel{eq:vISCdEf}
\phI_t(0, t) \geq \lambda(D^2\phi(0, t)).$$ wE fIRst show that $\oveRline{u}$ ItSElF SAtiSfiEs. InsertinG $x=0$ Into aND, we see tHAt WE MUst HAve $$f'(t)+g'(t) \geq \lamBda(-\frac{f(t)}{2t}\mAThbB{Id}_n) =-\frAc{F(t)}{2t} = \FRac{1}{4t|\loG|t||^{\delTa+1}}.$$ wRitIng out the leFt haNd side, we gEt $$\frac{1}{T|\Log|t||^{\delTA+1}}\left(-\frAc{\deltA+1}{2}\frAc{1}{|\lOg|t||}+\dELtA\rIghT) \gEQ \frAC{1}{4t|\Log|T||^{\DelTa+1}},$$ which iS eQuIvaleNt to $$-\FRAC{\DeltA+1}{2}\frAc{1}{|\loG|t||}+\delTa \leq \frac{1}{4}.$$ SincE $\deLta$ wAS seT to $1/4$, we End up With $$-\FrAc{5}{8}\fraC{1}{|\log|t||} \lEq 0,$$ and So $\Overline{u}$ itself SatiSfies in $\OmEga$. leT $\phI$ bE the tESt funcTioN in definitIon \[def:vIScoSiTYDEfInition\] so that $\overlInE{U}-\PhI$ has a minImum at $(0, T)$. thIs IMplies thAt $$ | frac{1}{4}+\delta\mathrm{e }^{\frac{| x|^2} {4t }}\ ri ght) .
\end{aligned } $$ S ince $t<0$, we have $$ \math rm { e}^{ \ fr ac{|x |^2}{4t } }< 1 . $$He nc e $ $\ o ve rline {u} _t(x, t )-\Delta_{ \in ft y}^N\overlin e {u }(x, t) >\ mat hrm{e}^{\fra c{- |x|^2} {4 t}} \ frac{ 1}{ t|\lo g|t||^ { \delta +1}}\left (\ f rac{-( \ delta+1 ) } {2 |\lo g|t||}-\frac{1}{4 } +\ d elta\right).$$ Setti ng $\ d e lta =1 /4$ ensure sthatt his isp os i t i ve.
Then assume$x=0$. We n e edto che ck th a t forevery $ \ phi \in C^2(\O mega )$ so tha t $u-\ p hi$ has a minim um at$(0 , t )$,w eha ve$$ \ lab e l{ eq: v isc def}
\ ph i_t(0 , t) \ g e q \l amb da(D ^2\ph i(0, t)).$$ W e f irst sho w tha t $\o verl in e{u}$ itsel f sat is fies. Inserting $x= 0$ into a nd, w e s ee that we mus t h ave $$f'(t )+g'(t) \ge q\ l a mb da(-\frac{f(t)}{2t }\ m a th bb{Id}_n ) =-\f r ac {f ( t)}{2t}=\fr ac{1 } { 4t|\l og|t | |^ {\delta+ 1}}.$$ Wr it ing out t he lef than d s ide,w e ge t $$\f rac{1}{t |\log | t||^{\delta+1} } \left(-\frac{ \ de l t a+ 1 }{2} \fr ac{1}{|\log |t|| } +\de lta\ r ig ht) \geq \fra c{ 1 }{ 4 t|\log|t||^{\delta+ 1} },$$ w hichis equivalent to $$-\fr a c { \delta+1 }{2} \ fr a c{1}{|\log|t|| }+\de lta \leq \ f rac{1}{4 }.$$Since $\ delta$ wa s set to $ 1/4 $,weend u pwith $$-\frac { 5 }{8} \f rac{1}{ |\l og|t||} \l eq0,$ $ a nd so $\ove rline{u} $it se lf sa tisfi e s in $\O me ga$ .Let $\ph i $ be t he te st f un ct i onin Defi n it i o n \[ de f: visc osi ty defin itio n \]so that $\overli ne{ u }-\p hi $has a m inimum at $(0 ,t)$. Thisim pli es tha t $$ | frac{1}{4}+\delta\mathrm{e}^{\frac{|x|^2}{4t}}\right).
_ _\end{aligned}$$ Since $t<0$, we_have $$\mathrm{e}^{\frac{|x|^2}{4t}}<1.$$_Hence_$$\overline{u}_t(x, t)-\Delta_{\infty}^N\overline{u}(x,_t)_>\mathrm{e}^{\frac{-|x|^2}{4t}}\frac{1}{t|\log|t||^{\delta+1}}\left(\frac{-(\delta+1)}{2|\log|t||}-\frac{1}{4}+\delta\right).$$ Setting $\delta_=1/4$ ensures that_this is positive.
Then assume_$x=0$. We need_to_check that for every $\phi \in C^2(\Omega)$ so that $u-\phi$ has a minimum at_$(0,_t)$, we_have_$$\label{eq:viscdef}
_ \phi_t(0, t)_\geq \lambda(D^2\phi(0, t)).$$ We first_show that_$\overline{u}$ itself satisfies. Inserting $x=0$ into and, we_see_that we must_have $$f'(t)+g'(t) \geq \lambda(-\frac{f(t)}{2t}\mathbb{Id}_n) =-\frac{f(t)}{2t} = \frac{1}{4t|\log|t||^{\delta+1}}.$$ Writing out_the left hand side, we get_$$\frac{1}{t|\log|t||^{\delta+1}}\left(-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta\right) \geq _\frac{1}{4t|\log|t||^{\delta+1}},$$_which_is equivalent to $$-\frac{\delta+1}{2}\frac{1}{|\log|t||}+\delta_\leq \frac{1}{4}.$$ Since $\delta$ was set_to $1/4$, we end up with_$$-\frac{5}{8}\frac{1}{|\log|t||} \leq 0,$$ and so $\overline{u}$ itself_satisfies in $\Omega$. Let $\phi$ be_the test function in Definition_\[def:viscositydefinition\] so_that $\overline{u}-\phi$ has a minimum_at $(0, t)$._This implies_that $$ |
-e^{i\theta_j}}\log zdz,$$ where $C$ is a curve which runs cclw around the ball of radius $R$ around the origin, from angle $-\pi$ to $\pi$, then along the negative $x$-axis from $-R$ to $-r$, then clockwise around the ball of radius $r$ from angle $\pi$ to $-\pi$, and then back along the $x$-axis from $-r$ to $-R$.
The integral around the ball of radius $r$ is $$-\frac1{8\pi^2 i}\int_{\pi}^{-\pi}(-1)^k(1+O(Nr))(\log r+i\theta)id\theta
=\frac{(-1)^k\log r}{4\pi}(1+O(Nr)).$$ The difference between the value of $\log z$ above and below the $x$-axis is $2\pi i$, so that we can combine the two parts of the integral along the $x$-axis into the integral of $$-\frac1{4\pi}\int_{-R}^{-r}\frac1z\prod_{j=0}^{k-1}
\frac{z+e^{i\theta_j}}{z-e^{i\theta_j}}dz.$$ This integral we split into a part from $-R$ to $-\sqrt{N}$, from $-\sqrt{N}$ to $-1/\sqrt{N}$ and from $-1/\sqrt{N}$ to $-r$.
For small $t<0$ we have $$\prod_{j=0}^{k-1}\frac{t+e^{i\theta_j}}{t-e^{i\theta_j}}=
(-1)^ke^{2\sum e^{-i\theta_j}t+O(t^2)}=(-1)^ke^{2(\bar v_1-\bar v_0)t+O(t^2)}$$ so that the integral near the origin is $$-\frac{(-1)^k}{4\pi}
\int_{-1/\sqrt{N}}^{-r}\frac{e^{\bar\alpha t+O(t^2)}}t dt$$ where $\alpha=2(v_1-v_0),$ $$\begin{aligned}
&=&-\frac{(-1)^k}{4\pi}(1+O(\frac1N))
\int_{-\ | -e^{i\theta_j}}\log zdz,$$ where $ C$ is a curve which runs cclw around the ball of radius $ R$ around the beginning, from slant $ -\pi$ to $ \pi$, then along the negative $ x$-axis from $ -R$ to $ -r$, then clockwise around the ball of radius $ r$ from slant $ \pi$ to $ -\pi$, and then back along the $ x$-axis from $ -r$ to $ -R$.
The integral around the testis of radius $ r$ is $ $ -\frac1{8\pi^2 i}\int_{\pi}^{-\pi}(-1)^k(1+O(Nr))(\log r+i\theta)id\theta
= \frac{(-1)^k\log r}{4\pi}(1+O(Nr)).$$ The difference between the value of $ \log z$ above and below the $ x$-axis is $ 2\pi i$, so that we can combine the two region of the integral along the $ x$-axis into the integral of $ $ -\frac1{4\pi}\int_{-R}^{-r}\frac1z\prod_{j=0}^{k-1 }
\frac{z+e^{i\theta_j}}{z - e^{i\theta_j}}dz.$$ This integral we divide into a part from $ -R$ to $ -\sqrt{N}$, from $ -\sqrt{N}$ to $ -1/\sqrt{N}$ and from $ -1/\sqrt{N}$ to $ -r$.
For small $ t<0 $ we get $ $ \prod_{j=0}^{k-1}\frac{t+e^{i\theta_j}}{t - e^{i\theta_j}}=
(-1)^ke^{2\sum e^{-i\theta_j}t+O(t^2)}=(-1)^ke^{2(\bar v_1-\bar v_0)t+O(t^2)}$$ so that the integral near the origin is $ $ -\frac{(-1)^k}{4\pi }
\int_{-1/\sqrt{N}}^{-r}\frac{e^{\bar\alpha t+O(t^2)}}t dt$$ where $ \alpha=2(v_1 - v_0),$ $ $ \begin{aligned }
& = & -\frac{(-1)^k}{4\pi}(1+O(\frac1N) )
\int_{-\ | -e^{i\tjeta_j}}\log zdz,$$ where $C$ is x curve which ronw cclw arouns the baul of radius $R$ around the ormgin, from angle $-\pi$ to $\pi$, then auong the jegative $x$-axms from $-R$ to $-r$, tisn clockwise zvound vhe ball of radlus $r$ from dngle $\pi$ to $-\pi$, avd then back along the $x$-axis from $-r$ tj $-R$.
The onhegral around jhe bsjl or radius $r$ is $$-\frac1{8\pi^2 i}\int_{\pi}^{-\pi}(-1)^k(1+O(Nr))(\lkg r+i\thtta)id\theta
=\frac{(-1)^k\log r}{4\pi}(1+O(Nr)).$$ The difference betaeen the value of $\log x$ above and beljq the $x$-axis ks $2\pi i$, so that we can combine the two parts of the ivtegrcl along thg $s$-adhs into the intedral of $$-\frac1{4\ip}\int_{-R}^{-r}\fsac1z\proc_{j=0}^{k-1}
\frac{z+e^{i\thets_j}}{z-x^{i\thwta_j}}dz.$$ This integral xe split into a part from $-R$ tm $-\aqrt{N}$, from $-\sqrt{N}$ ro $-1/\sqrt{T}$ ang frun $-1/\sdrt{H}$ vo $-d$.
For slalm $t<0$ we havs $$\prod_{j=0}^{k-1}\frax{t+e^{i\theta_j}}{t-e^{i\theta_j}}=
(-1)^le^{2\flk e^{-i\theta_j}t+O(f^2)}=(-1)^ke^{2(\bar v_1-\far v_0)t+O(t^2)}$$ so that the integral near the mrifin is $$-\frac{(-1)^k}{4\pi}
\int_{-1/\sqrt{N}}^{-r}\drac{e^{\bar\alpha t+O(t^2)}}t dt$$ where $\al[ha=2(v_1-v_0),$ $$\begin{aligned}
&=&-\frac{(-1)^k}{4\pi}(1+O(\frac1N))
\int_{-\ | -e^{i\theta_j}}\log zdz,$$ where $C$ is a curve cclw the ball radius $R$ around to then along the $x$-axis from $-R$ $-r$, then clockwise around the ball radius $r$ from angle $\pi$ to $-\pi$, and then back along the $x$-axis $-r$ to $-R$. The integral around the ball of radius $r$ is $$-\frac1{8\pi^2 r+i\theta)id\theta r}{4\pi}(1+O(Nr)).$$ difference the value of $\log z$ above and below the $x$-axis is $2\pi i$, so that we combine the two parts of the integral along $x$-axis into the integral $$-\frac1{4\pi}\int_{-R}^{-r}\frac1z\prod_{j=0}^{k-1} \frac{z+e^{i\theta_j}}{z-e^{i\theta_j}}dz.$$ This integral we into part from to from to $-1/\sqrt{N}$ and $-1/\sqrt{N}$ to $-r$. For small $t<0$ we have $$\prod_{j=0}^{k-1}\frac{t+e^{i\theta_j}}{t-e^{i\theta_j}}= (-1)^ke^{2\sum e^{-i\theta_j}t+O(t^2)}=(-1)^ke^{2(\bar v_1-\bar v_0)t+O(t^2)}$$ so that the integral the origin \int_{-1/\sqrt{N}}^{-r}\frac{e^{\bar\alpha t+O(t^2)}}t where $$\begin{aligned} \int_{-\ | -e^{i\theta_j}}\log zdz,$$ where $C$ is a cuRve which ruNs cclW arOunD tHe baLl of Radius $R$ around tHE oriGin, from angle $-\pi$ to $\pi$, then Along ThE NegaTIvE $x$-axiS from $-R$ tO $-R$, tHEN clOcKwIse ArOUnD the bAll Of radiuS $r$ from anglE $\pi$ To $-\Pi$, and then bacK AlOng the $x$-axiS frOm $-r$ to $-R$.
The intEgrAl arouNd The BAll of RadIus $r$ iS $$-\frac1{8\pI^2 I}\int_{\pi}^{-\Pi}(-1)^k(1+O(Nr))(\log R+i\THeta)id\THeta
=\fraC{(-1)^K\LoG r}{4\pi}(1+o(Nr)).$$ The difference bETwEEn the value of $\loG z$ abovE aND bELOw tHe $x$-Axis is $2\pi i$, sO tHat we CAn combiNE tHE TWo pARts of the integRal along the $X$-AxiS into tHe IntEGral of $$-\Frac1{4\pI}\iNT_{-R}^{-r}\Frac1z\prod_{j=0}^{k-1}
\Frac{Z+e^{i\theta_j}}{Z-e^{i\theTA_j}}dz.$$ ThiS IntegraL we splIt iNto A parT FrOm $-r$ to $-\SqRT{N}$, fROm $-\SqrT{n}$ to $-1/\Sqrt{N}$ and FrOm $-1/\Sqrt{N}$ To $-r$.
FOR SMAll $t<0$ We hAve $$\pRod_{j=0}^{k-1}\Frac{t+e^{i\theta_j}}{T-e^{i\ThetA_J}}=
(-1)^ke^{2\Sum e^{-i\Theta_J}t+O(t^2)}=(-1)^Ke^{2(\Bar v_1-\bAr v_0)t+O(t^2)}$$ So thaT tHe integral near tHe orIgin is $$-\fraC{(-1)^k}{4\pI}
\iNt_{-1/\sQrT{N}}^{-r}\frAC{e^{\bar\aLphA t+O(T^2)}}t dt$$ wheRe $\alpha=2(V_1-V_0),$ $$\beGiN{ALIgNed}
&=&-\frac{(-1)^k}{4\pi}(1+O(\frac1N))
\inT_{-\ | -e^{i\theta_j}}\log zdz,$$ where $C$ is a cu rve w hich run s cclw aroundt he b all of radius $R$ arou nd th eo rigi n ,fromangle $ - \p i $ to $ \p i$, t h en alon g t he nega tive $x$-a xis f rom $-R$ to$ -r $, then cl ock wise aroundthe ballof ra d ius $ r$fromangle$ \pi$ t o $-\pi$, a n d then back al o n gthe$x$-axis from $-r $ t o $-R$.
The in tegral a r ou n d th e b all of rad iu s $r$ is $$-\ f ra c 1 { 8\p i ^2 i}\int_{\p i}^{-\pi}(- 1 )^k (1+O(N r) )(\ l og r+i \thet a) i d\t heta
=\frac {(-1 )^k\log r }{4\pi } (1+O(Nr ) ).$$ Th e diff ere nce bet w ee nthe v a lue of $\ l ogz$ above a nd belo w th e $ x $-ax isis $ 2\pii$, so that w e c an c o mbi ne th e two par ts of t he int egral a long the $x$-ax is i nto the i nte gr alof $$-\ f rac1{4 \pi }\i nt_{-R} ^{-r}\f r ac1 z\ p r o d_ {j=0}^{k-1}
\frac{ z+ e ^ {i \theta_j }}{z-e ^ {i \t h eta_j}}d z. $$This i ntegr al w e s plit int o a pa r tfr om $-R$ t o $-\s qr t{N }$, from $-\s qrt{N} $ to $-1 /\sqr t {N}$ and from$ -1/\sqrt{N}$t o$ - r$ .
Fo r s mall $t<0$we h a ve $ $\pr o d_ {j= 0 }^{k- 1}\fr ac { t+ e ^{i\theta_j}}{t-e^{ i\ theta_ j}}=(-1)^ke^{2\su m e^{-i\th e t a _j}t+O(t ^2)} = (- 1 )^ke^{2(\bar v _1-\b ar v_0)t+O ( t^2)}$$so th at the i ntegral n e a r the or igi n i s $ $-\ f r ac {(-1)^k}{4\pi } \int _{ -1/\sqr t{N }}^{-r} \fr ac{ e^{ \ba r\ alpha t+O (t^2)}}t d t$ $wh ere $\al p ha=2(v_1 -v _0) ,$ $$ \begi n {align ed}
& =&-\ fr ac { (-1 )^k}{4\ p i} ( 1 +O(\ fr ac 1N))
\i nt _{-\ | -e^{i\theta_j}}\log zdz,$$_where $C$_is a curve which_runs cclw_around_the ball_of_radius $R$ around_the origin, from_angle $-\pi$ to $\pi$,_then along the_negative_$x$-axis from $-R$ to $-r$, then clockwise around the ball of radius $r$ from_angle_$\pi$ to_$-\pi$,_and_then back along the $x$-axis_from $-r$ to $-R$.
The integral_around the_ball of radius $r$ is $$-\frac1{8\pi^2 i}\int_{\pi}^{-\pi}(-1)^k(1+O(Nr))(\log r+i\theta)id\theta
=\frac{(-1)^k\log_r}{4\pi}(1+O(Nr)).$$_The difference between_the value of $\log z$ above and below the_$x$-axis is $2\pi i$, so that_we can combine_the_two_parts of the integral_along the $x$-axis into the integral_of $$-\frac1{4\pi}\int_{-R}^{-r}\frac1z\prod_{j=0}^{k-1}
\frac{z+e^{i\theta_j}}{z-e^{i\theta_j}}dz.$$ This integral we split_into a part from $-R$ to $-\sqrt{N}$,_from $-\sqrt{N}$ to $-1/\sqrt{N}$ and from_$-1/\sqrt{N}$ to $-r$.
For small $t<0$_we have_$$\prod_{j=0}^{k-1}\frac{t+e^{i\theta_j}}{t-e^{i\theta_j}}=
(-1)^ke^{2\sum e^{-i\theta_j}t+O(t^2)}=(-1)^ke^{2(\bar v_1-\bar v_0)t+O(t^2)}$$ so_that the integral_near the_origin is $$-\frac{(-1)^k}{4\pi}
\int_{-1/\sqrt{N}}^{-r}\frac{e^{\bar\alpha_t+O(t^2)}}t dt$$ where $\alpha=2(v_1-v_0),$ $$\begin{aligned}
&=&-\frac{(-1)^k}{4\pi}(1+O(\frac1N))
\int_{-\ |
emptyset = +\infty$. We then define the Markov chain $(Z^{u,\varepsilon}_n)_{n \in {{\mathbb N}}}$ as $$Z^{u,\varepsilon}_n:=U^{u,\varepsilon}(\eta_n,\cdot)$$ for each $n \in {{\mathbb N}}_0$ and set $\vartheta^u_\varepsilon := \min \{ n \in {{\mathbb N}}: Z^{u,\varepsilon}_n \in {\partial}G\}$. Just as in the previous section, for our purposes we will not need to worry about the possibility of $Z_n^\varepsilon$ not being well defined. Also, the constants $r_1$ and $r_2$ will be later taken conveniently small It is not hard to see that for any $u \in {\partial}B_{r_1}$ the strong Markov property yields $$\label{cociente}
P_u \left( U^{\varepsilon}(\tau_\varepsilon(\partial G),\cdot) \notin \partial^{\pm z} \right) \leq \sup_{v \in {\partial}B_{r_1}} \frac{ P_v ( Z_1^\varepsilon \in {\partial}G - {\partial}^{\pm z} )}{P_v ( Z_1^\varepsilon \in {\partial}G)}$$ from which we conclude that in order to show (ii) it will suffice to give a lower bound for the denominator and an upper bound for the numerator such that the quotient of these bounds goes to zero with $\varepsilon$. The following lemma, whose proof can be found in [@B1], provides the desired lower bound.
\[lemab0\] Let us suppose that $r_1$ is sufficiently small so as to guarantee that for any $u \in {\partial}B_{r_1}$ the deterministic orbit $\{ U^u(t,\cdot) : t \geq 0 \}$ does not intersect ${\partial}B_{\frac{r_2}{2}}$. Then for all $\varepsilon > 0$ sufficiently small $$\inf_{u \in {\partial}B_{r_1}} P_u( Z_1^\varepsilon \in {\partial}G) \geq e^{- \frac{\Delta + k r_2}{\varepsilon^2} },$$ where $k > 0$ is a constant which does not depend on the choice of $r_1$ and $r_2$.
The upper bound on the numerator in is more involved | emptyset = + \infty$. We then define the Markov chain $ (Z^{u,\varepsilon}_n)_{n \in { { \mathbb N}}}$ as $ $ Z^{u,\varepsilon}_n:=U^{u,\varepsilon}(\eta_n,\cdot)$$ for each $ n \in { { \mathbb N}}_0 $ and set $ \vartheta^u_\varepsilon: = \min \ { n \in { { \mathbb N } }: Z^{u,\varepsilon}_n \in { \partial}G\}$. barely as in the former section, for our purposes we will not necessitate to concern about the possibility of $ Z_n^\varepsilon$ not being well defined. Also, the constant $ r_1 $ and $ r_2 $ will be later taken handily humble It is not hard to experience that for any $ u \in { \partial}B_{r_1}$ the strong Markov place yields $ $ \label{cociente }
P_u \left (U^{\varepsilon}(\tau_\varepsilon(\partial G),\cdot) \notin \partial^{\pm z } \right) \leq \sup_{v \in { \partial}B_{r_1 } } \frac { P_v (Z_1^\varepsilon \in { \partial}G - { \partial}^{\pm z }) } { P_v (Z_1^\varepsilon \in { \partial}G)}$$ from which we conclude that in order to show (ii) it will suffice to give a lower oblige for the denominator and an upper bound for the numerator such that the quotient of these bounds goes to zero with $ \varepsilon$. The following lemma, whose validation can be found in [ @B1 ], provides the desire lower bound.
\[lemab0\ ] permit us suppose that $ r_1 $ is sufficiently small so as to guarantee that for any $ u \in { \partial}B_{r_1}$ the deterministic orbit $ \ { U^u(t,\cdot): t \geq 0 \}$ does not intersect $ { \partial}B_{\frac{r_2}{2}}$. Then for all $ \varepsilon > 0 $ sufficiently belittled $ $ \inf_{u \in { \partial}B_{r_1 } } P_u (Z_1^\varepsilon \in { \partial}G) \geq e^{- \frac{\Delta + k r_2}{\varepsilon^2 } }, $ $ where $ k > 0 $ is a constant which does not depend on the choice of $ r_1 $ and $ r_2$.
The upper bound on the numerator in is more involved | emphyset = +\infty$. We then deflne the Markov ckqin $(Z^{u,\tarepsimon}_n)_{n \in {{\mathbb N}}}$ as $$Z^{u,\varepsilon}_n:=U^{u,\tarepsilob}(\eta_n,\cdot)$$ for each $n \iv {{\mathbb J}}_0$ and ser $\vactheta^u_\varepsiloi := \min \{ k \in {{\jwthby I}}: Z^{u,\varepsilon}_n \in {\partian}G\}$. Just as in dhd 'revious section, for our purposes we will npt need to worry abolt the iowsibility of $Z_n^\varepsilon$ not being xell defined. Alxo, the constants $r_1$ and $r_2$ wlll he later taken congeniently snall Ut is not hafd to see uhct for any $o \in {\partial}B_{r_1}$ the strong Markov propzrty yields $$\lqbep{wociente}
P_u \oeft( L^{\varepsilon}(\tam_\narepsinon(\partoal G),\cdot) \notik \parvial^{\pm z} \right) \leq \sup_{v \ii {\partial}B_{r_1}} \frac{ P_v ( Z_1^\varepsinou \in {\partial}G - {\partiao}^{\pn z} )}{P_e ( Z_1^\eareowilun \jn {\pzrtial}H)}$$ fcom which ws conclude rhat in order to shpw (po) it will surfice eo give a lower bound for the denominator ans an upper bound for thw numerator such that the quotyent of these bounds goes to zero with $\varepsilon$. The houloclkg ldnmw, whose proof can be found in [@B1], provides the qssorvd lower bound.
\[lemcb0\] Let us supposr hhsj $r_1$ is sufficigntly smalm so as to guarantfe that for qny $u \in {\kartisl}B_{r_1}$ the deterministic orbir $\{ U^u(t,\cdot) : n \gew 0 \}$ does not interdect ${\partiau}B_{\frsc{r_2}{2}}$. Tnen for all $\varepsilon > 0$ surficiently dmall $$\inf_{h \in {\partial}B_{r_1}} P_u( Z_1^\vsrapsilon \in {\partial}G) \geq e^{- \srac{\Delta + k x_2}{\varepsiuon^2} },$$ where $k > 0$ is a fonstant which does not fepenb on dhe choice of $r_1$ and $r_2$.
The upper bound on tix numerator im hs kore invjlved | emptyset = +\infty$. We then define the $(Z^{u,\varepsilon}_n)_{n {{\mathbb N}}}$ $$Z^{u,\varepsilon}_n:=U^{u,\varepsilon}(\eta_n,\cdot)$$ for each set := \min \{ \in {{\mathbb N}}: \in {\partial}G\}$. Just as in the section, for our purposes we will not need to worry about the possibility $Z_n^\varepsilon$ not being well defined. Also, the constants $r_1$ and $r_2$ will be taken small is hard to see that for any $u \in {\partial}B_{r_1}$ the strong Markov property yields $$\label{cociente} P_u U^{\varepsilon}(\tau_\varepsilon(\partial G),\cdot) \notin \partial^{\pm z} \right) \leq \sup_{v {\partial}B_{r_1}} \frac{ P_v ( \in {\partial}G - {\partial}^{\pm z} ( \in {\partial}G)}$$ which conclude in order to (ii) it will suffice to give a lower bound for the denominator and an upper bound for numerator such quotient of bounds to with $\varepsilon$. The whose proof can be found in desired lower bound. \[lemab0\] Let us suppose that is sufficiently so as to guarantee that for $u \in {\partial}B_{r_1}$ the deterministic orbit $\{ U^u(t,\cdot) t \geq 0 \}$ does not intersect ${\partial}B_{\frac{r_2}{2}}$. Then for all $\varepsilon > 0$ sufficiently \in {\partial}B_{r_1}} P_u( Z_1^\varepsilon {\partial}G) \geq e^{- + r_2}{\varepsilon^2} where > 0$ a constant which does not depend on the choice of $r_1$ $r_2$. The upper bound on the numerator in is more | emptyset = +\infty$. We then define The Markov cHain $(Z^{U,\vaRepSiLon}_n)_{N \in {{\mAthbb N}}}$ as $$Z^{u,\varePSiloN}_n:=U^{u,\varepsilon}(\eta_n,\cdot)$$ For eaCh $N \In {{\maTHbB N}}_0$ and Set $\vartHEtA^U_\VarEpSiLon := \MiN \{ N \iN {{\mathBb N}}: z^{u,\varepSilon}_n \in {\paRtiAl}g\}$. Just as in the PReVious sectiOn, fOr our purposeS we Will noT nEed TO worrY abOut thE possiBIlity oF $Z_n^\varepsIlON$ not beINg well dEFInEd. AlSo, the constants $r_1$ anD $R_2$ wILl be later taken ConvenIeNTlY SMalL It Is not hard tO sEe thaT For any $u \IN {\pARTIal}b_{R_1}$ the strong MarKov property YIelDs $$\labeL{cOciENte}
P_u \lEft( U^{\vArEPsiLon}(\tau_\varepSiloN(\partial G),\Cdot) \noTIn \partiAL^{\pm z} \rigHt) \leq \sUp_{v \In {\pArtiAL}B_{R_1}} \fRac{ p_v ( z_1^\VarEPsIloN \In {\pArtial}G - {\pArTiAl}^{\pm z} )}{p_v ( Z_1^\vAREPSiloN \in {\PartIal}G)}$$ fRom which we conCluDe thAT in Order To shoW (ii) iT wIll suFfice tO give A lOwer bound for the DenoMinator anD an UpPer BoUnd foR The numEraTor Such thaT the quoTIenT oF THEsE bounds goes to zero wItH $\VArEpsilon$. THe follOWiNg LEmma, whosE pRooF can BE Found In [@B1], pROvIdes the dEsired LOwEr Bound.
\[leMaB0\] Let us SuPpoSe tHat $r_1$ iS SuffIcientLy small sO as to GUarantee that foR Any $u \in {\partial}b_{R_1}$ tHE DeTErmiNisTic orbit $\{ U^u(t,\Cdot) : T \Geq 0 \}$ dOes nOT iNteRSect ${\pArtiaL}B_{\FRaC{R_2}{2}}$. Then for all $\varepsilOn > 0$ SufficIentlY small $$\inf_{u \in {\pArtial}B_{r_1}} P_u( z_1^\VARepsilon \In {\paRTiAL}G) \geq e^{- \frac{\DeltA + k r_2}{\vaRepsilon^2} },$$ whERe $k > 0$ is a coNstanT which doEs not depeND On the choIce Of $r_1$ And $R_2$.
ThE UPpEr bound on the nUMEratOr In is morE inVolved | emptyset = +\infty$. We th en definethe M ark ovch ain$(Z^ {u,\varepsilon } _n)_ {n \in {{\mathbb N}}}$ as $ $Z ^ {u,\ v ar epsil on}_n:= U ^{ u , \va re ps ilo n} ( \e ta_n, \cd ot)$$ f or each $n \i n{{\mathbb N} } _0 $ and set$\v artheta^u_\v are psilon : = \ m in \{ n\in { {\math b b N}}: Z^{u,\va re p silon} _ n \in { \ p ar tial }G\}$. Just as in th e previous sect ion, f or ou r pur pos es we will n ot ne e d to wo r ry a b out the possibili ty of $Z_n^ \ var epsilo n$ no t being well d e fin ed. Also, t he c onstants$r_1$a nd $r_2 $ will b e late r t ake n co n ve ni ent ly sma l lIti s n ot hardto s ee th at f o r a ny $ u \ in { \part ial}B_{r_1}$the str o ngMarko v pro pert yyield s $$\l abel{ co ciente}
P_u \le ft(U^{\varep sil on }(\ ta u_\va r epsilo n(\ par tial G) ,\cdot) \no ti n \ pa rtial^{\pm z} \rig ht ) \l eq \sup_ {v \in {\ pa r tial}B_{ r_ 1}} \fr a c { P_v ( Z _ 1^ \varepsi lon \i n { \p artial} G- {\pa rt ial }^{ \pm z } )}{ P_v (Z_1^\var epsil o n \in {\partia l }G)}$$ from w h ic h we conc lud e that in o rder to s how( ii ) i t will suff ic e t o give a lower bound f or the deno minator and a n upper bo u n d for the num e ra t or such that t he qu otient oft hese bou nds g oes to z ero with$ \ varepsil on$ . T hefol l o wi ng lemma, who s e pro of can be fo und in[@B 1], pr ovi de s the des ired low er b ou nd .
\[lem a b0\] Let u s s up pos e tha t $r_1$ is s uffi ci en t lysmall s o a s to g ua ra ntee th at forany$ u \ in {\pa rtial}B_{ r_1 } $ th ede termini stic orbit $\ {U^u(t,\cdo t) :t \geq 0 \}$ doe s not intersect ${\part i al}B_{\ fra c{r_2 }{2} }$. Thenfor all $ \va r epsilo n > 0$ suff ic ien t l y sma l l $ $\i nf _{u \in {\ p a rti al}B_ {r _1}} P_u( Z _1^\varepsilon \in {\p artial}G) \ge q e ^{-\ f ra c{\ D el t a + k r_2 } { \varepsilon^2}},$$ where $ k > 0$ is a c o nst an t which does n ot de p end onthe choic e of $r_1 $and$ r _2$ .
The upp er bound on the n u merat o rin is mo re inv ol ved | emptyset =_+\infty$. We_then define the Markov_chain $(Z^{u,\varepsilon}_n)_{n_\in_{{\mathbb N}}}$_as_$$Z^{u,\varepsilon}_n:=U^{u,\varepsilon}(\eta_n,\cdot)$$ for each_$n \in {{\mathbb_N}}_0$ and set $\vartheta^u_\varepsilon_:= \min \{_n_\in {{\mathbb N}}: Z^{u,\varepsilon}_n \in {\partial}G\}$. Just as in the previous section, for our_purposes_we will_not_need_to worry about the possibility_of $Z_n^\varepsilon$ not being well_defined. Also,_the constants $r_1$ and $r_2$ will be later_taken_conveniently small It_is not hard to see that for any $u_\in {\partial}B_{r_1}$ the strong Markov property_yields $$\label{cociente}
P_u \left(_U^{\varepsilon}(\tau_\varepsilon(\partial_G),\cdot)_\notin \partial^{\pm z} \right)_\leq \sup_{v \in {\partial}B_{r_1}} \frac{ P_v_( Z_1^\varepsilon \in {\partial}G - {\partial}^{\pm_z} )}{P_v ( Z_1^\varepsilon \in {\partial}G)}$$ from_which we conclude that in order_to show (ii) it will_suffice to_give a lower bound for_the denominator and_an upper_bound for the_numerator such that the quotient of_these bounds goes_to zero with $\varepsilon$. The following_lemma,_whose proof can_be_found_in [@B1],_provides the desired_lower_bound.
\[lemab0\] Let_us_suppose that $r_1$ is sufficiently small_so_as to guarantee that for any $u_\in {\partial}B_{r_1}$ the deterministic_orbit_$\{ U^u(t,\cdot) : t_\geq 0 \}$ does not_intersect ${\partial}B_{\frac{r_2}{2}}$. Then for all $\varepsilon_> 0$_sufficiently small_$$\inf_{u \in {\partial}B_{r_1}} P_u( Z_1^\varepsilon \in {\partial}G) \geq e^{- \frac{\Delta +_k r_2}{\varepsilon^2} },$$ where $k >_0$ is a constant_which does_not_depend on the_choice_of $r_1$_and $r_2$.
The upper bound on the numerator_in is_more involved |
\
ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russia\
E-mail:
- |
A.I.Veselov\
ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russia\
E-mail:
title: Upper bound on the cutoff in the Standard Model
---
Introduction
============
According to the conventional point of view the upper bound $\Lambda$ on the cutoff in the Electroweak theory (without fermions) depends on the Higgs mass. It is decreased when the Higgs mass is increased. And at the Higgs mass around $1$ Tev $\Lambda$ becomes of the order of $M_H$. At the same time for $M_H \sim
200$ Gev the value of $\Lambda$ can be made almost infinite[^1]. This conclusion is made basing on the perturbation expansion around trivial vacuum. In our presentation we demonstrate that the vacuum of the lattice Weinberg - Salam model is rather complicated, which means that the application of the perturbation expansion around trivial vacuum may be limited.
Namely, we investigate the behavior of the topological defects composed of the lattice gauge fields that are to be identified with quantum Nambu monopoles [@Nambu; @BVZ; @Chernodub_Nambu]. We show that their lattice density increases along the lines of constant physics when the ultraviolet cutoff in increased. At sufficiently large values of the cutoff these objects begin to dominate.
Moving further along the line of constant physics we reach the point on the phase diagram where the monopole worldlines begin to percolate. This point roughly coincides with the position of the transition between the physical Higgs phase and the unphysical symmetric phase of the lattice model. At infinite bare scalar self coupling $\lambda$ the transition is a crossover and the ultraviolet cutoff achieves its maximal value around $1.4$ Tev at the transition point. At smaller bare values of $\lambda$ correspondent to small Higgs masses the phase transition becomes stronger. Still we do not know the order of the phase transition at small values of $\lambda$. We have estimated the maximal value of the cutoff in the vicinity of the transition point at $\lambda = 0.009$. The obtained value of the cutoff appears to be around $1.4$ Tev.
The lattice model under investigation
=====================================
The lattice Weinberg - Salam | \
ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russia\
E - mail:
- |
A.I.Veselov\
ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russia\
E - mail:
claim: Upper tie on the cutoff in the Standard Model
---
Introduction
= = = = = = = = = = = =
According to the ceremonious point of view the upper bound $ \Lambda$ on the shortcut in the Electroweak hypothesis (without fermions) depends on the Higgs mass. It is decreased when the Higgs mass is increase. And at the Higgs mass around $ 1 $ Tev $ \Lambda$ becomes of the decree of $ M_H$. At the same time for $ M_H \sim
200 $ Gev the value of $ \Lambda$ can be made almost infinite[^1 ]. This conclusion is made establish on the perturbation expansion around trivial vacuum. In our display we demonstrate that the vacuum of the wicket Weinberg - Salam model is quite complicated, which means that the lotion of the perturbation expansion around trivial vacuum may be limited.
Namely, we investigate the demeanor of the topological defects composed of the lattice gauge fields that are to be identified with quantum Nambu monopoles [ @Nambu; @BVZ; @Chernodub_Nambu ]. We show that their lattice density increases along the lines of constant physics when the ultraviolet cutoff in increased. At sufficiently large values of the shortcut these object begin to overshadow.
move further along the line of constant purgative we reach the distributor point on the phase diagram where the monopole worldlines begin to percolate. This distributor point roughly coincides with the position of the transition between the physical Higgs phase and the unphysical symmetric phase of the wicket model. At infinite bare scalar self coupling $ \lambda$ the transition is a crossover and the ultraviolet cutoff achieve its maximal value around $ 1.4 $ Tev at the passage point. At smaller bare values of $ \lambda$ correspondent to belittled Higgs masses the phase transition becomes potent. even we do not know the order of the phase conversion at small value of $ \lambda$. We have estimated the maximal value of the cutoff in the vicinity of the passage point at $ \lambda = 0.009$. The obtain value of the cutoff appear to be around $ 1.4 $ Tev.
The lattice mannequin under investigation
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
The lattice Weinberg - Salam | \
ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russic\
E-meil:
- |
A.I.Veseluv\
ITEP, B.Cheremushkinskaye 25, Miscow, 117259, Russia\
E-mail:
title: Upper bolnd on thw cuuoff in the Standedd Model
---
Introsmctiou
============
Arcording to the conventiotal point of vhed che upper bound $\Lambda$ on the cutoff in the Epectroweak thejry (eythohn nermions) depends on the Higgs maas. It iv decreased wnen the Higgs mass is incrfasef. And at the Higgs mass aroune $1$ Tqc $\Lambda$ becumes of tht mrder of $M_G$. At the same time for $M_H \sim
200$ Gdv thz value of $\Oanbdw$ can be madx almoft infinite[^1]. Bnis cotclusiom is made basikg on thw perturbation expansmon around trivial vwcuum. In mux presentation we deminwtratg thad thd vazuuj pf the lwttmce Weinberf - Salam moeel is rather complocwnrd, which meahs thae ehe application of the perturbation expdnsjon around trivial vacuym may be limited.
Namepy, we invqstigate the behavior of the topological defects wompoadd if gye lattice gauge fields that are to be identifiqs einh quantum Nambu ionopoles [@Nsmhu; @FVZ; @Chernodub_Vambu]. Cs ahow that their lahtice dgnsity increasef alpng the lines of constant pyysics when nhe yltraviolet cutoff in increaszd. At xuffiviently large values of the cutoff thede objecta begin to dominage.
Mpvhng further along the line of constent pkysics wd resch thq point on the ivase diagram where the lotopole worpdlines begin to percolate. This point roughly cminbides witk the iosition of the transition bejween the physkcal Higgs phase end the unphrsical symmetslc phase of vhe lattise midel. At infkvite bare scalsr self coupling $\lanbda$ the transitiok is x crossover and cke ultraviolet cuyofw ashpevxs itf maximal valge afouvc $1.4$ Tex at tht tvanritipn point. At smaller tare values of $\lambda$ vovrespondebt to smwll Higgs masxes the phase tranditioi becooes syrogger. Still we do not know the krder of hhe phase transytiok at small valbes of $\lambda$. We have estimated the maxinal value of the cytoff in the vicinijy pf the tranxitiog point ad $\lambda = 0.009$. The obtaibed value of the gutoff appears to be adound $1.4$ Tev.
Tje lattice model under investigation
=====================================
The lattice Weinberg - Salam | \ ITEP, B.Cheremushkinskaya 25, Moscow, 117259, Russia\ | ITEP, B.Cheremushkinskaya Moscow, 117259, Russia\ the in the Standard --- Introduction ============ to the conventional point of view upper bound $\Lambda$ on the cutoff in the Electroweak theory (without fermions) depends the Higgs mass. It is decreased when the Higgs mass is increased. And the mass $1$ $\Lambda$ becomes of the order of $M_H$. At the same time for $M_H \sim 200$ Gev value of $\Lambda$ can be made almost infinite[^1]. conclusion is made basing the perturbation expansion around trivial In presentation we that vacuum the lattice Weinberg Salam model is rather complicated, which means that the application of the perturbation expansion around trivial vacuum be limited. investigate the of topological composed of the fields that are to be identified monopoles [@Nambu; @BVZ; @Chernodub_Nambu]. We show that their density increases the lines of constant physics when ultraviolet cutoff in increased. At sufficiently large values the cutoff these objects begin to dominate. Moving further along the line of constant physics the point on the diagram where the worldlines to This roughly coincides the position of the transition between the physical Higgs phase and unphysical symmetric phase of the lattice model. At infinite bare coupling the transition is crossover and the ultraviolet achieves maximal value around $1.4$ the point. values $\lambda$ to small Higgs masses phase transition becomes stronger. Still do not know the at small values of $\lambda$. We have estimated maximal value of the cutoff in the of the transition point at $\lambda = 0.009$. The obtained value of cutoff appears around $1.4$ Tev. The lattice model under investigation The lattice Weinberg - | \
ITEP, B.Cheremushkinskaya 25, MosCow, 117259, Russia\
E-Mail:
- |
A.i.VeSelOv\
iTEP, b.CheRemushkinskaya 25, mOscoW, 117259, Russia\
E-mail:
title: Upper Bound On THe cuTOfF in thE StandaRD MODEl
---
INtRoDucTiON
============
ACcordIng To the coNventional PoiNt Of view the uppER bOund $\Lambda$ On tHe cutoff in thE ElEctrowEaK thEOry (wiThoUt ferMions) dEPends oN the Higgs MaSS. It is dECreased WHEn The HIggs mass is increasED. AND at the Higgs masS arounD $1$ TEV $\LAMBda$ BecOmes of the oRdEr of $M_h$. at the saME tIME For $m_h \sim
200$ Gev the valUe of $\Lambda$ cAN be Made alMoSt iNFinite[^1]. this cOnCLusIon is made baSing On the pertUrbatiON expansIOn arounD triviAl vAcuUm. In OUr PrEseNtATioN We DemONstRate that ThE vAcuum Of thE LATTice weiNberG - SalaM model is ratheR coMpliCAteD, whicH meanS thaT tHe appLicatiOn of tHe Perturbation expAnsiOn around tRivIaL vaCuUm may BE limitEd.
NAmeLy, we invEstigatE The BeHAVIoR of the topological dEfECTs Composed Of the lATtIcE Gauge fieLdS thAt arE TO be idEntiFIeD with quaNtum NaMBu MoNopoles [@naMbu; @BVZ; @chErnOduB_NambU]. we shOw that Their latTice dENsity increases ALong the lines oF CoNSTaNT phySicS when the ultRaviOLet cUtofF In IncREased. at sufFiCIeNTly large values of the CuToff thEse obJects begin to dOminate.
MovING Further aLong THe LIne of constant pHysicS we reach thE Point on tHe phaSe diagraM where the MONopole woRldLinEs bEgiN TO pErcolate. This pOINt roUgHly coinCidEs with tHe pOsiTioN of ThE transitiOn betweeN tHe PhYsIcaL HiggS Phase and ThE unPhYsiCal syMMetric Phase Of thE lAtTIce Model. At INfINIte bArE sCalaR seLf CouplIng $\lAMbdA$ the traNsition is A crOSsovEr AnD the ultRaviolet cutofF aChieves its MaXimAl valuE ARound $1.4$ Tev At the transition point. At sMAller baRe vAlues Of $\laMbda$ correSpoNdent tO smALl HiggS masseS the pHaSe tRANsitiON BeComEs Stronger. StILL we Do not KnOw thE order oF the phase transitioN At sMall values of $\lAmbDa$. We HAVe EstIMaTEd tHe MAxiMAL value of the cutoFf in the vicInITy Of the transITioN pOint at $\lAmbda = 0.009$. ThE obtaINed valuE of the cutOff appearS tO be aROUnd $1.4$ tev.
The lattIce model Under inveSTigatIOn
=====================================
the laTtiCe WeinBeRg - SAlam | \
ITEP, B.Cheremushkin skaya 25,Mosco w,117 25 9, R ussi a\
E-mail: - | A.I.Veselov\
I TEP,B. C here m us hkins kaya 25 , M o s cow ,11 725 9, Ru ssia\
E-mai l:
title:Upp er bound on th e c utoff in t heStandard Mod el---
I nt rod u ction
== ===== =====Accord ing to th ec onvent i onal po i n tof v iew the upper bou n d$ \Lambda$ on th e cuto ff in t heEle ctroweak t he ory ( w ithoutf er m i o ns) depends on th e Higgs mas s . I t is d ec rea s ed whe n the H i ggs mass is in crea sed. Andat the Higgs m a ss arou nd $1$ Te v $ \Lam b da $bec om e s o f t heo rde r of $M_ H$ .At th e sa m e t imefor $M_ H \si m
200$ Gev th e v alue of$\Lam bda$canbe made almos t inf in ite[^1]. This c oncl usion ismad ebas in g ont he per tur bat ion exp ansiona rou nd t r iv ial vacuum. In our p r e se ntationwe dem o ns tr a te thatth e v acuu m of th e la t ti ce Weinb erg -S al am modelis rathe rcom pli cated , whi ch mea ns thatthe a p plication of t h e perturbatio n e x p an s ionaro und trivial vac u um m ay b e l imi t ed.
Namel y, we investigate the beh av ior of thetopological d efects com p o s ed of th e la t ti c e gauge fields that are to be identifi ed wi th quant um Nambum o nopoles[@N amb u;@BV Z ; @ Chernodub_Nam b u ]. W eshow th attheir l att ice de nsi ty increase s alongth eli ne s o f con s tant phy si cswh enthe u l travio let c utof fin inc reased. At s uffi ci en tlylar ge valu es o f th e cutof f these o bje c ts b eg in to dom inate.
Movin gfurther al on g t he lin e of const ant physics we reach th e pointonthe p hase diagramwhe re the mo n opoleworldl inesbe gin t o per c o la te. T his pointr o ugh ly co in cide s withthe position of th e tr ansition betw een the p hy sic a lH igg sp has e and the unphysi cal symmet ri c p hase of th e la tt ice mod el. Atinfin i te bare scalar s elf coupl in g $\ l a mbd a$ the tra nsitionis a cros s overa nd theult raviol et cu toffachiev e s i ts ma ximalva lue ar ound$1 .4$ Tevat the transition point . At s malle r b are value s o f $\ lambda$ c orre spondent t o s mal l Hig gsm asses the ph ase trans itio n becomess tr ong e r .Still we do n o t k now t heo rder o f th e phase transitio n at small valu es o f $\l amb d a$.We have estimate d t he m aximal v al ue of the c utoff in t h e vic inityof the transi t i on pointat $ \la mbda = 0. 009 $. The obt ai ne d value ofth e cuto ff app e arst o be around $1.4$ Tev. The l a tti ce mo de l under inve stigation=========== ====== ==== ===== ======= == ==
Th e l at tice Weinb e rg - Sala m | \
_ _ITEP, B.Cheremushkinskaya 25, Moscow,_117259, Russia\
__ _E-mail:
-_|
_ A.I.Veselov\
_ ITEP, B.Cheremushkinskaya_25, Moscow, 117259,_Russia\
_ E-mail:
title: Upper bound on the cutoff in the Standard Model
---
Introduction
============
According to_the_conventional point_of_view_the upper bound $\Lambda$ on_the cutoff in the Electroweak_theory (without_fermions) depends on the Higgs mass. It is_decreased_when the Higgs_mass is increased. And at the Higgs mass around_$1$ Tev $\Lambda$ becomes of the_order of $M_H$._At_the_same time for $M_H_\sim
200$ Gev the value of $\Lambda$_can be made almost infinite[^1]. This_conclusion is made basing on the perturbation_expansion around trivial vacuum. In our_presentation we demonstrate that the_vacuum of_the lattice Weinberg - Salam_model is rather_complicated, which_means that the_application of the perturbation expansion around_trivial vacuum may_be limited.
Namely, we investigate the behavior_of_the topological defects_composed_of_the lattice_gauge fields that_are_to be_identified_with quantum Nambu monopoles [@Nambu; @BVZ;_@Chernodub_Nambu]._We show that their lattice density increases_along the lines of_constant_physics when the ultraviolet_cutoff in increased. At sufficiently_large values of the cutoff these_objects begin_to dominate.
Moving_further along the line of constant physics we reach the point_on the phase diagram where the_monopole worldlines begin to_percolate. This_point_roughly coincides with_the_position of_the transition between the physical Higgs phase_and the_unphysical symmetric phase of the lattice_model. At infinite bare_scalar_self coupling $\lambda$ the transition is_a crossover and the ultraviolet cutoff_achieves its maximal value around_$1.4$_Tev_at the transition point. At_smaller bare values of $\lambda$ correspondent_to small Higgs_masses the phase transition becomes stronger. Still_we_do not know the order of_the_phase transition at small values of_$\lambda$._We_have estimated the maximal value_of the cutoff in the vicinity_of the transition point at $\lambda = 0.009$. The_obtained value of_the cutoff appears to be_around_$1.4$_Tev.
The lattice model under investigation
=====================================
The lattice Weinberg - Salam |
right}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\kappa\setminus \bigcup_{\alpha<\beta<\kappa}[x_{\beta}{\mathord{\upharpoonright}}\xi_{\beta}]$$ is a nowhere dense set for any $\alpha<\kappa$. Hence, $A$ is $\kappa$-meagre.$\square$
\[lusin-hur\] If $\kappa<\lambda\leq 2^\kappa$, and $L\subseteq {{\mathfrak{K}}}^\kappa$ is a $\lambda$-$\kappa$-Lusin set, then $L$ does not have $\kappa$-Hurewicz property.
$\square$
### $\kappa$-Menger property
A set has [[**$\kappa$-Menger property**]{}]{} if it satisfies $U_{<\kappa}^\kappa({{\mathcal{O}}}_\kappa,{{\mathcal{O}}}_\kappa)$ principle.
Despite that every Lusin set for $\kappa$ lacks $\kappa$-Hurewicz property (see Corollary \[lusin-hur\]), it has $\kappa$-Menger property.
\[lus-men\] Let $L\subseteq {{\mathfrak{K}}}^\kappa$ be a Lusin set for $\kappa$. Then $L$ has $\kappa$-Menger property.
Proof: Let $\{s_\alpha\colon \alpha<\kappa\}=\{s\in {{\mathfrak{K}}}^{<\kappa}\colon [s]\cap L\neq 0\}$, and let $\{x_\alpha\colon \alpha<\kappa\}$ be such that $x_\alpha\in [s_\alpha]\cap L$ for all $\alpha<\kappa$.
Let $\langle{{\mathcal{U}}}_\alpha\rangle_{\alpha<\kappa}$ be a sequence of open covers essentially of size $\kappa$. For $\alpha<\kappa$, let $U_\alpha\in{{\mathcal{U}}}_\kappa$ be such that $x_\alpha\in U_\alpha$. Then, $L\setminus \bigcup_{\alpha<\kappa}U_\alpha$ is nowhere dense, hence $|L\setminus \bigcup_{\alpha<\kappa}U_\alpha|\leq\kappa$. Thus let $$L\setminus \bigcup_{\alpha<\kappa}U_\alpha=\{y_\alpha\colon \alpha<\kappa\}.$$ For all $\alpha<\kappa$, let $V_\alpha\in {{\mathcal{U}}}_\alpha | right}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\kappa\setminus \bigcup_{\alpha<\beta<\kappa}[x_{\beta}{\mathord{\upharpoonright}}\xi_{\beta}]$$ is a nowhere dense set for any $ \alpha<\kappa$. Hence, $ A$ is $ \kappa$-meagre.$\square$
\[lusin - hur\ ] If $ \kappa<\lambda\leq 2^\kappa$, and $ L\subseteq { { \mathfrak{K}}}^\kappa$ is a $ \lambda$-$\kappa$-Lusin set, then $ L$ does not have $ \kappa$-Hurewicz place.
$ \square$
# # # $ \kappa$-Menger place
ampere set has [ [ * * $ \kappa$-Menger place * * ] { } ] { } if it satisfies $ U_{<\kappa}^\kappa({{\mathcal{O}}}_\kappa,{{\mathcal{O}}}_\kappa)$ principle.
Despite that every Lusin set for $ \kappa$ miss $ \kappa$-Hurewicz place (see Corollary \[lusin - hur\ ]), it has $ \kappa$-Menger property.
\[lus - men\ ] Let $ L\subseteq { { \mathfrak{K}}}^\kappa$ be a Lusin set for $ \kappa$. Then $ L$ suffer $ \kappa$-Menger property.
Proof: get $ \{s_\alpha\colon \alpha<\kappa\}=\{s\in { { \mathfrak{K}}}^{<\kappa}\colon [ s]\cap L\neq 0\}$, and lease $ \{x_\alpha\colon \alpha<\kappa\}$ be such that $ x_\alpha\in [ s_\alpha]\cap L$ for all $ \alpha<\kappa$.
Let $ \langle{{\mathcal{U}}}_\alpha\rangle_{\alpha<\kappa}$ be a sequence of open covers essentially of size $ \kappa$. For $ \alpha<\kappa$, permit $ U_\alpha\in{{\mathcal{U}}}_\kappa$ be such that $ x_\alpha\in U_\alpha$. Then, $ L\setminus \bigcup_{\alpha<\kappa}U_\alpha$ is nowhere dense, hence $ |L\setminus \bigcup_{\alpha<\kappa}U_\alpha|\leq\kappa$. Thus let $ $ L\setminus \bigcup_{\alpha<\kappa}U_\alpha=\{y_\alpha\colon \alpha<\kappa\}.$$ For all $ \alpha<\kappa$, lease $ V_\alpha\in { { \mathcal{U}}}_\alpha | rigjt}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\kakpa\setminus \bigcuk_{\aopha<\beva<\kappa}[s_{\beta}{\matford{\upharpoonright}}\xi_{\beta}]$$ is e nowyere eense set for any $\alphx<\kappa$. Hejce, $A$ is $\kapka$-meagre.$\square$
\[lusmh-hur\] If $\kappa<\mwmbdc\lxq 2^\kappa$, and $L\sobseteq {{\mathxrak{K}}}^\kappa$ is d $\lxmyda$-$\kappa$-Lusin set, then $L$ does not hade $\kapps$-Hkrewicz properjy.
$\squswe$
### $\kzipc$-Menger property
A set has [[**$\kappa$-Msnger pgoperty**]{}]{} if it satosfies $U_{<\kappa}^\kappa({{\mathcal{O}}}_\nappw,{{\mathcal{O}}}_\kappa)$ prijciple.
Despijs trqt every Luskn set for $\kappa$ lacks $\kappa$-Hurewicz property (see Corullarv \[lusin-hur\]), ij gad $\kappa$-Mengec propvrty.
\[lus-men\] Leb $L\subsateq {{\mayhfrak{K}}}^\kappa$ bc a Luvin set for $\kappa$. Then $L$ has $\kappa$-Menger ptoperty.
Promf: Let $\{s_\alpha\colon \qlpha<\kakpa\}=\{s\it {{\magyfrxk{K}}}^{<\ia'pa}\dolon [d]\ca' L\neq 0\}$, and let $\{x_\alpha\xolon \alpha<\kappa\}$ be stbn that $x_\alphz\in [s_\ajpra]\cap L$ for all $\alpha<\kappa$.
Let $\langle{{\matvcam{U}}}_\alpha\rangle_{\alpha<\kappa}$ be a sequence of open covers efsentially of size $\kappa$. For $\alpha<\kappa$, let $U_\alphd\in{{\mavhzal{B}}}_\jappa$ ve such that $x_\alpha\in U_\alpha$. Then, $L\setminus \bigshp_{\sliha<\kappa}U_\alpha$ is nowhere demsf, ngnce $|L\setminus \bigcu'_{\zllha<\kappa}U_\alpha|\leq\kwppa$. Thos let $$L\setminuf \bibcup_{\alpha<\kappa}U_\alpha=\{y_\alpha\cilon \alpha<\kaipa\}.$$ Dor all $\alpha<\kappa$, let $V_\alpha\nn {{\matncal{U}}}_\slpha | right}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\kappa\setminus \bigcup_{\alpha<\beta<\kappa}[x_{\beta}{\mathord{\upharpoonright}}\xi_{\beta}]$$ is a nowhere dense set $\alpha<\kappa$. $A$ is \[lusin-hur\] If $\kappa<\lambda\leq a set, then $L$ not have $\kappa$-Hurewicz $\square$ ### $\kappa$-Menger property A set [[**$\kappa$-Menger property**]{}]{} if it satisfies $U_{<\kappa}^\kappa({{\mathcal{O}}}_\kappa,{{\mathcal{O}}}_\kappa)$ principle. Despite that every Lusin set for lacks $\kappa$-Hurewicz property (see Corollary \[lusin-hur\]), it has $\kappa$-Menger property. \[lus-men\] Let $L\subseteq be Lusin for Then $L$ has $\kappa$-Menger property. Proof: Let $\{s_\alpha\colon \alpha<\kappa\}=\{s\in {{\mathfrak{K}}}^{<\kappa}\colon [s]\cap L\neq 0\}$, and let $\{x_\alpha\colon be such that $x_\alpha\in [s_\alpha]\cap L$ for all Let $\langle{{\mathcal{U}}}_\alpha\rangle_{\alpha<\kappa}$ be a of open covers essentially of $\kappa$. $\alpha<\kappa$, let be that U_\alpha$. Then, $L\setminus is nowhere dense, hence $|L\setminus \bigcup_{\alpha<\kappa}U_\alpha|\leq\kappa$. Thus let $$L\setminus \bigcup_{\alpha<\kappa}U_\alpha=\{y_\alpha\colon \alpha<\kappa\}.$$ For all $\alpha<\kappa$, let $V_\alpha\in {{\mathcal{U}}}_\alpha | right}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\Kappa\setmiNus \biGcuP_{\alPhA<\betA<\kapPa}[x_{\beta}{\mathord{\UPharPoonright}}\xi_{\beta}]$$ is a nowhEre deNsE Set fOR aNy $\alpHa<\kappa$. hEnCE, $a$ is $\KaPpA$-meAgRE.$\sQuare$
\[LusIn-hur\] If $\Kappa<\lambdA\leQ 2^\kAppa$, and $L\subsETeQ {{\mathfrak{K}}}^\KapPa$ is a $\lambda$-$\kAppA$-Lusin SeT, thEN $L$ doeS noT have $\Kappa$-HURewicz Property.
$\sQuARe$
### $\kappA$-menger pROPeRty
A Set has [[**$\kappa$-Menger PRoPErty**]{}]{} if it satisfIes $U_{<\kaPpA}^\KaPPA({{\maThcAl{O}}}_\kappa,{{\maThCal{O}}}_\kAPpa)$ prinCIpLE.
dEspITe that every LuSin set for $\kaPPa$ lAcks $\kaPpA$-HuREwicz pRoperTy (SEe COrollary \[lusIn-huR\]), it has $\kapPa$-MengER properTY.
\[lus-men\] let $L\suBseTeq {{\MathFRaK{K}}}^\KapPa$ BE a LUSiN seT For $\Kappa$. TheN $L$ HaS $\kappA$-MenGER PRopeRty.
prooF: Let $\{s_\Alpha\colon \alpHa<\kAppa\}=\{S\In {{\mAthfrAk{K}}}^{<\kaPpa}\cOlOn [s]\caP L\neq 0\}$, aNd let $\{X_\aLpha\colon \alpha<\kAppa\}$ Be such thaT $x_\aLpHa\iN [s_\Alpha]\CAp L$ for All $\AlpHa<\kappa$.
let $\langLE{{\maThCAL{u}}}_\aLpha\rangle_{\alpha<\kapPa}$ BE A sEquence oF open cOVeRs ESsentialLy Of sIze $\kAPPa$. For $\AlphA<\KaPpa$, let $U_\aLpha\in{{\MAtHcAl{U}}}_\kappA$ bE such tHaT $x_\aLphA\in U_\aLPha$. THen, $L\seTminus \biGcup_{\aLPha<\kappa}U_\alpha$ IS nowhere dense, HEnCE $|l\sETminUs \bIgcup_{\alpha<\kAppa}u_\AlphA|\leq\KApPa$. THUs let $$l\setmInUS \bIGcup_{\alpha<\kappa}U_\alphA=\{y_\Alpha\cOlon \aLpha<\kappa\}.$$ For aLl $\alpha<\kapPA$, LEt $V_\alpha\In {{\maTHcAL{U}}}_\alpha | right}}\xi_{\beta}]\right) ={{\mathfr ak{K} }}^ \ka pp a\se tmin us \bigcup_{\a l pha< \beta<\kappa}[x_{\beta }{\ma th o rd{\ u ph arpoo nright} } \x i _ {\b et a} ]$$ i s a nowh ere denseset for an y $ \a lpha<\kappa$ . H ence, $A$is$\kappa$-mea gre .$\squ ar e$\[lus in- hur\] If $\ k appa<\ lambda\le q2 ^\kapp a $, and$ L \s ubse teq {{\mathfrak{K } }} ^ \kappa$ is a $ \lambd a$ - $\ k a ppa $-L usin set,th en $L $ does n o th a v e $ \ kappa$-Hurewi cz property .
$ \squar e$
# # # $\ka ppa$- Me n ger property
A se t has [[* *$\kap p a$-Meng e r prope rty**] {}] {}if i t s at isf ie s $U _ {< \ka p pa} ^\kappa( {{ \m athca l{O} } } _ \ kapp a,{ {\ma thcal {O}}}_\kappa) $ p rinc i ple .
De spite tha tevery Lusin setfo r $\kappa$ lack s $\ kappa$-Hu rew ic z p ro perty (see C oro lla ry \[lu sin-hur \ ]), i t h as $\kappa$-Menger p ro p e rt y.
\[lu s-men\ ] L et $L\subse te q { {\ma t h frak{ K}}} ^ \k appa$ be a Lus i nse t for $ \k appa$. T hen $L $ has $\ka ppa$-M enger pr opert y .
Proof: Let$ \{s_\alpha\co l on \ al p ha<\ kap pa\}=\{s\in {{\ m athf rak{ K }} }^{ < \kapp a}\co lo n [ s ]\cap L\neq 0\}$, a nd let $ \{x_\ alpha\colon \ alpha<\kap p a \ }$ be su ch t h at $x_\alpha\in [ s_\al pha]\cap L $ for all $\al pha<\kap pa$.
Let $ \langle{ {\m ath cal {U} } } _\ alpha\rangle_ { \ alph a< \kappa} $ b e a seq uen ceofope ncovers es sentiall yof s iz e $ \kapp a $. For $ \a lph a< \ka ppa$, let $U _\alp ha\i n{ {\ m ath cal{U}} } _\ k a ppa$ b esuch th at $x_\ alph a \in U_\alp ha$. Then , $ L \set mi nu s \bigc up_{\alpha<\k ap pa}U_\alph a$ is nowhe r e dense,hence $|L\setminus \big c up_{\al pha <\kap pa}U _\alpha|\ leq \kappa $.T hus le t $$L\ setmi nu s \ b i gcup_ { \ al pha <\ kappa}U_\a l p ha= \{y_\ al pha\ colon \ alpha<\kappa\}.$$F orall $\alpha<\ kap pa$, l et $V _ \a l pha \i n {{ \ m athcal{U}}}_\al pha | right}}\xi_{\beta}]\right)={{\mathfrak{K}}}^\kappa\setminus \bigcup_{\alpha<\beta<\kappa}[x_{\beta}{\mathord{\upharpoonright}}\xi_{\beta}]$$_is a nowhere_dense set for any_$\alpha<\kappa$. Hence,_$A$_is $\kappa$-meagre.$\square$
\[lusin-hur\]_If_$\kappa<\lambda\leq 2^\kappa$, and_$L\subseteq {{\mathfrak{K}}}^\kappa$ is_a $\lambda$-$\kappa$-Lusin set, then $L$_does not have_$\kappa$-Hurewicz_property.
$\square$
### $\kappa$-Menger property
A set has [[**$\kappa$-Menger property**]{}]{} if it satisfies $U_{<\kappa}^\kappa({{\mathcal{O}}}_\kappa,{{\mathcal{O}}}_\kappa)$ principle.
Despite that every Lusin_set_for $\kappa$_lacks_$\kappa$-Hurewicz_property (see Corollary \[lusin-hur\]), it has_$\kappa$-Menger property.
\[lus-men\] Let $L\subseteq {{\mathfrak{K}}}^\kappa$_be a Lusin_set for $\kappa$. Then $L$ has $\kappa$-Menger property.
Proof:_Let_$\{s_\alpha\colon \alpha<\kappa\}=\{s\in {{\mathfrak{K}}}^{<\kappa}\colon_[s]\cap L\neq 0\}$, and let $\{x_\alpha\colon \alpha<\kappa\}$ be such_that $x_\alpha\in [s_\alpha]\cap L$ for all_$\alpha<\kappa$.
Let $\langle{{\mathcal{U}}}_\alpha\rangle_{\alpha<\kappa}$ be_a sequence_of_open covers essentially of_size $\kappa$. For $\alpha<\kappa$, let $U_\alpha\in{{\mathcal{U}}}_\kappa$_be such that $x_\alpha\in U_\alpha$. Then,_$L\setminus \bigcup_{\alpha<\kappa}U_\alpha$ is nowhere dense, hence $|L\setminus_\bigcup_{\alpha<\kappa}U_\alpha|\leq\kappa$. Thus let $$L\setminus \bigcup_{\alpha<\kappa}U_\alpha=\{y_\alpha\colon \alpha<\kappa\}.$$_For all $\alpha<\kappa$, let $V_\alpha\in_{{\mathcal{U}}}_\alpha |
partitions of $n$ with distinct parts, see [@jozef1989; @stemb1988; @stemb1989]. For any $\lambda\in \mathcal{D}(n)$, let $\langle\lambda\rangle$ denote the character of the irreducible projective or spin representation indexed by $\lambda$. Morris [@morri1965] has found a combinatorial rule for calculating the characters, which is the projective analogue of the Murnaghan-Nakayama rule. In terms of bar tableaux, Morris’s theorem reads as follows:
\[mnrule\] Let $\lambda\in \mathcal{D}(n)$ and $\pi\in
\mathcal{P}^o(n)$. Then $$\label{mnruleeq}
\langle\lambda\rangle(\pi)=\sum_{T}wt(T)$$ where the sum ranges over all bar tableaux of shape $\lambda$ and type $\pi$.
The above theorem for projective characters implies the following formula, which will be used later in the proof of Lemma \[len2\].
\[2odd\] Let $\lambda$ be a strict partition of length $2$. Suppose that the two parts $\lambda_1,\lambda_2$ are both odd. Then we have $$\langle\lambda\rangle(\lambda)=-1.$$
Let $T$ be the bar tableau obtained by filling the last $\lambda_2$ squares in the first row of $S(\lambda)$ with $2$’s and the remaining squares with $1$’s, and let $T'$ be the bar tableau obtained by filling the first row of $S(\lambda)$ with $1$’s and the second row with $2$’s. Clearly, $T$ and $T'$ are of the same type $\lambda$. Let us first consider the weight of $T$. If $\lambda_1-\lambda_2<\lambda_2$, then $$wt(T)=(-1)^{2-1} 2^{1-0}\cdot (-1)^{2-1+\lambda_2}2^{1-1}=-2.$$ If $\lambda_1-\lambda_2>\lambda_2$, then $$wt(T)=(-1)^{1-1} 2^{1-0}\cdot
(-1)^{2-1+\lambda_1-\lambda_2}2^{1-1}=-2.$$ In both cases, the weight of $T'$ equals $$wt(T')=(-1)^{2-2 | partitions of $ n$ with distinct parts, see [ @jozef1989; @stemb1988; @stemb1989 ]. For any $ \lambda\in \mathcal{D}(n)$, lease $ \langle\lambda\rangle$ announce the character of the irreducible projective or spin theatrical performance indexed by $ \lambda$. Morris [ @morri1965 ] has discover a combinatorial principle for forecast the characters, which is the projective analogue of the Murnaghan - Nakayama convention. In terms of bar tableaux, Morris ’s theorem reads as keep up:
\[mnrule\ ] Let $ \lambda\in \mathcal{D}(n)$ and $ \pi\in
\mathcal{P}^o(n)$. Then $ $ \label{mnruleeq }
\langle\lambda\rangle(\pi)=\sum_{T}wt(T)$$ where the union range over all bar tableaux of shape $ \lambda$ and type $ \pi$.
The above theorem for projective characters implies the following formula, which will be use later in the proof of Lemma \[len2\ ].
\[2odd\ ] Let $ \lambda$ be a rigid partition of length $ 2$. Suppose that the two parts $ \lambda_1,\lambda_2 $ are both curious. Then we have $ $ \langle\lambda\rangle(\lambda)=-1.$$
Let $ T$ be the bar tableau obtained by filling the final $ \lambda_2 $ squares in the first row of $ S(\lambda)$ with $ 2 $ ’s and the remaining squares with $ 1 $ ’s, and let $ T'$ be the bar tableau obtained by filling the first row of $ S(\lambda)$ with $ 1 $ ’s and the second row with $ 2 $ ’s. Clearly, $ T$ and $ T'$ are of the same type $ \lambda$. Let us first see the weight of $ T$. If $ \lambda_1-\lambda_2<\lambda_2 $, then $ $ wt(T)=(-1)^{2 - 1 } 2^{1 - 0}\cdot (-1)^{2 - 1+\lambda_2}2^{1 - 1}=-2.$$ If $ \lambda_1-\lambda_2>\lambda_2 $, then $ $ wt(T)=(-1)^{1 - 1 } 2^{1 - 0}\cdot
(-1)^{2 - 1+\lambda_1-\lambda_2}2^{1 - 1}=-2.$$ In both cases, the weight unit of $ T'$ equal $ $ wt(T')=(-1)^{2 - 2 | pagtitions of $n$ with distikct parts, see [@joeed1989; @stemu1988; @stemb1989]. For any $\lambda\in \mathcal{D}(n)$, let $\langpe\oambdq\rangle$ denote the chafacter of the irrwducmble projective or spin vzpressktatimi indexed by $\lakbda$. Morriv [@morri1965] has fogna c combinatorial rule for calculating the chsrwcters, which if tht pwojedniye analogue of the Murnaghan-Nakagama rune. In terms og bar tableaux, Morris’s thelrem reads as follows:
\[mjrule\] Let $\lqmbdw\un \mathcal{D}(n)$ and $\pi\in
\mathcal{P}^o(n)$. Thgn $$\label{mnruleeq}
\langle\lambda\rangue(\pi)=\sbm_{T}wt(T)$$ wherg fhf sum ranges over all bar tableaux of shape $\kambda$ and typc $\pi$.
Tie avove theorem for projxctive characters imklies the xomlowing formula, wyixh winl ba usde lxted mn fhe prlof of Lemma \[men2\].
\[2odd\] Let $\oambda$ be a strict karepyion of lengfh $2$. Su[pjse that the two parts $\lambda_1,\lambda_2$ are bofh odd. Then we have $$\lantle\lambda\rangle(\lambda)=-1.$$
Pet $T$ be ehe bar tableau obtained by filling the last $\lambga_2$ sqnafes ik thd flrst row of $S(\lambda)$ with $2$’s and the remaining fsusrvs with $1$’s, and let $T'$ be the bsr tsfleau obtainea by fnmljng the first row lf $S(\lamfda)$ wuth $1$’s and the second row with $2$’s. Clearly, $R$ and $T'$ are jd the same type $\lalbda$. Let us firxt comsider the weight of $T$. Nf $\lamgda_1-\lambda_2<\lalbda_2$, then $$dt(T)=(-1)^{2-1} 2^{1-0}\cdot (-1)^{2-1+\lambda_2}2^{1-1}=-2.$$ Kf $\kakbda_1-\lambda_2>\lambda_2$, then $$wt(T)=(-1)^{1-1} 2^{1-0}\sdot
(-1)^{2-1+\lambde_1-\lambba_2}2^{1-1}=-2.$$ In bogh csses, tre weight lf $T'$ cxuals $$wt(T')=(-1)^{2-2 | partitions of $n$ with distinct parts, see @stemb1989]. any $\lambda\in let $\langle\lambda\rangle$ denote projective spin representation indexed $\lambda$. Morris [@morri1965] found a combinatorial rule for calculating characters, which is the projective analogue of the Murnaghan-Nakayama rule. In terms of tableaux, Morris’s theorem reads as follows: \[mnrule\] Let $\lambda\in \mathcal{D}(n)$ and $\pi\in \mathcal{P}^o(n)$. $$\label{mnruleeq} where sum over all bar tableaux of shape $\lambda$ and type $\pi$. The above theorem for projective characters the following formula, which will be used later the proof of Lemma \[2odd\] Let $\lambda$ be a partition length $2$. that two $\lambda_1,\lambda_2$ are both Then we have $$\langle\lambda\rangle(\lambda)=-1.$$ Let $T$ be the bar tableau obtained by filling the last $\lambda_2$ squares the first $S(\lambda)$ with and remaining with $1$’s, and be the bar tableau obtained by row of $S(\lambda)$ with $1$’s and the second with $2$’s. $T$ and $T'$ are of the type $\lambda$. Let us first consider the weight $T$. If $\lambda_1-\lambda_2<\lambda_2$, then $$wt(T)=(-1)^{2-1} 2^{1-0}\cdot (-1)^{2-1+\lambda_2}2^{1-1}=-2.$$ If $\lambda_1-\lambda_2>\lambda_2$, then $$wt(T)=(-1)^{1-1} 2^{1-0}\cdot (-1)^{2-1+\lambda_1-\lambda_2}2^{1-1}=-2.$$ In both weight of $T'$ equals | partitions of $n$ with distinct Parts, see [@joZef1989; @stEmb1988; @SteMb1989]. for aNy $\laMbda\in \mathcal{D}(N)$, Let $\lAngle\lambda\rangle$ denotE the cHaRActeR Of The irReducibLE pROJecTiVe Or sPiN RePreseNtaTion indExed by $\lambDa$. MOrRis [@morri1965] has fOUnD a combinatOriAl rule for calCulAting tHe ChaRActerS, whIch is The proJEctive Analogue oF tHE MurnaGHan-NakaYAMa Rule. in terms of bar tableAUx, mOrris’s theorem rEads as FoLLoWS:
\[MnrUle\] let $\lambda\iN \mAthcaL{d}(n)$ and $\pi\IN
\mATHCal{p}^O(n)$. Then $$\label{mnRuleeq}
\langlE\LamBda\ranGlE(\pi)=\SUm_{T}wt(T)$$ Where ThE Sum Ranges over aLl baR tableaux Of shapE $\Lambda$ aND type $\pi$.
the aboVe tHeoRem fOR pRoJecTiVE chARaCteRS imPlies the FoLlOwing FormULA, WHich WilL be uSed laTer in the proof Of LEmma \[LEn2\].
\[2oDd\] Let $\LambdA$ be a StRict pArtitiOn of lEnGth $2$. Suppose that tHe twO parts $\lamBda_1,\LaMbdA_2$ aRe botH Odd. TheN we HavE $$\langle\Lambda\rANglE(\lAMBDa)=-1.$$
let $T$ be the bar tableaU oBTAiNed by filLing thE LaSt $\LAmbda_2$ squArEs iN the FIRst roW of $S(\LAmBda)$ with $2$’s And the REmAiNing squArEs with $1$’S, aNd lEt $T'$ Be the BAr taBleau oBtained bY fillINg the first row oF $s(\lambda)$ with $1$’s aND tHE SeCOnd rOw wIth $2$’s. Clearly, $t$ and $t'$ Are oF the SAmE tyPE $\lambDa$. Let Us FIrST consider the weight oF $T$. if $\lambDa_1-\lamBda_2<\lambda_2$, then $$Wt(T)=(-1)^{2-1} 2^{1-0}\cdot (-1)^{2-1+\lamBDA_2}2^{1-1}=-2.$$ if $\lambda_1-\LambDA_2>\lAMbda_2$, then $$wt(T)=(-1)^{1-1} 2^{1-0}\cdoT
(-1)^{2-1+\lambDa_1-\lambda_2}2^{1-1}=-2.$$ In BOth cases, The weIght of $T'$ eQuals $$wt(T')=(-1)^{2-2 | partitions of $n$ with di stinct par ts, s ee[@j oz ef19 89;@stemb1988; @s t emb1 989]. For any $\lambda \in \ ma t hcal { D} (n)$, let $\ l an g l e\l am bd a\r an g le $ den ote the ch aracter of th eirreduciblep ro jective or sp in represent ati on ind ex edb y $\l amb da$.Morris [@morr i1965] ha sf ound a combina t o ri al r ule for calculati n gt he characters, which i s t h e pr oje ctive anal og ue of the Mur n ag h a n -Na k ayama rule. I n terms ofb artablea ux , M o rris’s theo re m re ads as foll ows:
\[mnrul e\] Le t $\lamb d a\in \m athcal {D} (n) $ an d $ \p i\i n\ mat h ca l{P } ^o( n)$. The n$$ \labe l{mn r u l e eq}\la ngle \lamb da\rangle(\pi )=\ sum_ { T}w t(T)$ $ whe re t he sumranges over a ll bar tableaux ofshape $\l amb da $ a nd type $\pi$.
T heabove t heoremf orpr o j e ct ive characters imp li e s t he follo wing f o rm ul a , whichwi llbe u s e d lat er i n t he proof of Le m ma \ [len2\] .
\[2od d\ ] L et$\lam b da$be a s trict pa rtiti o n of length $2 $ . Suppose tha t t h e t w o pa rts $\lambda_1 ,\la m bda_ 2$ a r ebot h odd. Then w e h a ve $$\langle\lambda \r angle( \lamb da)=-1.$$
Le t $T$ be t h e bar tabl eauo bt a ined by fillin g the last $\la m bda_2$ s quare s in the first ro w of $S(\l amb da) $ w ith $ 2$ ’s and the re m a inin gsquares wi th $1$’ s,and le t $ T' $ be thebar tabl ea uob ta ine d byf illing t he fi rs t r ow of $S(\la mbda) $ wi th $ 1 $’s and th e s e c ondro wwith $2 $’ s. Cl earl y , $ T$ and$T'$ areoft he s am etype $\ lambda$. Letus first con si der the w e i ght of $ T$. If $\lambda_1-\lamb d a_2<\la mbd a_2$, the n $$wt(T) =(- 1)^{2- 1}2 ^{1-0} \cdot(-1)^ {2 -1+ \ l ambda _ 2 }2 ^{1 -1 }=-2.$$ If $ \la mbda_ 1- \lam bda_2>\ lambda_2$, then $$ w t(T )=(-1)^{1-1}2^{ 1-0} \ c do t
( - 1) ^ {2- 1+ \ lam b d a_1-\lambda_2}2 ^{1-1}=-2. $$ In both case s , t he weight of $T' $ equ a ls $$wt (T')=(-1) ^{2-2 | partitions_of $n$_with distinct parts, see_[@jozef1989; @stemb1988;_@stemb1989]._For any_$\lambda\in_\mathcal{D}(n)$, let $\langle\lambda\rangle$_denote the character_of the irreducible projective_or spin representation_indexed_by $\lambda$. Morris [@morri1965] has found a combinatorial rule for calculating the characters, which_is_the projective_analogue_of_the Murnaghan-Nakayama rule. In terms_of bar tableaux, Morris’s theorem_reads as_follows:
\[mnrule\] Let $\lambda\in \mathcal{D}(n)$ and $\pi\in
\mathcal{P}^o(n)$. Then $$\label{mnruleeq}
\langle\lambda\rangle(\pi)=\sum_{T}wt(T)$$_where_the sum ranges_over all bar tableaux of shape $\lambda$ and type_$\pi$.
The above theorem for projective characters_implies the following_formula,_which_will be used later_in the proof of Lemma \[len2\].
\[2odd\]_Let $\lambda$ be a strict partition_of length $2$. Suppose that the two_parts $\lambda_1,\lambda_2$ are both odd. Then_we have $$\langle\lambda\rangle(\lambda)=-1.$$
Let $T$ be_the bar_tableau obtained by filling the_last $\lambda_2$ squares_in the_first row of_$S(\lambda)$ with $2$’s and the remaining_squares with $1$’s,_and let $T'$ be the bar_tableau_obtained by filling_the_first_row of_$S(\lambda)$ with $1$’s_and_the second_row_with $2$’s. Clearly, $T$ and $T'$_are_of the same type $\lambda$. Let us_first consider the weight_of_$T$. If $\lambda_1-\lambda_2<\lambda_2$, then_$$wt(T)=(-1)^{2-1} 2^{1-0}\cdot (-1)^{2-1+\lambda_2}2^{1-1}=-2.$$ If $\lambda_1-\lambda_2>\lambda_2$,_then $$wt(T)=(-1)^{1-1} 2^{1-0}\cdot
(-1)^{2-1+\lambda_1-\lambda_2}2^{1-1}=-2.$$ In both cases,_the weight_of $T'$_equals $$wt(T')=(-1)^{2-2 |
$V\ \in\ \textrm{Lie}({\cal G})$ maps a configuration with zero news to a configuration with zero news because the definition of the news before and after the action of $V$ refer to different frames. Hence choosing a $q_{AB}$ (and working with $\Gamma^{q}$) implies breaking the ${\cal G}$ symmetry spontaneously to $\textrm{BMS}$ and the subleading soft gravitons can be thought of as goldstone modes associated to this symmetry breaking as they map one family of vacua (associated to a given $q_{AB}$) to a distinct family of vacua associated to a different $q_{AB}$.
Evaluating $\delta_{V}N_{AB}$ {#ssb1}
-----------------------------
From (\[deltaV\]) and the definition of the news tensor (\[newsdef\]) we have \_V N\_[AB]{} & =& - \_u \_V C\_[AB]{}- \_V \^[(0)]{}\_[AB]{}\
& =& -Ł\_V \_[AB]{} - u \_[AB]{} + 2 (D\_A D\_B )\^ - \_V \^[(0)]{}\_[AB]{}. \[delVN2\] We now evaluate the last term in (\[delVN2\]). From $\delta_V q_{AB}= \L_Vq_{AB}- 2 \alpha q_{AB}$ we have that $\delta_V \rho_{AB}$ is a sum of a Lie derivative term plus a scale transformation term. Since the behaviour of $\rho_{AB}$ under scale transformation is known [@geroch] the effect of the second term can be obtained explicitly. The total change is found to be: \_[V]{}\_[AB]{}= Ł\_V \_[AB]{} + 2 D\_A D\_B, \[delrho\] which in turn implies, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V \^[(0)]{}\_[AB]{} + 2 (D\_A D\_B )\^.\[delrho0\] When substituting (\[delrho0\]) in (\[delVN2\]) the ‘soft’ factors cancel out and one obtains: \_V N\_[AB]{} & =& - Ł\_V \_[AB]{} - u \_[AB]{} - Ł\_V \^[(0)]{}\_[AB]{}\
&=& Ł\_V N\_[AB]{} + u \_[AB]{}, where in the second line we used the definition of the news tensor (\[newsdef\]) and the fact that $-\ddot{C}_{AB}= | $ V\ \in\ \textrm{Lie}({\cal G})$ maps a configuration with zero news to a configuration with zero news program because the definition of the news program before and after the action of $ V$ refer to different frame. Hence choosing a $ q_{AB}$ (and work with $ \Gamma^{q}$) implies breaking the $ { \cal G}$ isotropy ad lib to $ \textrm{BMS}$ and the subleading soft gravitons can be think of as goldstone modes associated to this symmetry break as they map one family of vacua (associated to a given $ q_{AB}$) to a distinct family of vacua associate to a different $ q_{AB}$.
Evaluating $ \delta_{V}N_{AB}$ { # ssb1 }
-----------------------------
From (\[deltaV\ ]) and the definition of the news tensor (\[newsdef\ ]) we consume \_V N\_[AB ] { } & = & - \_u \_V C\_[AB]{}- \_V \^[(0)]{}\_[AB]{}\
& = & -Ł\_V \_[AB ] { } - u \_[AB ] { } + 2 (D\_A D\_B) \^ - \_V \^[(0)]{}\_[AB ] { }. \[delVN2\ ] We now evaluate the last term in (\[delVN2\ ]). From $ \delta_V q_{AB}= \L_Vq_{AB}- 2 \alpha q_{AB}$ we experience that $ \delta_V \rho_{AB}$ is a sum of a Lie derivative term plus a scale transformation term. Since the behaviour of $ \rho_{AB}$ under plate transformation is known [ @geroch ] the effect of the second term can be obtained explicitly. The total change is find to be: \_[V]{}\_[AB]{}= Ł\_V \_[AB ] { } + 2 D\_A D\_B, \[delrho\ ] which in turn implies, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V \^[(0)]{}\_[AB ] { } + 2 (D\_A D\_B) \^.\[delrho0\ ] When substituting (\[delrho0\ ]) in (\[delVN2\ ]) the ‘ soft ’ factors cancel out and one obtains: \_V N\_[AB ] { } & = & - Ł\_V \_[AB ] { } - u \_[AB ] { } - Ł\_V \^[(0)]{}\_[AB]{}\
& = & Ł\_V N\_[AB ] { } + u \_[AB ] { }, where in the second line we used the definition of the news program tensor (\[newsdef\ ]) and the fact that $ -\ddot{C}_{AB}= | $V\ \ln\ \textrm{Lie}({\cal G})$ maps a configuration cuth zeco news to a covfiguration with zero news bxcauwe tht definition of the news beflre and qftec the action of $T$ refer bj dirnerenc hrames. Hence chposing a $q_{DB}$ (and working wktk $\Gamma^{q}$) implies breaking the ${\cal G}$ fymmetru dpontaneously jo $\teqtwm{BMA}$ and the subleading soft gravitona can bt thought of as gokdstone modes associated tl thls symmetry breakijg as they nap jbe family of vacua (associated to a given $q_{AB}$) to a distinct family of vccua associqtwd hm a differeit $q_{AB}$.
Vvaluating $\delta_{V}N_{AB}$ {#vsb1}
-----------------------------
From (\[deltaV\]) and thc defmnituon of the news tensoc (\[newsdef\]) we have \_V G\_[AB]{} & =& - \_u \_E E\_[AB]{}- \_V \^[(0)]{}\_[AB]{}\
& =& -Ł\_V \_[AB]{} - u \_[AB]{} + 2 (E\_A D\_B )\^ - \_V \^[(0)]{}\_[AB]{}. \[awlVV2\] Wt npw evaluwte the last ferm in (\[delCN2\]). From $\delta_V q_{AB}= \K_Vz_{QB}- 2 \alpha q_{AB}$ we hade that $\delta_V \rho_{AB}$ is a sum of a Lie degivafive term plus a scale rransformation term. Slnce the fehaviour of $\rho_{AB}$ under scale transformation is nnown [@eericm] thd evfect of the second term can be obtained explydiuly. The total chakge is found to be: \_[V]{}\_[WB]{}= Ł\_D \_[AB]{} + 2 D\_A D\_B, \[aelrho\] whjch in turn implied, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V \^[(0)]{}\_[AB]{} + 2 (D\_A D\_B )\^.\[dejrho0\] When substituting (\[delrho0\]) ib (\[delVN2\]) the ‘fift’ factors cancel out and onz obtaons: \_V N\_[AB]{} & =& - Ł\_V \_[AB]{} - u \_[AB]{} - Ł\_V \^[(0)]{}\_[AY]{}\
&=& Ł\_V N\_[AG]{} + u \_[AB]{}, whege in the recond line we ured tve definition of the news eensor (\[nexsdef\]) and thd favt thae $-\ddot{C}_{AB}= | $V\ \in\ \textrm{Lie}({\cal G})$ maps a configuration news a configuration zero news because before after the action $V$ refer to frames. Hence choosing a $q_{AB}$ (and with $\Gamma^{q}$) implies breaking the ${\cal G}$ symmetry spontaneously to $\textrm{BMS}$ and the soft gravitons can be thought of as goldstone modes associated to this symmetry as map family vacua (associated to a given $q_{AB}$) to a distinct family of vacua associated to a different Evaluating $\delta_{V}N_{AB}$ {#ssb1} ----------------------------- From (\[deltaV\]) and the of the news tensor we have \_V N\_[AB]{} & - \_V C\_[AB]{}- \^[(0)]{}\_[AB]{}\ =& \_[AB]{} - u + 2 (D\_A D\_B )\^ - \_V \^[(0)]{}\_[AB]{}. \[delVN2\] We now evaluate the last term in (\[delVN2\]). $\delta_V q_{AB}= \alpha q_{AB}$ have $\delta_V is a sum Lie derivative term plus a scale the behaviour of $\rho_{AB}$ under scale transformation is [@geroch] the of the second term can be explicitly. The total change is found to be: Ł\_V \_[AB]{} + 2 D\_A D\_B, \[delrho\] which in turn implies, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V \^[(0)]{}\_[AB]{} + D\_B )\^.\[delrho0\] When substituting in (\[delVN2\]) the factors out one \_V N\_[AB]{} =& - Ł\_V \_[AB]{} - u \_[AB]{} - Ł\_V \^[(0)]{}\_[AB]{}\ &=& N\_[AB]{} + u \_[AB]{}, where in the second line we definition the news tensor and the fact that | $V\ \in\ \textrm{Lie}({\cal G})$ maps a confIguration wIth zeRo nEws To A conFiguRation with zero NEws bEcause the definition of tHe newS bEFore ANd After The actiON oF $v$ RefEr To DifFeREnT framEs. HEnce choOsing a $q_{AB}$ (aNd wOrKing with $\GammA^{Q}$) iMplies breaKinG the ${\cal G}$ symmEtrY spontAnEouSLy to $\tExtRm{BMS}$ And the SUbleadIng soft grAvITons caN Be thougHT Of As goLdstone modes assocIAtED to this symmetrY breakInG As THEy mAp oNe family of VaCua (asSOciated TO a GIVEn $q_{ab}$) to a distinct fAmily of vacuA AssOciateD tO a dIFferenT $q_{AB}$.
EVaLUatIng $\delta_{V}N_{Ab}$ {#ssb1}
-----------------------------
from (\[deltav\]) and thE DefinitIOn of the News teNsoR (\[neWsdeF\]) We HaVe \_V n\_[Ab]{} & =& - \_U \_V C\_[ab]{}- \_V \^[(0)]{}\_[aB]{}\
& =& -Ł\_v \_[aB]{} - u \_[aB]{} + 2 (D\_A D\_B )\^ - \_V \^[(0)]{}\_[Ab]{}. \[dElvN2\] We nOw evALUATe thE laSt teRm in (\[dElVN2\]). From $\delta_v q_{Ab}= \L_Vq_{ab}- 2 \alPha q_{Ab}$ we haVe thAt $\Delta_v \rho_{AB}$ Is a suM oF a Lie derivative Term Plus a scalE trAnSfoRmAtion TErm. SinCe tHe bEhaviouR of $\rho_{Ab}$ UndEr SCALe Transformation is knOwN [@GErOch] the efFect of THe SeCOnd term cAn Be oBtaiNED explIcitLY. THe total cHange iS FoUnD to be: \_[V]{}\_[Ab]{}= Ł\_v \_[AB]{} + 2 D\_A D\_b, \[dElrHo\] wHich iN Turn ImplieS, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V \^[(0)]{}\_[AB]{} + 2 (d\_A D\_B )\^.\[dELrho0\] When substiTUting (\[delrho0\]) in (\[DElvn2\]) ThE ‘Soft’ FacTors cancel oUt anD One oBtaiNS: \_V n\_[AB]{} & =& - ł\_v \_[AB]{} - u \_[Ab]{} - Ł\_V \^[(0)]{}\_[AB]{}\
&=& ł\_V n\_[aB]{} + U \_[aB]{}, where in the second lInE we useD the dEfinition of thE news tensoR (\[NEWsdef\]) and The fACt THat $-\ddot{C}_{AB}= | $V\ \in\ \textrm{Lie}({\c al G})$ ma ps acon fig ur atio n wi th zero news t o a c onfiguration with zero news b e caus e t he de finitio n o f the n ew s b ef o re andaft er theaction of$V$ r efer to diff e re nt frames. He nce choosing a$q_{AB }$ (a n d wor kin g wit h $\Ga m ma^{q} $) implie sb reakin g the ${ \ c al G}$ symmetry spontan e ou s ly to $\textrm {BMS}$ a n dt h e s ubl eading sof tgravi t ons can be t h oug h t of as golds tone modesa sso ciated t o t h is sym metry b r eak ing as they map one fami ly ofv acua (a s sociate d to a gi ven $q_ { AB }$ ) t oa di s ti nct fam ily of v ac ua asso ciat e d t o adif fere nt $q _{AB}$.
Eval uat ing$ \de lta_{ V}N_{ AB}$ { #ssb1 }
---- ----- -- --------------- ---
From (\[ del ta V\] )and t h e defi nit ion of the news t e nso r( \ [ ne wsdef\]) we have \ _V N \_ [AB]{} & =& -\ _u \ _ V C\_[AB ]{ }-\_V\ ^ [(0)] {}\_ [ AB ]{}\
& = & -Ł\_ V \ _[ AB]{} - u \_[AB ]{ } + 2(D\_A D\_B )\^ - \_V \^[ (0)]{ } \_[AB]{}. \[de l VN2\] We nowe va l u at e the la st term in(\[d e lVN2 \]). Fr om$ \delt a_V q _{ A B} = \L_Vq_{AB}- 2 \alp ha q_{AB }$ we have that $\ delta_V \r h o _ {AB}$ is a s u mo f a Lie deriva tiveterm plusa scale t ransf ormation term. Si n c e the be hav iou r o f $ \ r ho _{AB}$ unders c aletr ansform ati on is k now n [ @ge roc h] the effe ct of th ese co nd te rm ca n be obta in edex pli citly . The t otalchan ge i s fo und tob e: \ _[V] {} \_ [AB] {}= Ł \_V \ _[AB ] {}+ 2 D\_ A D\_B, \ [de l rho\ ]wh ich inturn implies, \ _[V]{}\^[( 0) ]{} \_[AB] { } = Ł\_V \ ^[(0)]{}\_[AB]{} + 2 (D \ _A D\_B )\ ^.\[d elrh o0\] When su bstitu tin g (\[de lrho0\ ]) in ( \[d e l VN2\] ) th e ‘ so ft’ factor s can cel o ut and one ob tains: \_V N\_[AB] { } & =& - Ł\_V \_ [AB ]{}- u\_[ A B] { } - Ł \ _V\ ^ [(0)]{}\_[AB]{} \
&=& Ł\_V N \ _[ AB]{} + u\ _[A B] {}, whe re in t he se c ond lin e we used the defi ni tion o f t he news te nsor (\[ newsdef\] ) andt he fact th at $-\ dd ot{ C}_{A B}= | $V\_\in\ \textrm{Lie}({\cal_G})$ maps a configuration_with zero_news_to a_configuration_with zero news_because the definition_of the news before_and after the_action_of $V$ refer to different frames. Hence choosing a $q_{AB}$ (and working with $\Gamma^{q}$)_implies_breaking the_${\cal_G}$_symmetry spontaneously to $\textrm{BMS}$ and_the subleading soft gravitons can_be thought_of as goldstone modes associated to this symmetry_breaking_as they map_one family of vacua (associated to a given $q_{AB}$)_to a distinct family of vacua_associated to a_different_$q_{AB}$.
Evaluating_$\delta_{V}N_{AB}$ {#ssb1}
-----------------------------
From (\[deltaV\]) and_the definition of the news tensor_(\[newsdef\]) we have \_V N\_[AB]{} &_=& - \_u \_V C\_[AB]{}- \_V \^[(0)]{}\_[AB]{}\
&_=& -Ł\_V \_[AB]{} - u \_[AB]{}_+ 2 (D\_A D\_B )\^_- \_V_\^[(0)]{}\_[AB]{}. \[delVN2\] We now evaluate_the last term_in (\[delVN2\])._From $\delta_V q_{AB}=_\L_Vq_{AB}- 2 \alpha q_{AB}$ we have_that $\delta_V \rho_{AB}$_is a sum of a Lie_derivative_term plus a_scale_transformation_term. Since_the behaviour of_$\rho_{AB}$_under scale_transformation_is known [@geroch] the effect of_the_second term can be obtained explicitly. The_total change is found_to_be: \_[V]{}\_[AB]{}= Ł\_V \_[AB]{}_+ 2 D\_A D\_B, \[delrho\]_which in turn implies, \_[V]{}\^[(0)]{}\_[AB]{}= Ł\_V_\^[(0)]{}\_[AB]{} +_2 (D\_A_D\_B )\^.\[delrho0\] When substituting (\[delrho0\]) in (\[delVN2\]) the ‘soft’ factors cancel_out and one obtains: \_V N\_[AB]{}_& =& - Ł\_V_\_[AB]{} -_u_\_[AB]{} - Ł\_V_\^[(0)]{}\_[AB]{}\
&=&_Ł\_V N\_[AB]{}_+ u \_[AB]{}, where in the second_line we_used the definition of the news_tensor (\[newsdef\]) and the_fact_that $-\ddot{C}_{AB}= |
}^{\infty} KL_{\alpha}[f](\tau)
\frac{\left|
\Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
\Gamma\left(\beta+\frac{i(\tau-\mu)}{2}\right)
\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)\right|^{2}}
{ \left| \Gamma\left(i\tau\right)\right|^{2}}
d\tau \.
\end{array}$$
Considering in Proposition \[prop: Parseval\] the exponential function instead of the modified Bessel function, we can as well find a Parseval-like relation also valid when one of the integrands is a polynomial.
Upon the choice of $f(x)=x^{n}$, $n\geqslant 0$, in, we obtain, while yields $$\label{proof Parseval epsilon exp rel2}
\int_{0}^{\infty} x^{\alpha+\beta+n} {{\rm e}}^{-x} dx
=\frac{1}{4\pi} \int_{0}^{\infty} KL_{\alpha}[x^{n}](\tau) \frac{
\left|\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)
\Gamma\left(\beta+\frac{i\tau}{2}\right)\right|^{2}}
{\left|\Gamma\left( i\tau \right)\right|^{2}}
W_{ -\beta+\frac{1}{2},\frac{i\tau}{2}}(1)
d\tau \ ,\ n\geqslant 0.$$
Both equalities can be read as integral relations between the Gamma functions, precisely can be rewritten as $$\begin{aligned}
&& \frac{1}{2}\left| \Gamma\left( \alpha+\beta+n + \frac{i\mu}{2}\right)
\Gamma\left( \alpha+\beta + \frac{i\mu}{2}\right)\right|^{2} \\
&& = \frac{1}{8\pi \Gamma(2\beta)} \int_{0}^{\infty}
{\left|
\Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
\ | } ^{\infty } KL_{\alpha}[f](\tau)
\frac{\left|
\Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
\Gamma\left(\beta+\frac{i(\tau-\mu)}{2}\right)
\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)\right|^{2 } }
{ \left| \Gamma\left(i\tau\right)\right|^{2 } }
d\tau \.
\end{array}$$
Considering in Proposition \[prop: Parseval\ ] the exponential function instead of the modified Bessel affair, we can equally well find a Parseval - like relation back besides valid when one of the integrands is a polynomial.
Upon the choice of $ f(x)=x^{n}$, $ n\geqslant 0 $, in, we obtain, while yields $ $ \label{proof Parseval epsilon exp rel2 }
\int_{0}^{\infty } x^{\alpha+\beta+n } { { \rm e}}^{-x } dx
= \frac{1}{4\pi } \int_{0}^{\infty } KL_{\alpha}[x^{n}](\tau) \frac {
\left|\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)
\Gamma\left(\beta+\frac{i\tau}{2}\right)\right|^{2 } }
{ \left|\Gamma\left (i\tau \right)\right|^{2 } }
W _ { -\beta+\frac{1}{2},\frac{i\tau}{2}}(1)
d\tau \, \ n\geqslant 0.$$
Both equality can be read as integral relation between the Gamma functions, precisely can be rewritten as $ $ \begin{aligned }
& & \frac{1}{2}\left| \Gamma\left (\alpha+\beta+n + \frac{i\mu}{2}\right)
\Gamma\left (\alpha+\beta + \frac{i\mu}{2}\right)\right|^{2 } \\
& & = \frac{1}{8\pi \Gamma(2\beta) } \int_{0}^{\infty }
{ \left|
\Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
\ | }^{\infhy} KL_{\alpha}[f](\tau)
\fvac{\left|
\Yqmma\leht(\beta+\fdac{i(\tau+\mj)}{2}\right)
\Gamma\left(\beta+\fcac{i(\rau-\mu)}{2}\eight)
\Gamma\lewt(\alpha+1+\frwc{i\tau}{2}\ritht)\rmght|^{2}}
{ \lefv| \Gamma\lcyt(i\tah\vight)\xijht|^{2}}
d\tau \.
\end{array}$$
Cmnsidering in [ruplsition \[prop: Parseval\] the exponentiwl funcyiln instead of jhe mpqifisd Bessel function, we can as well rind a Karseval-like relatoon also valid when one of the integrands is a pllynomial.
Upin trw choice of $w(x)=x^{n}$, $n\geqslant 0$, in, we kbtain, while yields $$\label{proof Oarseral epsilon ezp ggl2}
\int_{0}^{\inhty} x^{\ajpha+\beta+n} {{\rm e}}^{-x} dx
=\frac{1}{4\pi} \int_{0}^{\infby} KN_{\alpha}[x^{n}](\tau) \frac{
\lefv|\Gamma\left(\alpha+1+\frac{i\jau}{2}\right)
\Gamma\left(\beta+\frqc{u\tau}{2}\rhght)\sighg|^{2}}
{\lert|\Jamja\left( i\teu \right)\riggt|^{2}}
W_{ -\bwta+\frac{1}{2},\frac{i\tau}{2}}(1)
d\equ \ ,\ n\geqslaht 0.$$
Botr qqualities can be read as integral relauions between the Gamma funcrions, precisely can bg rewritteg as $$\begin{aligned}
&& \frac{1}{2}\left| \Gamma\left( \alpha+\betd+n + \fcaz{i\mb}{2}\vlght)
\Gamma\left( \alpha+\beta + \frac{i\mu}{2}\right)\right|^{2} \\
&& = \ftab{1}{8\pi \Gamma(2\beta)} \int_{0}^{\lnfty}
{\left|
\Gsima\left(\beta+\frxc{i(\tau+\mu)}{2}\rjght)
\ | }^{\infty} KL_{\alpha}[f](\tau) \frac{\left| \Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right) \Gamma\left(\beta+\frac{i(\tau-\mu)}{2}\right) \Gamma\left(\alpha+1+\frac{i\tau}{2}\right)\right|^{2}} { d\tau \end{array}$$ Considering Proposition \[prop: Parseval\] the Bessel function, we as well find Parseval-like relation also valid when one the integrands is a polynomial. Upon the choice of $f(x)=x^{n}$, $n\geqslant 0$, in, obtain, while yields $$\label{proof Parseval epsilon exp rel2} \int_{0}^{\infty} x^{\alpha+\beta+n} {{\rm e}}^{-x} dx \int_{0}^{\infty} \frac{ \Gamma\left(\beta+\frac{i\tau}{2}\right)\right|^{2}} i\tau \right)\right|^{2}} W_{ -\beta+\frac{1}{2},\frac{i\tau}{2}}(1) d\tau \ ,\ n\geqslant 0.$$ Both equalities can be read as integral between the Gamma functions, precisely can be rewritten $$\begin{aligned} && \frac{1}{2}\left| \Gamma\left( + \frac{i\mu}{2}\right) \Gamma\left( \alpha+\beta + \\ = \frac{1}{8\pi \int_{0}^{\infty} \Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right) | }^{\infty} KL_{\alpha}[f](\tau)
\frac{\left|
\GAmma\left(\beTa+\fraC{i(\tAu+\mU)}{2}\rIght)
\gammA\left(\beta+\frac{i(\TAu-\mu)}{2}\Right)
\Gamma\left(\alpha+1+\fraC{i\tau}{2}\RiGHt)\riGHt|^{2}}
{ \Left| \GAmma\lefT(I\tAU\RigHt)\RiGht|^{2}}
D\tAU \.
\eNd{arrAy}$$
COnsiderIng in PropoSitIoN \[prop: ParsevaL\] ThE exponentiAl fUnction insteAd oF the moDiFieD besseL fuNctioN, we can AS well fInd a ParseVaL-Like reLAtion alSO VaLid wHen one of the integrANdS Is a polynomial.
UPon the ChOIcE OF $f(x)=X^{n}$, $n\Geqslant 0$, in, We ObtaiN, While yiELdS $$\LABel{PRoof Parseval ePsilon exp reL2}
\Int_{0}^{\Infty} x^{\AlPha+\BEta+n} {{\rm E}}^{-x} dx
=\fRaC{1}{4\Pi} \iNt_{0}^{\infty} KL_{\alPha}[x^{N}](\tau) \frac{
\lEft|\GamMA\left(\alPHa+1+\frac{i\Tau}{2}\rigHt)
\GAmmA\lefT(\BeTa+\FraC{i\TAu}{2}\rIGhT)\riGHt|^{2}}
{\lEft|\Gamma\LeFt( I\tau \rIght)\RIGHT|^{2}}
W_{ -\beTa+\fRac{1}{2},\fRac{i\tAu}{2}}(1)
d\tau \ ,\ n\geqslaNt 0.$$
BOth eQUalIties Can be Read As IntegRal relAtionS bEtween the Gamma fUnctIons, preciSelY cAn bE rEwritTEn as $$\beGin{AliGned}
&& \fraC{1}{2}\left| \GaMMa\lEfT( \ALPhA+\beta+n + \frac{i\mu}{2}\right)
\gaMMA\lEft( \alpha+\Beta + \frAC{i\Mu}{2}\RIght)\righT|^{2} \\
&& = \fRac{1}{8\Pi \GaMMA(2\beta)} \Int_{0}^{\iNFtY}
{\left|
\GamMa\left(\BEtA+\fRac{i(\tau+\Mu)}{2}\Right)
\ | }^{\infty} KL_{\alpha}[f]( \tau)
\fr ac{ \l eft|
\Gamma\l e ft(\ beta+\frac{i(\tau+\mu) }{2}\ ri g ht) \Gamma\ l ef t ( \be ta +\ fra c{ i (\ tau-\ mu) }{2}\ri ght)
\Gamma\left( \ al pha+1+\fra c{i \tau}{2}\rig ht) \right |^ {2} }
{ \l eft| \ G amma\l eft(i\tau \r i ght)\r i ght|^{2 } }
d\tau \.
\end{ a rr a y}$$
Consider ing in P r op o s iti on\[prop: Pa rs eval\ ] the ex p on e n t ial function inst ead of them odi fied B es sel functi on, w ec anas well fin d aParseval- like r e lationa lso val id whe n o neof t h ein teg ra n dsi sa p o lyn omial.
Up on thechoi c e o f $f (x) =x^{ n}$,$n\geqslant 0 $,in,w e o btain , whi le y ie lds $ $\labe l{pro of Parseval epsil on e xp rel2} \in t_ {0}^{ \ infty} x^ {\a lpha+\b eta+n}{{\ rm e } }^ {-x} dx
= \f r a c{ 1}{4\pi} \int_ { 0} ^{ \ infty} KL _{\ alph a } [x^{n }](\ t au ) \frac {
\ le ft |\Gamma \l eft(\a lp ha+ 1+\ frac{ i \tau }{2}\r ight) \ G amma\left(\bet a +\frac{i\tau} { 2} \ r ig h t)\r igh t|^{2}}
{\l e ft|\ Gamm a \l eft ( i\ta u \ri gh t )\ r ight|^{2}}
W_ {-\beta +\fra c{1}{2},\frac {i\tau}{2} } ( 1 )
d \tau \, \ n\geqslant0.$$
Both equa l ities ca n beread asintegralr e lationsbet wee n t heG a mm a functions,p r ecis el y can b e r ewritte n a s $ $\b egi n{ aligned} && \ fr ac {1 }{ 2}\ left| \Gamma\l ef t(\a lph a+\be t a+n +\frac {i\m u} {2 } \ri ght)
\ G amma \l ef t( \ alp ha +\bet a +\ fra c{i\mu} {2}\right )\r i ght| ^{ 2} \\
&& = \frac{1 }{ 8\pi \Gamm a( 2\b eta)}\ i nt_{0}^{ \infty}
{\left| \Ga mma \left (\be ta+\frac{ i(\ tau+\m u)} { 2}\rig ht)
\ | }^{\infty} KL_{\alpha}[f](\tau)
_ _ _ _\frac{\left|_
__ _ \Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
_ _ _\Gamma\left(\beta+\frac{i(\tau-\mu)}{2}\right)_
\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)\right|^{2}}
__ _{_\left|_\Gamma\left(i\tau\right)\right|^{2}}
_ d\tau \.
\end{array}$$
Considering_in Proposition_\[prop: Parseval\] the exponential function instead of the_modified_Bessel function, we_can as well find a Parseval-like relation also valid_when one of the integrands is_a polynomial.
Upon the_choice_of_$f(x)=x^{n}$, $n\geqslant 0$, in,_we obtain, while yields $$\label{proof Parseval_epsilon exp rel2}
_ \int_{0}^{\infty} x^{\alpha+\beta+n} {{\rm e}}^{-x} dx_
_ =\frac{1}{4\pi} \int_{0}^{\infty} _KL_{\alpha}[x^{n}](\tau) _\frac{
\left|\Gamma\left(\alpha+1+\frac{i\tau}{2}\right)_
_ _\Gamma\left(\beta+\frac{i\tau}{2}\right)\right|^{2}}
_ {\left|\Gamma\left( i\tau \right)\right|^{2}}
_ _W_{ -\beta+\frac{1}{2},\frac{i\tau}{2}}(1)
_d\tau_\ ,\ _n\geqslant_0.$$
Both_equalities can_be read as_integral_relations between_the_Gamma functions, precisely can be rewritten_as_$$\begin{aligned}
&& \frac{1}{2}\left| \Gamma\left(_\alpha+\beta+n + \frac{i\mu}{2}\right)
__ \Gamma\left( \alpha+\beta_+ \frac{i\mu}{2}\right)\right|^{2} \\
_ && = \frac{1}{8\pi \Gamma(2\beta)} \int_{0}^{\infty}_
_ _ {\left|
\Gamma\left(\beta+\frac{i(\tau+\mu)}{2}\right)
_\ |
$ is then at most $\lambda_a-2$. In the latter case, $ab$ anticommutes with itself, meaning that $a$ and $b$ have the same charge. This violates (iii).
It remains to show that the boundary operators coincide. Consider the $i$th term $\epsilon_i = (-1)^i\cdot x_0y_0 \wedge \cdots \wedge \hat{x_iy_i}
\wedge x_dy_d$ in the boundary of $\hat{\gamma}$ in $\mathcal{C}({\mathsf{M}_{N}};R)$. Write $\delta_i = [\epsilon_i]$. If $a_i \neq b_i$, then let $(z_1 = x_i, z_2, \ldots, z_q)$ be the sequence of elements appearing in $U_{a_i} \cap \{x_j,y_j : i \le j \le d\}$ and let $(z'_1 = y_i, z'_2, \ldots, z'_{q'})$ be the sequence of elements appearing in $U_{b_i} \cap \{x_j,y_j : i \le j \le d\}$, arranged in increasing order. To obtain the lexicographically smallest element $\hat{\delta_i}$ from $\epsilon_i$, we need to act on $\epsilon_i$ with the permutation $(z_q, \cdots, z_2, z_1)(z'_{q'}, \cdots, z'_2,
z'_1)$. There are three cases.
- $a_i$ and $b_i$ are both negatively charged. By construction, $a_ib_i$ appears only once in $\gamma$, meaning that there are $q+q'-2$ edges $a_jb_j$ commuting with $a_ib_i$ such that $j>i$; hence $\eta_i = d-i-(q+q'-2)$, where $\eta_i$ is the number of indices $j > i$ such that $a_ib_i$ and $a_jb_j$ anticommute. We conclude that the sign of $\hat{\delta_i}$ equals $(-1)^{i+q+q'-2} = (-1)^{d-\eta_i}$, which aligns with (\[partial-eq\]).
| $ is then at most $ \lambda_a-2$. In the latter case, $ ab$ anticommutes with itself, meaning that $ a$ and $ b$ have the like cathexis. This violates (iii).
It remains to picture that the boundary operator coincide. Consider the $ i$th terminus $ \epsilon_i = (-1)^i\cdot x_0y_0 \wedge \cdots \wedge \hat{x_iy_i }
\wedge x_dy_d$ in the limit of $ \hat{\gamma}$ in $ \mathcal{C}({\mathsf{M}_{N}};R)$. Write $ \delta_i = [ \epsilon_i]$. If $ a_i \neq b_i$, then let $ (z_1 = x_i, z_2, \ldots, z_q)$ be the sequence of elements appearing in $ U_{a_i } \cap \{x_j, y_j: i \le j \le d\}$ and lease $ (z'_1 = y_i, z'_2, \ldots, z'_{q'})$ be the sequence of elements appear in $ U_{b_i } \cap \{x_j, y_j: i \le j \le d\}$, arranged in increasing order. To obtain the lexicographically smallest component $ \hat{\delta_i}$ from $ \epsilon_i$, we need to act on $ \epsilon_i$ with the permutation $ (z_q, \cdots, z_2, z_1)(z'_{q' }, \cdots, z'_2,
z'_1)$. There exist three cases.
- $ a_i$ and $ b_i$ are both negatively charged. By construction, $ a_ib_i$ appears merely once in $ \gamma$, meaning that there are $ q+q'-2 $ edges $ a_jb_j$ commuting with $ a_ib_i$ such that $ j > i$; hence $ \eta_i = vitamin d - i-(q+q'-2)$, where $ \eta_i$ is the number of indices $ j > i$ such that $ a_ib_i$ and $ a_jb_j$ anticommute. We conclude that the sign of $ \hat{\delta_i}$ equals $ (-1)^{i+q+q'-2 } = (-1)^{d-\eta_i}$, which aligns with (\[partial - eq\ ]). | $ is then at most $\lambda_a-2$. In the latter casg, $qb$ antmcommutss with ktself, meaning that $a$ and $b$ iave the wame charge. This violages (iii).
It remains to wyow that tis boundary opsvatorv coincide. Conslder the $i$tv term $\epsilon_h = (-1)^i\edot x_0y_0 \wedge \cdots \wedge \hat{x_iy_i}
\wedge c_dj_d$ in the bounqary jf $\hzn{\gcmma}$ in $\mathcal{C}({\mathsf{M}_{N}};R)$. Write $\delta_i = [\epsilon_i]$. If $a_i \neq b_i$, then let $(z_1 = x_i, x_2, \ldlts, z_q)$ be the sequfnce of elenentf appearing iv $U_{a_i} \cap \{q_l,y_j : i \le j \le d\}$ and let $(z'_1 = y_i, z'_2, \ldots, z'_{q'})$ ce thz sequence if elfkents appeacing ig $U_{b_i} \cap \{x_j,n_k : i \la j \le c\}$, arranged in lncreesint order. To obtain the lexicographically smallest emement $\hat{\delta_i}$ drim $\epvilot_i$, wd nedd uo ect on $\epdilkn_i$ with tge permutatuon $(z_q, \cdots, z_2, z_1)(z'_{q'}, \vdjnx, z'_2,
z'_1)$. Thede are trree cases.
- $a_i$ and $b_i$ are both negativtly cgarged. By construction, $q_ib_i$ appears only oncg in $\gamma$, meaning that there are $q+q'-2$ edges $a_jb_j$ commuting whth $a_mb_k$ sbgm thxr $u>i$; hence $\eta_i = d-i-(q+q'-2)$, where $\eta_i$ is the number jr oncices $j > i$ sucm that $a_ib_i$ and $a_jn_j$ amjicommute. We cuncludz tgat the sign of $\hah{\delta_i}$ equaos $(-1)^{i+q+q'-2} = (-1)^{d-\tta_i}$, ehich aligns with (\[partial-eq\]).
| $ is then at most $\lambda_a-2$. In case, anticommutes with meaning that $a$ charge. violates (iii). It to show that boundary operators coincide. Consider the $i$th $\epsilon_i = (-1)^i\cdot x_0y_0 \wedge \cdots \wedge \hat{x_iy_i} \wedge x_dy_d$ in the boundary $\hat{\gamma}$ in $\mathcal{C}({\mathsf{M}_{N}};R)$. Write $\delta_i = [\epsilon_i]$. If $a_i \neq b_i$, then let = z_2, z_q)$ the sequence of elements appearing in $U_{a_i} \cap \{x_j,y_j : i \le j \le d\}$ and $(z'_1 = y_i, z'_2, \ldots, z'_{q'})$ be the of elements appearing in \cap \{x_j,y_j : i \le \le arranged in order. obtain lexicographically smallest element from $\epsilon_i$, we need to act on $\epsilon_i$ with the permutation $(z_q, \cdots, z_2, z_1)(z'_{q'}, \cdots, z'_2, There are - $a_i$ $b_i$ both charged. By construction, only once in $\gamma$, meaning that edges $a_jb_j$ commuting with $a_ib_i$ such that $j>i$; $\eta_i = where $\eta_i$ is the number of $j > i$ such that $a_ib_i$ and $a_jb_j$ We conclude that the sign of $\hat{\delta_i}$ equals $(-1)^{i+q+q'-2} = (-1)^{d-\eta_i}$, which aligns with (\[partial-eq\]). | $ is then at most $\lambda_a-2$. In the lAtter case, $aB$ antiComMutEs With ItseLf, meaning that $a$ ANd $b$ hAve the same charge. This viOlateS (iII).
It rEMaIns to Show thaT ThE BOunDaRy OpeRaTOrS coinCidE. ConsidEr the $i$th teRm $\ePsIlon_i = (-1)^i\cdot x_0y_0 \WEdGe \cdots \wedGe \hAt{x_iy_i}
\wedge x_Dy_d$ In the bOuNdaRY of $\haT{\gaMma}$ in $\MathcaL{c}({\mathsF{M}_{N}};R)$. Write $\DeLTa_i = [\epsILon_i]$. If $a_I \NEq B_i$, thEn let $(z_1 = x_i, z_2, \ldots, z_q)$ bE ThE Sequence of elemEnts apPeARiNG In $U_{A_i} \cAp \{x_j,y_j : i \le j \Le D\}$ and lET $(z'_1 = y_i, z'_2, \ldOTs, Z'_{Q'})$ BE thE Sequence of eleMents appearINg iN $U_{b_i} \caP \{x_J,y_j : I \Le j \le d\}$, ArranGeD In iNcreasing orDer. TO obtain thE lexicOGraphicALly smalLest elEmeNt $\hAt{\deLTa_I}$ fRom $\EpSIloN_I$, wE neED to Act on $\epsIlOn_I$ with The pERMUTatiOn $(z_Q, \cdoTs, z_2, z_1)(z'_{Q'}, \cdots, z'_2,
z'_1)$. There Are ThreE CasEs.
- $a_i$ aNd $b_i$ aRe boTh NegatIvely cHargeD. BY construction, $a_iB_i$ apPears only OncE iN $\gaMmA$, meanINg that TheRe aRe $q+q'-2$ edgEs $a_jb_j$ cOMmuTiNG WItH $a_ib_i$ such that $j>i$; henCe $\ETA_i = D-i-(q+q'-2)$, wherE $\eta_i$ iS ThE nUMber of inDiCes $J > i$ suCH That $a_Ib_i$ aND $a_Jb_j$ anticOmmute. wE cOnClude thAt The sigN oF $\haT{\deLta_i}$ eQUals $(-1)^{I+q+q'-2} = (-1)^{d-\etA_i}$, which aLigns WIth (\[partial-eq\]).
| $ is then at most $\lambda _a-2$. Inthe l att erca se,$ab$ anticommutesw ithitself, meaning that $ a$ an d$ b$ h a ve thesame ch a rg e . Th is v iol at e s(iii) .
It rema ins to sho w t ha t the bounda r yoperatorscoi ncide. Consi der the $ i$ tht erm $ \ep silon _i = ( - 1)^i\c dot x_0y_ 0\ wedge\ cdots \ w e dg e \h at{x_iy_i}
\w e dg e x_dy_d$ in th e boun da r yo f $\ hat {\gamma}$in $\ma t hcal{C} ( {\ m a t hsf { M}_{N}};R)$.Write $\del t a_i = [\e ps ilo n _i]$.If $a _i \ne q b_i$, the n le t $(z_1 = x_i,z _2, \ld o ts, z_q )$ bethe se quen c eof el em e nts ap pea r ing in $U_{ a_ i} \cap \{x _ j , y _j : i\lej \le d\}$ and let $( z'_1 = y _i, z '_2,\ldo ts , z'_ {q'})$ be t he sequence of el emen ts appear ing i n $ U_ {b_i} \cap \ {x_ j,y _j : i\le j \ l e d \} $ , ar ranged in increasi ng o rd er. To o btaint he l e xicograp hi cal ly s m a llest ele m en t $\hat{ \delta _ i} $from $\ ep silon_ i$ , w e n eed t o act on $\ epsilon_ i$ wi t h the permutat i on $(z_q, \cd o ts , z_ 2 , z_ 1)( z'_{q'}, \c dots , z'_ 2,
z'_ 1 )$. T herear e t h ree cases.
- $a_ i$ and $ b_i$are both nega tively cha r g e d. By co nstr u ct i on, $a_ib_i$ a ppear s only onc e in $\ga mma$, meaning that the r e are $q+ q'- 2$edg es$ a _j b_j$ commutin g with $ a_ib_i$ su ch that $j >i$ ; h enc e$\eta_i = d-i-(q+ q' -2 )$ ,whe re $\ e ta_i$ is t henu mbe r ofi ndices $j > i$su ch tha t $a_ib _ i$ a nd $ a_ jb _j$ant ic ommut e. W e co ncludethat thesig n of$\ ha t{\delt a_i}$ equals$( -1)^{i+q+q '- 2}= (-1) ^ { d-\eta_i }$, which aligns with ( \ [partia l-e q\]).
| $ is_then at_most $\lambda_a-2$. In the_latter case,_$ab$_anticommutes with_itself,_meaning that $a$_and $b$ have_the same charge. This_violates (iii).
It remains_to_show that the boundary operators coincide. Consider the $i$th term $\epsilon_i = (-1)^i\cdot x_0y_0_\wedge_\cdots \wedge_\hat{x_iy_i}
__ \wedge x_dy_d$ in_the boundary of $\hat{\gamma}$ in_$\mathcal{C}({\mathsf{M}_{N}};R)$. Write_$\delta_i = [\epsilon_i]$. If $a_i \neq b_i$, then_let_$(z_1 = x_i,_z_2, \ldots, z_q)$ be the sequence of elements appearing_in $U_{a_i} \cap \{x_j,y_j : i_\le j \le_d\}$_and_let $(z'_1 = y_i,_z'_2, \ldots, z'_{q'})$ be the sequence_of elements appearing in $U_{b_i} \cap_\{x_j,y_j : i \le j \le d\}$,_arranged in increasing order. To obtain_the lexicographically smallest element $\hat{\delta_i}$_from $\epsilon_i$,_we need to act on_$\epsilon_i$ with the_permutation $(z_q,_\cdots, z_2, z_1)(z'_{q'},_\cdots, z'_2,
z'_1)$._There are three_cases.
- $a_i$ and $b_i$_are_both negatively charged._By_construction,_$a_ib_i$ appears_only once in_$\gamma$,_meaning that_there_are $q+q'-2$ edges $a_jb_j$ commuting with_$a_ib_i$_such that $j>i$; hence $\eta_i = d-i-(q+q'-2)$,_where $\eta_i$ is the_number_of indices $j >_i$ such that $a_ib_i$ and_$a_jb_j$ anticommute. We conclude that the_sign of_$\hat{\delta_i}$ equals_$(-1)^{i+q+q'-2} = (-1)^{d-\eta_i}$, which aligns with (\[partial-eq\]).
|
eq:o2}
\tilde V_1(\tau) = \tilde V \, \delta(\tau-1), \qquad {\rm where}
\qquad
\tilde V = \frac{1}{2}\ \eta\ r^2,$$
where the parameter $\eta$ has been defined previously. All the information about the dynamics of the system is contained within this Floquet operator, since $\Phi(r,\phi,\tau+1) = U\,
\Phi(r,\phi,\tau)$, where $\Phi$ is the total wave function. Because of the periodic $\delta$-kicked dynamics, we can immediately integrate equation (\[eq:o1\]) to get $$\label{eq:p}
U_\ell(1,0) = \exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde V\right)\,
\exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde H_{dc}\right).$$ The subscript $\ell$ has been attached to $U$ to emphasize that the evolution operator has been restricted to that single $\ell$ value. In other words, states with different values of $\ell$ evolve independently, an immediate consequence of the conservation of angular momentum in this system. The rightmost exponential operator in equation (\[eq:p\]) evolves the wave function from just after the ‘kick’ at $\tau=0$ to just before the kick at one period under the influence of $\tilde H_{dc}$, while the operator to it’s left propagates it from just before to just after the kick at a period.
Since $U$ is an Unitary operator, the spectrum of it’s eigenvalues can be represented as $$\label{eq:q}
U_\ell\,\phi_{n\ell} = e^{i\,\varepsilon_{n\ell}}\,\phi_{n\ell}.$$ The set of eigenvalues $\{\varepsilon_{n\ell} \in (0,2\pi]\}$, are collectively known as the Quasi-energy Eigenvalues (QEE), and the eigenfunctions $\{\phi_{n\ell}\}$ as the Quasi-energy Eigenfunctions (QEF) of $U$. The investigation of the quantum dynamics of the system is completely equivalent to determining the nature of the QEE and QEF. The fundamental task is thus to obtain the Quasi-energy Spectrum (QES) of the evolution operator given by equation (\[eq:p\]).
Classical Dynamics | eq: o2 }
\tilde V_1(\tau) = \tilde V \, \delta(\tau-1), \qquad { \rm where }
\qquad
\tilde V = \frac{1}{2}\ \eta\ r^2,$$
where the parameter $ \eta$ has been defined previously. All the information about the dynamics of the arrangement is control within this Floquet operator, since $ \Phi(r,\phi,\tau+1) = U\,
\Phi(r,\phi,\tau)$, where $ \Phi$ is the total wave function. Because of the periodic $ \delta$-kicked moral force, we can immediately integrate equation (\[eq: o1\ ]) to get $ $ \label{eq :p }
U_\ell(1,0) = \exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde V\right)\,
\exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde H_{dc}\right).$$ The subscript $ \ell$ has been impound to $ U$ to emphasize that the evolution hustler has been restrict to that single $ \ell$ value. In early words, states with different values of $ \ell$ develop independently, an immediate consequence of the conservation of angular momentum in this system. The rightmost exponential hustler in equation (\[eq: p\ ]) develop the wave affair from just after the ‘ kick ’ at $ \tau=0 $ to just before the gripe at one period under the influence of $ \tilde H_{dc}$, while the operator to it ’s left propagate it from just before to just after the kick at a period.
Since $ U$ is an Unitary hustler, the spectrum of it ’s eigenvalues can be represented as $ $ \label{eq: q }
U_\ell\,\phi_{n\ell } = e^{i\,\varepsilon_{n\ell}}\,\phi_{n\ell}.$$ The set of eigenvalues $ \{\varepsilon_{n\ell } \in (0,2\pi]\}$, are collectively known as the Quasi - department of energy Eigenvalues (QEE), and the eigenfunctions $ \{\phi_{n\ell}\}$ as the Quasi - department of energy Eigenfunctions (QEF) of $ U$. The probe of the quantum dynamics of the system is completely equivalent to determining the nature of the QEE and QEF. The fundamental task is thus to receive the Quasi - department of energy Spectrum (QES) of the development operator given by equation (\[eq: p\ ]).
Classical Dynamics | eq:o2}
\hilde V_1(\tau) = \tilde V \, \delua(\tau-1), \qquad {\rm whgrw}
\qquav
\tilde V = \frac{1}{2}\ \eta\ r^2,$$
where the parameter $\ete$ haw beeb defined previously. Aul the invormatiob abiyt the dynejics of the sgdtem ms contained wijhin this Flmquet operator, sknee $\Phi(r,\phi,\tau+1) = U\,
\Phi(r,\phi,\tau)$, where $\Phi$ is the tltal wave funcjion. Nqcauav if the periodic $\delta$-kicked dgnamics, we can immedoately integrate equation (\[fq:o1\]) ho get $$\label{eq:p}
U_\elp(1,0) = \exp\left(-\ftzc{i}{{\eulde\hbar}}\,\tildd V\right)\,
\exk\lzft(-\frac{i}{{\tilse\hbar}}\,\tilde H_{dc}\right).$$ The subscrkpt $\ekl$ has beeb qttwwhed to $U$ ti emprasize that bne evonution pperator has bcen rxstructed to that single $\xll$ value. In other wjrds, statas with different vqlyes ox $\eln$ evuove inse'ensently, an immediate consequencw of the conservatipn if angular mojentum ig this system. The rightmost exponential opsrator in equation (\[eq:p\]) wvolves the wave funcjion from tust after the ‘kick’ at $\tau=0$ to just before the kicn at kve kevlod jbdfr the influence of $\tilde H_{dc}$, while the operaekr tp it’s left proiagates it from juxt brsore to just xfter cge kick at a period.
Slnce $U$ ys an Unitary jperstor, the spectrum of it’s eitenvalues cau bw represented as $$\lcbel{eq:q}
U_\ell\,\pki_{n\ell} = e^{i\,\vsrepsilon_{n\ell}}\,\phi_{n\ell}.$$ Thz set kf eigenvalkes $\{\varepaklon_{n\ell} \in (0,2\pi]\}$, afe bollactively known as the Quasy-energy Emgenvclues (QED), anc the qigenfunctlons $\{\ivi_{n\ell}\}$ as the Quasl-eneryy Eicenfunctiojs (QEF) of $U$. The investigation oh the quantum dfnakics of che syxtem is complqtely equivaleut to decerminkng the nanure of tie QEE and QQF. The fundamajtal task is thus to obtqin rhe Quark-energy Spectrim (QES) of the evolurion operator givek by gqhation (\[eq:p\]).
Classneao Dynamics | eq:o2} \tilde V_1(\tau) = \tilde V \, {\rm \qquad \tilde = \frac{1}{2}\ \eta\ has defined previously. All information about the of the system is contained within Floquet operator, since $\Phi(r,\phi,\tau+1) = U\, \Phi(r,\phi,\tau)$, where $\Phi$ is the total wave Because of the periodic $\delta$-kicked dynamics, we can immediately integrate equation (\[eq:o1\]) to $$\label{eq:p} = V\right)\, H_{dc}\right).$$ The subscript $\ell$ has been attached to $U$ to emphasize that the evolution operator has restricted to that single $\ell$ value. In other states with different values $\ell$ evolve independently, an immediate of conservation of momentum this The rightmost exponential in equation (\[eq:p\]) evolves the wave function from just after the ‘kick’ at $\tau=0$ to just before kick at under the of H_{dc}$, the operator to propagates it from just before to kick at a period. Since $U$ is an operator, the of it’s eigenvalues can be represented $$\label{eq:q} U_\ell\,\phi_{n\ell} = e^{i\,\varepsilon_{n\ell}}\,\phi_{n\ell}.$$ The set of eigenvalues \in (0,2\pi]\}$, are collectively known as the Quasi-energy Eigenvalues (QEE), and the eigenfunctions $\{\phi_{n\ell}\}$ as Eigenfunctions (QEF) of $U$. investigation of the dynamics the is equivalent to the nature of the QEE and QEF. The fundamental task is to obtain the Quasi-energy Spectrum (QES) of the evolution operator equation Classical Dynamics | eq:o2}
\tilde V_1(\tau) = \tilde V \, \delta(\taU-1), \qquad {\rm whEre}
\qqUad
\TilDe v = \fraC{1}{2}\ \eta\ R^2,$$
where the paramETer $\eTa$ has been defined previoUsly. ALl THe inFOrMatioN about tHE dYNAmiCs Of The SySTeM is coNtaIned witHin this FloQueT oPerator, since $\pHi(R,\phi,\tau+1) = U\,
\PhI(r,\pHi,\tau)$, where $\PhI$ is The totAl WavE FunctIon. becauSe of thE PeriodIc $\delta$-kiCkED dynamICs, we can IMMeDiatEly integrate equatIOn (\[EQ:o1\]) to get $$\label{eq:P}
U_\ell(1,0) = \eXp\LEfT(-\FRac{I}{{\tiLde\hbar}}\,\tilDe v\righT)\,
\Exp\left(-\FRaC{I}{{\TIldE\Hbar}}\,\tilde H_{dc}\rIght).$$ The subsCRipT $\ell$ haS bEen ATtacheD to $U$ tO eMPhaSize that the EvolUtion operAtor haS Been resTRicted tO that sIngLe $\eLl$ vaLUe. in OthEr WOrdS, StAteS WitH differeNt VaLues oF $\ell$ EVOLVe inDepEndeNtly, aN immediate conSeqUencE Of tHe conServaTion Of AngulAr momeNtum iN tHis system. The rigHtmoSt exponenTiaL oPerAtOr in eQUation (\[Eq:p\]) EvoLves the Wave funCTioN fROM JuSt after the ‘kick’ at $\taU=0$ tO JUsT before tHe kick AT oNe PEriod undEr The InflUENce of $\TildE h_{dC}$, while thE operaTOr To It’s left PrOpagatEs It fRom Just bEFore To just After the Kick aT A period.
Since $U$ iS An Unitary operAToR, THe SPectRum Of it’s eigenvAlueS Can bE repREsEntED as $$\laBel{eq:Q}
U_\ELl\,\PHi_{n\ell} = e^{i\,\varepsilon_{n\ElL}}\,\phi_{n\eLl}.$$ The Set of eigenvalUes $\{\varepsiLON_{N\ell} \in (0,2\pi]\}$, Are cOLlECtively known as The QuAsi-energy EIGenvalueS (QEE), aNd the eigEnfunctioNS $\{\Phi_{n\ell}\}$ aS thE QuAsi-EneRGY EIgenfunctions (qef) of $U$. thE investIgaTion of tHe qUanTum DynAmIcs of the sYstem is cOmPlEtElY eqUivalENt to deteRmIniNg The NaturE Of the QeE and qEF. THe FuNDamEntal taSK iS THus tO oBtAin tHe QUaSi-eneRgy SPEctRum (QES) oF the evoluTioN OperAtOr Given by Equation (\[eq:p\]).
ClAsSical DynamIcS | eq:o2}
\tilde V_1(\tau) =\tilde V \ , \de lta (\t au -1), \qq uad {\rm where }
\q quad
\tilde V = \frac {1}{2 }\ \eta \ r ^2,$$
where th e par am et er$\ e ta $ has be en defi ned previo usl y. All the inf o rm ation abou t t he dynamicsofthe sy st emi s con tai ned w ithint his Fl oquet ope ra t or, si n ce $\Ph i ( r, \phi ,\tau+1) = U\,
\P h i( r ,\phi,\tau)$,where$\ P hi $ isthe total wav efunct i on. Bec a us e o f t h e periodic $\ delta$-kick e d d ynamic s, we can im media te l y i ntegrate eq uati on (\[eq: o1\])t o get $ $ \label{ eq:p}U_\ ell (1,0 ) = \ exp \l e ft( - \f rac { i}{ {\tilde\ hb ar }}\,\ tild e V \ righ t)\ ,
\e xp\le ft(-\frac{i}{ {\t ilde \ hba r}}\, \tild e H_ {d c}\ri ght).$ $ The s ubscript $\ell$ has been att ach ed to $ U$ to emphas ize th at theevoluti o n o pe r a t or has been restrict ed t othat sin gle $\ e ll $v alue. In o the r wo r d s, st ates wi th diffe rent v a lu es of $\e ll $ evol ve in dep enden t ly,an imm ediate c onseq u ence of the co n servation ofa ng u l ar mome ntu m in this s yste m . Th e ri g ht mos t expo nenti al op e rator in equation ( \[ eq:p\] ) evo lves the wave functionf r o m just a fter th e ‘kick’ at $\t au=0$ to just b e fore the kick at oneperiod un d e r the in flu enc e o f $ \ t il de H_{dc}$, w h i le t he operat orto it’s le ftpro pag at es it fro m just b ef or eto ju st af t er the k ic k a ta p eriod .
Sinc e $U$ isan U n ita ry oper a to r , the s pe ctru m o fit’seige n val ues can be repre sen t ed a s$$ \label{ eq:q}
U_\ell\ ,\ phi_{n\ell }= e ^{i\,\ v a repsilon _{n\ell}}\,\phi_{n\ell} . $$ Theset of e igen values $\ {\v arepsi lon _ {n\ell } \in(0,2\ pi ]\} $ , arec o ll ect iv ely knowna s th e Qua si -ene rgy Eig envalues (QEE), an d th e eigenfuncti ons $\{ \ p hi _{n \ el l }\} $a s t h e Quasi-energy E igenfuncti on s ( QEF) of $U $ . T he invest igation of t h e quant um dynami cs of the s yste m iscompletely equival ent to de t ermin i ng thenat ure of t heQEE a nd QEF . Th e fun dament al taskis th us to obta in the Quasi-energy Spe ctrum(QES) of the evol uti o n o perator g iven by equati on(\[ eq:p\ ]).
Clas sica l D yna m ics | eq:o2}
\tilde V_1(\tau)_= \tilde_V \, \delta(\tau-1), \qquad_{\rm where}_
\qquad_
\tilde V_=_\frac{1}{2}\ \eta\ r^2,$$
where_the parameter $\eta$_has been defined previously._All the information_about_the dynamics of the system is contained within this Floquet operator, since $\Phi(r,\phi,\tau+1) =_U\,
\Phi(r,\phi,\tau)$,_where $\Phi$_is_the_total wave function. Because of_the periodic $\delta$-kicked dynamics, we_can immediately_integrate equation (\[eq:o1\]) to get $$\label{eq:p}
U_\ell(1,0) = \exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde_V\right)\,
\exp\left(-\frac{i}{{\tilde\hbar}}\,\tilde_H_{dc}\right).$$ The subscript_$\ell$ has been attached to $U$ to emphasize that_the evolution operator has been restricted_to that single_$\ell$_value._In other words, states_with different values of $\ell$ evolve_independently, an immediate consequence of the_conservation of angular momentum in this system._The rightmost exponential operator in equation_(\[eq:p\]) evolves the wave function_from just_after the ‘kick’ at $\tau=0$_to just before_the kick_at one period_under the influence of $\tilde H_{dc}$,_while the operator_to it’s left propagates it from_just_before to just_after_the_kick at_a period.
Since $U$_is_an Unitary_operator,_the spectrum of it’s eigenvalues can_be_represented as $$\label{eq:q}
U_\ell\,\phi_{n\ell} = e^{i\,\varepsilon_{n\ell}}\,\phi_{n\ell}.$$ The set_of eigenvalues $\{\varepsilon_{n\ell} \in_(0,2\pi]\}$,_are collectively known as_the Quasi-energy Eigenvalues (QEE), and_the eigenfunctions $\{\phi_{n\ell}\}$ as the Quasi-energy_Eigenfunctions (QEF)_of $U$._The investigation of the quantum dynamics of the system is completely_equivalent to determining the nature of_the QEE and QEF._The fundamental_task_is thus to_obtain_the Quasi-energy_Spectrum (QES) of the evolution operator given_by equation_(\[eq:p\]).
Classical Dynamics |
positive-definite
(resp. negative-definite) if $Q(X)\geq 0$ (resp. $Q(X)\leq 0$) for all $X\in {\mathbb{Q}}^n$ and $Q(X)=0$ if and only if $m=0$. In this paper, we are mostly interested in positive-definite integral ternary quadratic forms. For further information about ternary quadratic forms, a good survey may be found in [@Hanke].
We split the quadratic forms into
classes
, sets of quadratic forms which are equivalent under the action of ${\operatorname{GL}}_3({\mathbb{Z}})$. Two forms $Q$ and $\mathcal{Q}$ in the same class are referred to as
globally-equivalent
and we simply write $Q\sim_Z \mathcal{Q}$ for this relation. Classes are then grouped together based on their local conditions. For a positive-definite integral quadratic form ($a_{ij}\in{\mathbb{Z}}$) $$Q(X) = \sum_{1\leq i\leq j\leq n } a_{ij} X_i X_j,$$ since ${\mathbb{Z}}$ embeds into the $\ell$-adic integers ${\mathbb{Z}}_{\ell}$, it is natural to allow $X\in{\mathbb{Z}}_{\ell}$ and consider $Q$ as a quadratic form over ${\mathbb{Z}}_{\ell}$ (equivalently, we may tensor the Gram matrix with ${\mathbb{Z}}_{\ell}$ over ${\mathbb{Z}}$). Considering $Q$ over all ${\mathbb{Z}}_{\ell}$ simultaneously leads to an adelic interpretation; we do not investigate this further here, but simply note that we obtain a quadratic form $Q_{\ell}$ for each prime $\ell$. Two quadratic forms $Q$ and $\mathcal{Q}$ are
locally-equivalent
at the prime $\ell$ if they are equal under the action of an element of ${\operatorname{GL}}_3({\mathbb{Z}}_{\ell})$, and we denote this equivalence by $Q\sim_{{\mathbb{Z}}_{\ell}} \mathcal{Q}$. The set of equivalence classes which are locally-equivalent at all primes we call the
genus
of $Q$, and (a set of representatives for) the classes in the genus we denote by ${\operatorname{gen}}(Q)$. For the ternary case, the genus is then further | positive - definite
(resp. negative - definite) if $ Q(X)\geq 0 $ (resp. $ Q(X)\leq 0 $) for all $ X\in { \mathbb{Q}}^n$ and $ Q(X)=0 $ if and only if $ m=0$. In this paper, we are largely concerned in positive - definite integral ternary quadratic form. For further data about ternary quadratic forms, a good survey may be establish in [ @Hanke ].
We split the quadratic forms into
classes
, sets of quadratic forms which are equivalent under the legal action of $ { \operatorname{GL}}_3({\mathbb{Z}})$. Two forms $ Q$ and $ \mathcal{Q}$ in the same course are referred to as
globally - equivalent
and we plainly write $ Q\sim_Z \mathcal{Q}$ for this relation. class are then grouped together based on their local condition. For a positive - definite integral quadratic form ($ a_{ij}\in{\mathbb{Z}}$) $ $ Q(X) = \sum_{1\leq i\leq j\leq n } a_{ij } X_i X_j,$$ since $ { \mathbb{Z}}$ embeds into the $ \ell$-adic integer $ { \mathbb{Z}}_{\ell}$, it is natural to admit $ X\in{\mathbb{Z}}_{\ell}$ and consider $ Q$ as a quadratic form over $ { \mathbb{Z}}_{\ell}$ (equivalently, we may tensor the Gram matrix with $ { \mathbb{Z}}_{\ell}$ over $ { \mathbb{Z}}$). Considering $ Q$ over all $ { \mathbb{Z}}_{\ell}$ simultaneously leave to an adelic interpretation; we do not investigate this further here, but simply note that we obtain a quadratic mannequin $ Q_{\ell}$ for each prime $ \ell$. Two quadratic forms $ Q$ and $ \mathcal{Q}$ are
locally - equivalent
at the prime $ \ell$ if they are equal under the action of an element of $ { \operatorname{GL}}_3({\mathbb{Z}}_{\ell})$, and we denote this equivalence by $ Q\sim_{{\mathbb{Z}}_{\ell } } \mathcal{Q}$. The set of equivalence class which are locally - equivalent at all primes we call the
genus
of $ Q$, and (a hardening of representatives for) the classes in the genus we denote by $ { \operatorname{gen}}(Q)$. For the ternary case, the genus is then far |
posltive-definite
(resp. negatiye-definite) if $Q(X)\ywq 0$ (revp. $Q(X)\lsq 0$) for xll $X\in {\mathbb{Q}}^n$ and $Q(X)=0$ if aid obly id $m=0$. In this paper, we afe mostly interesred mn positive-definmfe intenxal tsvnary wuadratic formx. For furtver informatiot xblut ternary quadratic forms, a good furvey kaj be found in [@Rankt].
We spljn uhe quadratic forms into
classes
, ssts of xuadratic forks which are equivalent unfer hhe action of ${\operwtorname{GL}}_3({\mqthbf{Z}})$. Two forms $D$ and $\mathbcl{Q}$ in the aame class are referred to as
gluballv-equivalent
qne wf simply wrive $Q\sii_Z \mathcal{Q}$ npr thiv relatoon. Classes arc thei griuped together based mn their local conqitions. Fmr a positive-definire intecral quaaeatkc rocm ($z_{ij}\in{\mwthub{Z}}$) $$Q(X) = \sum_{1\meq i\leq j\lwq n } a_{ij} X_i X_j,$$ sinve ${\nathbb{Z}}$ embeda into tre $\ell$-adic integers ${\mathbb{Z}}_{\ell}$, it is nauural to allow $X\in{\mathbb{Z}}_{\ell}$ and consider $Q$ as a euadratic form over ${\mathbb{Z}}_{\ell}$ (equivalently, we may tensor dhe Gcao mcbvix dutj ${\mathbb{Z}}_{\ell}$ over ${\mathbb{Z}}$). Considering $Q$ over wml ${\msthbb{Z}}_{\ell}$ simujtaneously kewdx to an adelic inter'defation; we do not ijvestigwte tyis furthtr hete, but simply note that we ibtain a quabraric form $Q_{\ell}$ for zach prime $\eul$. Teo qusdratic forms $Q$ and $\matkcal{Q}$ zre
locally-eeuivalent
zg the prime $\ell$ kf nhey are equal under the actiog of an eoemeut of ${\opdratprname{DL}}_3({\mathbb{Z}}_{\epl})$, and we denote this eqkivalgnce bf $Q\sim_{{\mathhb{Z}}_{\ell}} \mathcal{Q}$. The set of equitelence classex fhibh are loeally-eauivalent at wll primes we eall the
yenus
ow $Q$, and (a aet of cepresentatides for) the cnwsses in the genus wq debote by ${\opefxtorname{gen}}(Q)$. Fpr the tegncry case, rhe genus is then nurthgr | positive-definite (resp. negative-definite) if $Q(X)\geq 0$ (resp. for $X\in {\mathbb{Q}}^n$ $Q(X)=0$ if and paper, are mostly interested positive-definite integral ternary forms. For further information about ternary forms, a good survey may be found in [@Hanke]. We split the quadratic into classes , sets of quadratic forms which are equivalent under the action ${\operatorname{GL}}_3({\mathbb{Z}})$. forms and in the same class are referred to as globally-equivalent and we simply write $Q\sim_Z \mathcal{Q}$ for relation. Classes are then grouped together based on local conditions. For a integral quadratic form ($a_{ij}\in{\mathbb{Z}}$) $$Q(X) \sum_{1\leq j\leq n a_{ij} X_j,$$ ${\mathbb{Z}}$ embeds into $\ell$-adic integers ${\mathbb{Z}}_{\ell}$, it is natural to allow $X\in{\mathbb{Z}}_{\ell}$ and consider $Q$ as a quadratic form over (equivalently, we the Gram with over Considering $Q$ over simultaneously leads to an adelic interpretation; investigate this further here, but simply note that obtain a form $Q_{\ell}$ for each prime $\ell$. quadratic forms $Q$ and $\mathcal{Q}$ are locally-equivalent at prime $\ell$ if they are equal under the action of an element of ${\operatorname{GL}}_3({\mathbb{Z}}_{\ell})$, and this equivalence by $Q\sim_{{\mathbb{Z}}_{\ell}} The set of classes are at primes we the genus of $Q$, and (a set of representatives for) the in the genus we denote by ${\operatorname{gen}}(Q)$. For the ternary genus then further |
positive-definite
(resp. negatIve-definitE) if $Q(X)\Geq 0$ (ResP. $Q(x)\leq 0$) For aLl $X\in {\mathbb{Q}}^n$ aND $Q(X)=0$ iF and only if $m=0$. In this paper, We are MoSTly iNTeResteD in posiTIvE-DEfiNiTe IntEgRAl TernaRy qUadratiC forms. For fUrtHeR information ABoUt ternary qUadRatic forms, a gOod Survey MaY be FOund iN [@HaNke].
We Split tHE quadrAtic forms InTO
classES
, sets of QUAdRatiC forms which are equIVaLEnt under the actIon of ${\oPeRAtORNamE{GL}}_3({\Mathbb{Z}})$. Two FoRms $Q$ aND $\mathcaL{q}$ iN THE saME class are refeRred to as
gloBAllY-equivAlEnt
ANd we siMply wRiTE $Q\sIm_Z \mathcal{Q}$ For tHis relatiOn. ClasSEs are thEN groupeD togetHer BasEd on THeIr LocAl COndITiOns. fOr a Positive-DeFiNite iNtegRAL QUadrAtiC forM ($a_{ij}\iN{\mathbb{Z}}$) $$Q(X) = \sum_{1\Leq I\leq J\Leq N } a_{ij} X_I X_j,$$ siNce ${\mAtHbb{Z}}$ eMbeds iNto thE $\eLl$-adic integers ${\mAthbB{Z}}_{\ell}$, it is NatUrAl tO aLlow $X\IN{\mathbB{Z}}_{\eLl}$ aNd consiDer $Q$ as a QUadRaTIC FoRm over ${\mathbb{Z}}_{\ell}$ (eqUiVALeNtly, we maY tensoR ThE GRAm matrix WiTh ${\mAthbB{z}}_{\Ell}$ ovEr ${\maTHbB{Z}}$). ConsidEring $Q$ OVeR aLl ${\mathbB{Z}}_{\Ell}$ simUlTanEouSly leADs to An adelIc interpRetatIOn; we do not invesTIgate this furtHEr HERe, BUt siMplY note that we ObtaIN a quAdraTIc ForM $q_{\ell}$ fOr eacH pRImE $\Ell$. Two quadratic formS $Q$ And $\matHcal{Q}$ Are
locally-equIvalent
at tHE PRime $\ell$ iF theY ArE Equal under the aCtion Of an elemenT Of ${\operatOrnamE{GL}}_3({\mathbB{Z}}_{\ell})$, and wE DEnote thiS eqUivAleNce BY $q\sIm_{{\mathbb{Z}}_{\ell}} \mATHcal{q}$. THe set of EquIvalencE clAssEs wHicH aRe locally-EquivaleNt At AlL pRimEs we cALl the
genUs
Of $Q$, AnD (a sEt of rEPresenTativEs foR) tHe CLasSes in thE GeNUS we dEnOtE by ${\oPerAtOrnamE{gen}}(q)$. for The ternAry case, thE geNUs is ThEn Further |
positive-definite
(resp. negative- defin ite ) i f$Q(X )\ge q 0$ (resp. $Q ( X)\l eq 0$) for all $X\in { \math bb { Q}}^ n $and $ Q(X)=0$ if a ndon ly if $ m =0 $. In th is pape r, we aremos tl y interested in positive- def inite integr alternar yqua d ratic fo rms.For fu r ther i nformatio na bout t e rnary q u a dr atic forms, a good su r ve y may be foundin [@H an k e] .
We sp lit the qu ad ratic forms i n to c las s es
, sets of quadraticf orm s whic hare equiva lentun d erthe actionof $ {\operato rname{ G L}}_3({ \ mathbb{ Z}})$. Tw o f orms $Q $and $ \ mat h ca l{Q } $ i n the sa me c lassarer e f e rred to as
glob ally-equivale nt
and wesimpl y wri te $ Q\ sim_Z \math cal{Q }$ for this relat ion. Classesare t hen g roupe d toget her ba sed ontheir l o cal c o n d it ions. For a positi ve - d ef inite in tegral qu ad r atic for m($a _{ij } \ in{\m athb b {Z }}$) $$Q (X) =\ su m_ {1\leqi\ leq j\ le q n }a_{ij } X_i X_j,$ $ since${\ma t hbb{Z}}$ embed s into the $\e l l$ - a di c int ege rs ${\mathb b{Z} } _{\e ll}$ , i t i s natu ral t oa ll o w $X\in{\mathbb{Z}} _{ \ell}$ andconsider $Q$as a quadr a t i c form o ver$ {\ m athbb{Z}}_{\el l}$ ( equivalent l y, we ma y ten sor theGram matr i x with ${ \ma thb b{Z }}_ { \ el l}$ over ${\m a t hbb{ Z} }$). Co nsi dering$Q$ ov erall $ {\mathbb{ Z}}_{\el l} $si mu lta neous l y leadsto an a del ic in t erpret ation ; we d on otinvesti g at e this f ur ther he re , but sim p lynote th at we obt ain a qu ad ra tic for m $Q_{\ell}$fo r each pri me $\ ell$.T w o quadra tic forms $Q$ and $\mat h cal{Q}$ ar e
lo call y-equival ent
at t hep rime $ \ell$if th ey ar e equal u nd erth e action o f aneleme nt of${\oper atorname{GL}}_3({\ m ath bb{Z}}_{\ell} )$, and w eden o te thi se qui v a lence by $Q\sim _{{\mathbb {Z } }_ {\ell}} \m a thc al {Q}$. T he setof eq u ivalenc e classes which ar eloca l l y-e quivalentat all p rimes wec all t h e
genu s
of $Q$ ,and (a s et ofr epr esent atives f or) th e cla ss es in th e genus we denote by ${ \opera torna me{ gen}}(Q)$ . F o r t he ternar y ca se, the ge nus is then fu r ther |
positive-definite
(resp. negative-definite)_if $Q(X)\geq_0$ (resp. $Q(X)\leq 0$)_for all_$X\in_{\mathbb{Q}}^n$ and_$Q(X)=0$_if and only_if $m=0$. In_this paper, we are_mostly interested in_positive-definite_integral ternary quadratic forms. For further information about ternary quadratic forms, a good survey_may_be found_in_[@Hanke].
We_split the quadratic forms into
classes
,_sets of quadratic forms which_are equivalent_under the action of ${\operatorname{GL}}_3({\mathbb{Z}})$. Two forms $Q$_and_$\mathcal{Q}$ in the_same class are referred to as
globally-equivalent
and we simply write_$Q\sim_Z \mathcal{Q}$ for this relation. Classes_are then grouped_together_based_on their local conditions._For a positive-definite integral quadratic form_($a_{ij}\in{\mathbb{Z}}$) $$Q(X) = \sum_{1\leq i\leq j\leq_n } a_{ij} X_i X_j,$$ since ${\mathbb{Z}}$_embeds into the $\ell$-adic integers ${\mathbb{Z}}_{\ell}$,_it is natural to allow_$X\in{\mathbb{Z}}_{\ell}$ and_consider $Q$ as a quadratic_form over ${\mathbb{Z}}_{\ell}$_(equivalently, we_may tensor the_Gram matrix with ${\mathbb{Z}}_{\ell}$ over ${\mathbb{Z}}$)._Considering $Q$ over_all ${\mathbb{Z}}_{\ell}$ simultaneously leads to an_adelic_interpretation; we do_not_investigate_this further_here, but simply_note_that we_obtain_a quadratic form $Q_{\ell}$ for each_prime_$\ell$. Two quadratic forms $Q$ and $\mathcal{Q}$_are
locally-equivalent
at the prime $\ell$_if_they are equal under_the action of an element_of ${\operatorname{GL}}_3({\mathbb{Z}}_{\ell})$, and we denote this_equivalence by_$Q\sim_{{\mathbb{Z}}_{\ell}} \mathcal{Q}$._The set of equivalence classes which are locally-equivalent at all primes_we call the
genus
of $Q$, and (a_set of representatives for)_the classes_in_the genus we_denote_by ${\operatorname{gen}}(Q)$._For the ternary case, the genus is_then further |
hopes of gaining the strengths of each; no attempt is made to prove optimality. This technique has also been used in graph colouring, where @Lewandowski:1996 observe that “in general, having the processors work in parallel yields better colorings faster than simply using multiple independent runs”.
This theme also shows up in the ManySAT parallel SAT solver [@Hamadi:2009], where parallelism is used to counter the sensitivity of SAT solvers to initial tuning parameters. Hamadi et al.remark that “the performance of parallel solvers is usually better on SAT problems than on UNSAT ones”. This is because multiple search trees may make it easier to find a solution, but they are less helpful in proving that there is no solution (although things are not this simple for SAT, where learned clauses are shared). In contrast, in cases where a good solution can be found quickly, our approach allows all of the work done by every thread to contribute to a proof of optimality—this is why we typically get at least near linear speedups. Some SAT solvers also make use of randomised restarts [@Gomes:1998], again with the aim of avoiding heavy commitment to any particular portion of the search space; the emphasis here is upon finding a solution to a satisfiable instance more quickly.
Finally, one could use these experiments to support the view that *sequential* maximum clique algorithms should use some form of early discrepancy search. We would not dispute this, although we note firstly that more complicated sequential search algorithms introduce overheads which cannot be offset during the proof of optimality stage of search, and secondly that either way, every modern processor has multiple cores and we should be making use of this [@Sutter:2005]. In any case, doing this would not invalidate our approach—we would simply see some of our superlinear speedups become linear speedups over a faster sequential baseline.
Conclusion
==========
We have considered some of the design choices available when parallelising a state-of-the-art branch and bound algorithm for the maximum clique problem. We have shown that the irregularity of subproblem sizes can cause workload balance problems—this should not be surprising. But we also saw that different parallel search orders will often produce substantially different speedups, to the extent that balance is usually *not* the deciding factor for performance.
This sheds new light on Lai and Sahni’s claim that “anomalous behaviour will be rarely witnessed in practice | hopes of gaining the strengths of each; no attempt is make to rise optimality. This technique has also been used in graph color, where @Lewandowski:1996 observe that “ in general, have the processors work in parallel yields better colorings quicker than simply using multiple independent runs ”.
This root also shows up in the ManySAT parallel SAT solver [ @Hamadi:2009 ], where parallelism is used to counter the sensitivity of SAT problem solver to initial tuning parameters. Hamadi et al.remark that “ the performance of parallel solvers is normally better on SAT problems than on UNSAT ones ”. This is because multiple search trees may make it easy to find a solution, but they are less helpful in proving that there is no solution (although things are not this simple for SAT, where learn clauses are shared). In contrast, in cases where a good solution can be find quickly, our approach allows all of the work done by every thread to contribute to a proof of optimality — this is why we typically have at least near linear speedups. Some SAT solvers besides form use of randomised restarts [ @Gomes:1998 ], again with the purpose of avoiding heavy commitment to any particular helping of the search space; the emphasis here is upon finding a solution to a satisfiable instance more quickly.
last, one could use these experiments to support the view that * consecutive * maximum clique algorithm should use some form of early discrepancy search. We would not dispute this, although we note firstly that more complicated sequential search algorithm introduce overheads which cannot be offset during the proof of optimality stage of search, and secondly that either way, every modern processor has multiple kernel and we should be making use of this [ @Sutter:2005 ]. In any case, doing this would not invalidate our access — we would simply interpret some of our superlinear speedups become analogue speedups over a faster sequential service line.
Conclusion
= = = = = = = = = =
We have considered some of the design choices available when parallelising a state - of - the - art branch and bound algorithm for the maximum clique problem. We have shown that the irregularity of subproblem sizes can induce workload balance problems — this should not be surprising. But we besides saw that different parallel search orders will much produce substantially different acceleration, to the extent that balance is usually * not * the deciding factor for performance.
This sheds new light on Lai and Sahni ’s title that “ anomalous behavior will be rarely witnessed in practice | hooes of gaining the strennths of each; no crtempt is mase to pruve optimality. This techniqux haw alsi been used in graph culouring, ahere @Leqandiqski:1996 obserts that “lu gensval, hcvmng the processprs work it parallel yiendr yetter colorings faster than simply tsing milhiple independgnt rlnf”.
Thia theme also shows up in the ManySZT paranlel SAT solvrr [@Hamadi:2009], where parallelisl is used to counter tje sensitivuty jd SAT solverr to initial tuning patameters. Hamadi et al.remark that “the 'erformance od pwtallel solvecs is lsually bettev on SAD problrms than on UNXAT onws”. This is because muntiple search treef may maka nt easier to find a silytion, but theh ard ltss hslpful in proving tgat there iw no solution (althoigr things are nkt thif fimple for SAT, where learned clauses art shaded). In contrast, in casew where a good solutiln can be found quickly, our approach allows all of the worn donx cy tvcvy tfeewd to contribute to a proof of optimality—this js wmy we typically net at least near kijest speedups. Somg SAT solvsrs also make use lf randjmisee restartf [@Gokes:1998], again with the aim of acoiding heavj conmitment to any paxticular porcion og the search space; the empharis gere is upoj finding x solution to a ratpsfidble insuxnce more quickly.
Sinally, oie cobld use ghesg experyments to duppovd the view that *seeuentnal* mdximum clieue algorithms should use some hprm of early dhscgepancy szarch. Ee would not qispute this, akthough we nute firstlj that moce complicatqd sequential dearch algormthms intwoduxe ocerheadr which cannot ne offset during thw proof of optimallty sjafe of search, anb wecondly that eothdr raj, etery imdern procesvor fas kultiole cortf cna we should be making usa of this [@Sutter:2005]. In anu gase, doint this wjuld not invakidate our approacj—we wmulv simpky fee some of our superlinear spsedups befomc linear speequps over a fastzr sequential baseline.
Conclusion
==========
We have ronsidered some of the eesign choices avaiksble when perallejising a vtate-of-the-art branch and bound algoriuhm for the maximum clisue prmblem. We have shown that the irregularity of subproblem sizes can cause workloqd balense problems—fhis shound uot be furpciring. But we also saw that different parallel seerch orderv cill often produce substantiaklh different skeedups, to the extent that balance is usually *not* the deciding favtor for performance.
This sheds nww lifht on Lai and Sahni’s claim yhat “auomaloys behaciouv will be rareuy eibnessed in prartiee | hopes of gaining the strengths of each; is to prove This technique has colouring, @Lewandowski:1996 observe that general, having the work in parallel yields better colorings than simply using multiple independent runs”. This theme also shows up in the parallel SAT solver [@Hamadi:2009], where parallelism is used to counter the sensitivity of solvers initial parameters. et al.remark that “the performance of parallel solvers is usually better on SAT problems than on ones”. This is because multiple search trees may it easier to find solution, but they are less in that there no (although are not this for SAT, where learned clauses are shared). In contrast, in cases where a good solution can be quickly, our all of work by thread to contribute proof of optimality—this is why we least near linear speedups. Some SAT solvers also use of restarts [@Gomes:1998], again with the aim avoiding heavy commitment to any particular portion of search space; the emphasis here is upon finding a solution to a satisfiable instance more one could use these to support the that maximum algorithms use some of early discrepancy search. We would not dispute this, although we firstly that more complicated sequential search algorithms introduce overheads which offset the proof of stage of search, and that way, every modern processor cores we use this In any case, doing would not invalidate our approach—we simply see some of speedups over a faster sequential baseline. Conclusion ========== have considered some of the design choices when parallelising a state-of-the-art branch and bound algorithm for the maximum clique We have the irregularity of subproblem sizes can cause workload problems—this should not be But we also saw that different parallel search orders often substantially different to the extent balance is usually the deciding factor This sheds light Lai that “anomalous behaviour will be rarely in | hopes of gaining the strengthS of each; no aTtempT is MadE tO proVe opTimality. This teCHniqUe has also been used in graPh colOuRIng, wHErE @LewaNdowski:1996 OBsERVe tHaT “iN geNeRAl, HavinG thE procesSors work in ParAlLel yields betTEr Colorings fAstEr than simply UsiNg multIpLe iNDepenDenT runs”.
this thEMe also Shows up in ThE manySAt ParalleL saT SolvEr [@Hamadi:2009], where paraLLeLIsm is used to couNter thE sENsITIviTy oF SAT solverS tO initIAl tuninG PaRAMEteRS. Hamadi et al.reMark that “the PErfOrmancE oF paRAllel sOlverS iS UsuAlly better oN SAT Problems tHan on UnsAT ones”. tHis is beCause mUltIplE seaRCh TrEes MaY MakE It EasIEr tO find a soLuTiOn, but They ARE LEss hElpFul iN provIng that there iS no SoluTIon (AlthoUgh thIngs ArE not tHis simPle foR SaT, where learned cLausEs are sharEd). IN cOntRaSt, in cASes wheRe a GooD solutiOn can be FOunD qUICKlY, our approach allows AlL OF tHe work doNe by evERy ThREad to conTrIbuTe to A PRoof oF optIMaLity—this Is why wE TyPiCally geT aT least NeAr lIneAr speEDups. some SAt solvers Also mAKe use of randomiSEd restarts [@GomES:1998], aGAIn WIth tHe aIm of avoidinG heaVY comMitmENt To aNY partIculaR pORtIOn of the search space; tHe EmphasIs herE is upon findinG a solution TO A SatisfiaBle iNStANce more quickly.
finalLy, one could USe these eXperiMents to sUpport the VIEw that *seQueNtiAl* mAxiMUM cLique algorithMS ShouLd Use some ForM of earlY diScrEpaNcy SeArch. We wouLd not disPuTe ThIs, AltHough WE note firStLy tHaT moRe comPLicateD sequEntiAl SeARch AlgoritHMs INTrodUcE oVerhEadS wHich cAnnoT Be oFfset duRing the prOof OF optImAlIty stagE of search, and sEcOndly that eItHer Way, eveRY Modern prOcessor has multiple cores ANd we shoUld Be makIng uSe of this [@SUttEr:2005]. In anY caSE, doing This woUld noT iNvaLIDate oUR ApProAcH—we would siMPLy sEe somE oF our SuperliNear speedups become LIneAr speedups oveR a fAsteR SEqUenTIaL BasElINe.
CONClusion
==========
We have coNsidered soMe OF tHe design chOIceS aVailablE when paRalleLIsing a sTate-of-the-Art branch AnD bouND AlgOrithm for tHe maximuM clique prOBlem. WE HaVe shoWn tHat the IrRegUlariTy of suBProBlem sIzes caN cAuse woRkloaD bAlance prOblems—this should not be suRprisiNg. But We aLso saw thaT diFFerEnt parallEl seArch orders WilL ofTen prOduCE subsTantIAlLy dIFfereNt spEEdups, to thE ExTenT THaT balance is uSUALly *Not* thE deCIding fActoR for performance.
ThIS sheds new light On LaI ANd SAhnI’S claIm That “anomalous bEhaViOUR will be rArEly witnesseD in practIcE | hopes of gaining the stre ngths of e ach;noatt em pt i s ma de to prove op t imal ity. This technique ha s als ob eenu se d ingraph c o lo u r ing ,wh ere @ L ew andow ski :1996 o bserve tha t “ in general, ha v in g the proc ess ors work inpar allelyi eld s bett ercolor ings f a ster t han simpl yu sing m u ltiplei n de pend ent runs”.
Thist he m e also shows u p in t he Ma n y SAT pa rallel SAT s olver [@Hamad i :2 0 0 9 ],w here parallel ism is used tocounte rthe sensit ivity o f SA T solvers t o in itial tun ing pa r ameters . Hamadi et al .re mar k th a t“t hepe r for m an ceo f p arallelso lv ers i s us u a l l y be tte r on SATproblems than on UNS A T o nes”. This isbe cause multi ple s ea rch trees may m akeit easier to f ind a solu t ion, b utthe y are l ess hel p ful i n p ro ving that there is n o so lution ( althou g hth i ngs areno t t hiss i mpleforS AT , wherelearne d c la uses ar eshared ). In co ntras t , in cases where a good solution can b e found quickl y ,o u ra ppro ach allows all oft he w orkd on e b y ever y thr ea d t o contribute to a pr oo f of o ptima lity—this iswhy we typ i c a lly getat l e as t near linear s peedu ps. Some S A T solver s als o make u se of ran d o mised re sta rts [@ Gom e s :1 998], again w i t h th eaim ofavo iding h eav y c omm itm en t to anyparticul ar p or ti onof th e searchsp ace ;the emph a sis he re is upo nfi n din g a sol u ti o n toasa tisf iab le inst ance mor e quick ly.
Fina lly , one c ou ld usethese experim en ts to supp or t t he vie w that *se quential* maximum cliqu e algori thm s sho ulduse somefor m of e arl y discr epancy sear ch . W e would n ot di sp ute this,a l tho ugh w enote firstl y that more compli c ate d sequentialsea rcha l go rit h ms int ro d uce o verheads whichcannot beof f se t during t h e p ro of of o ptimali ty st a ge of s earch, an d secondl ythat e ith er way, ev ery mode rn proces s or ha s m ultip lecoresan d w e sho uld be mak ing u se ofth is [@S utter :2 005]. In any case, doing this w ould n ot in val idate our ap p roa ch—we wou ld s imply seesom e o f our su p erlin ears pe edu p s bec omel inear spe e du pso v er a faster s e q u ent ial b ase l ine.
Conc lusion
==========
We have consi dere d som e o f the d esign choicesava il a b le whenpa rallelising a state -o f -the- art br anch a nd boun d al g orithm for th e maximum cl iq u e probl em .W e have sho wn thatthe ir r egul a r ity of subproble m siz e s canc aus e wor kl oad bal a nceproblems—t his shouldnot be sur prisi ng. But w e also sa wthat diffe r ent paral lel s earch o rd erswil l ofte n pr o d uce s ubst an tia lly diffe r e nt sp ee d ups , to theex tent that bal a nce is u sua l ly *not *the d ecidin g f a c tor for pe rfo rmanc e .
This she d s ne w l i ght o n Laiand Sa hni’s c l aim t hat “an oma l o us behavi our willb e r ar elywitnesse din p ra cti ce | hopes_of gaining_the strengths of each;_no attempt_is_made to_prove_optimality. This technique_has also been_used in graph colouring,_where @Lewandowski:1996 observe_that_“in general, having the processors work in parallel yields better colorings faster than simply_using_multiple independent_runs”.
This_theme_also shows up in the_ManySAT parallel SAT solver [@Hamadi:2009],_where parallelism_is used to counter the sensitivity of SAT_solvers_to initial tuning_parameters. Hamadi et al.remark that “the performance of parallel_solvers is usually better on SAT_problems than on_UNSAT_ones”._This is because multiple_search trees may make it easier_to find a solution, but they_are less helpful in proving that there_is no solution (although things are_not this simple for SAT,_where learned_clauses are shared). In contrast,_in cases where_a good_solution can be_found quickly, our approach allows all_of the work_done by every thread to contribute_to_a proof of_optimality—this_is_why we_typically get at_least_near linear_speedups._Some SAT solvers also make use_of_randomised restarts [@Gomes:1998], again with the aim_of avoiding heavy commitment_to_any particular portion of_the search space; the emphasis_here is upon finding a solution_to a_satisfiable instance_more quickly.
Finally, one could use these experiments to support the view_that *sequential* maximum clique algorithms should_use some form of_early discrepancy_search._We would not_dispute_this, although_we note firstly that more complicated sequential_search algorithms_introduce overheads which cannot be offset_during the proof of_optimality_stage of search, and secondly that_either way, every modern processor has_multiple cores and we should_be_making_use of this [@Sutter:2005]. In_any case, doing this would not_invalidate our approach—we_would simply see some of our superlinear_speedups_become linear speedups over a faster_sequential_baseline.
Conclusion
==========
We have considered some of the_design_choices_available when parallelising a state-of-the-art_branch and bound algorithm for the_maximum clique problem. We have shown that the irregularity_of subproblem sizes_can cause workload balance problems—this_should_not_be surprising. But we also saw that different parallel search_orders will_often produce substantially_different speedups, to the extent that balance is usually *not*_the deciding factor for performance.
This sheds new_light on Lai and Sahni’s claim that “anomalous behaviour will be_rarely witnessed in practice |
1961, Z. Physik 164, 1
Sunyaev R. A., Titarchuk L. G., 1980, A&A 86, 121
Sunyaev R. A., [*et al.*]{}, 1991, ApJ 383, L49
Sunyaev R. A., [*et al.*]{}, 1992, ApJ 389, L75
Svensson R., 1982, ApJ 258, 321
Svensson R., 1990, in Physical Processes in Hot Cosmic Plasmas, eds. W. Brinkmann, A.C. Fabian, F. Giovannelli, Kluwer, Dordrecht, p.357
[^1]: which follows from suggestion $f(\gamma)=f_1(\gamma)g(\gamma)$, where $f_1(\gamma)$ is a Maxwellian.
[^2]: We took $n_+\le
n_-\approx1.5\times10^8$ cm$^{-3}$ $(\Delta_d/10$ days$)^{-1}$.
[^3]: A possibility that Nova Muscae lies in front of a large gas cloud can not be totally excluded. In this case, particles could be injected into this cloud, away from the observer.
[^4]: The corresponding annihilation lifetime $\Delta_d$ (eq. \[\[8.2\]\]–\[\[8.3\]\]) was obtained for thermal plasma and is not valid for the cold medium where positrons mostly annihilate in the bound (positronium) state.
---
abstract: 'We present the exact solution of a Brownian inchworm model of a self-propelled elastic dimer which has recently been proposed in \[K. V. Kumar *et al*, Phys. Rev. E **77**, 020102(R) (2008)\] as a unifying model for the propulsion mechanisms of DNA helicase, polar rods on a vibrated surface, crawling keratocytes, and Myosin VI.'
author:
- 'A. Baule'
- 'K. Vijay Kumar'
- Sriram Ramaswamy
title: 'Exact solution of a Brownian inchworm model for self-propulsion'
---
Introduction
============
Methods of nonequilibrium statistical mechanics are particularly useful in describing transport processes in biological systems. In this article we investigate a none | 1961, Z. Physik 164, 1
Sunyaev R. A., Titarchuk L. G., 1980, A&A 86, 121
Sunyaev R. A., [ * et al. * ] { }, 1991, ApJ 383, L49
Sunyaev R. A., [ * et al. * ] { }, 1992, ApJ 389, L75
Svensson R., 1982, ApJ 258, 321
Svensson R., 1990, in Physical Processes in Hot Cosmic Plasmas, eds. W. Brinkmann, A.C. Fabian, F. Giovannelli, Kluwer, Dordrecht, p.357
[ ^1 ]: which follows from suggestion $ f(\gamma)=f_1(\gamma)g(\gamma)$, where $ f_1(\gamma)$ is a Maxwellian.
[ ^2 ]: We took $ n_+\le
n_-\approx1.5\times10 ^ 8 $ cm$^{-3}$ $ (\Delta_d/10 $ days$)^{-1}$.
[ ^3 ]: A possibility that Nova Muscae dwell in presence of a large gas cloud cannot be wholly excluded. In this case, particles could be inject into this cloud, away from the perceiver.
[ ^4 ]: The correspond annihilation lifetime $ \Delta_d$ (eq. \[\[8.2\]\]–\[\[8.3\]\ ]) was prevail for thermal plasma and is not valid for the cold medium where positrons largely annihilate in the bound (positronium) state.
---
abstract:' We present the exact solution of a Brownian inchworm exemplar of a self - propelled elastic dimer which has recently been proposed in \[K. V. Kumar * et al *, Phys. Rev. E * * 77 * *, 020102(R) (2008)\ ] as a unifying mannequin for the propulsion mechanisms of DNA helicase, polar rods on a vibrated surface, crawl keratocytes, and Myosin VI.'
author:
-' A. Baule'
-' K. Vijay Kumar'
- Sriram Ramaswamy
title:' Exact solution of a Brownian inchworm model for self - propulsion'
---
Introduction
= = = = = = = = = = = =
method of nonequilibrium statistical mechanics are particularly useful in describing transport processes in biological systems. In this article we investigate a none | 1961, Z. Ohysik 164, 1
Sunyaev R. A., Titarghuk L. G., 1980, A&A 86, 121
Suniawv R. A., [*xt al.*]{}, 1991, ZpJ 383, L49
Suvyaev R. A., [*et al.*]{}, 1992, ApJ 389, L75
Svenssln R., 1982, AkL 258, 321
Svensson R., 1990, in Phhsical Prlcesses un Hir Cosmic Plasmas, eds. W. Brihnmanu, E.C. Fabian, F. Giovaknelli, Kluwar, Dordrecht, p.357
[^1]: wfieh follows from suggestion $f(\gamma)=f_1(\gaima)g(\gamka)$, where $f_1(\gamma)$ ys a Iaxwsllian.
[^2]: We took $n_+\le
n_-\approx1.5\times10^8$ dm$^{-3}$ $(\Delta_v/10$ days$)^{-1}$.
[^3]: A possibikity that Nova Muscae lies in vront of a large gws cloud cab noe be totally dxcluded. In this case, karticles could be injected into this cloud, awai rrlk the obserter.
[^4]: Thv correspondikb annivilatiom lifetime $\Delba_d$ (ex. \[\[8.2\]\]–\[\[8.3\]\]) wqs obtained for thermel plasma and is not valid fos che cold medium where pisitrmns kostut avnigikafe in hhe bound (posjtronium) stqte.
---
abstract: 'We prexegn the exact sklutiog jf a Brownian inchworm model of a self-pgopemled elastic dimer whicy has recently been ptoposed in \[K. V. Kumar *et al*, Phys. Rev. E **77**, 020102(R) (2008)\] as a unifying modal foc ghe pvopuuwiln mechanisms of DNA helicase, polar rods on a binrsted surface, cvawling keratocytex, wnc Myosin VI.'
autfor:
- 'A. Yzume'
- 'K. Vijay Kumar'
- Sgiram Rwmaswqmy
title: 'Txact solution of a Brownian incyworm model yor self-propulsion'
---
Intxoduction
============
Metkods og nonrquilibrium statistical mecganics are oarticuladuy useful in deszrininc transport processes in byological syscems. In ghis articje we invedtigaba a none | 1961, Z. Physik 164, 1 Sunyaev R. L. 1980, A&A 121 Sunyaev R. 383, Sunyaev R. A., al.*]{}, 1992, ApJ L75 Svensson R., 1982, ApJ 258, Svensson R., 1990, in Physical Processes in Hot Cosmic Plasmas, eds. W. Brinkmann, Fabian, F. Giovannelli, Kluwer, Dordrecht, p.357 [^1]: which follows from suggestion $f(\gamma)=f_1(\gamma)g(\gamma)$, where is Maxwellian. We $n_+\le n_-\approx1.5\times10^8$ cm$^{-3}$ $(\Delta_d/10$ days$)^{-1}$. [^3]: A possibility that Nova Muscae lies in front of a gas cloud can not be totally excluded. In case, particles could be into this cloud, away from observer. The corresponding lifetime (eq. was obtained for plasma and is not valid for the cold medium where positrons mostly annihilate in the bound (positronium) --- abstract: the exact of Brownian model of a dimer which has recently been proposed Kumar *et al*, Phys. Rev. E **77**, 020102(R) as a model for the propulsion mechanisms of helicase, polar rods on a vibrated surface, crawling and Myosin VI.' author: - 'A. Baule' - 'K. Vijay Kumar' - Sriram Ramaswamy title: of a Brownian inchworm for self-propulsion' --- ============ of statistical are particularly in describing transport processes in biological systems. In this article we a none | 1961, Z. Physik 164, 1
Sunyaev R. A., Titarchuk l. G., 1980, A&A 86, 121
SunyaeV R. A., [*et Al.*]{}, 1991, APJ 383, L49
suNyaeV R. A., [*eT al.*]{}, 1992, ApJ 389, L75
SvenssoN r., 1982, ApJ 258, 321
svensson R., 1990, in Physical ProCesseS iN hot COSmIc PlaSmas, eds. w. brINKmaNn, a.C. fabIaN, f. GIovanNelLi, KluweR, Dordrecht, P.357
[^1]: whIcH follows from SUgGestion $f(\gaMma)=F_1(\gamma)g(\gamma)$, WheRe $f_1(\gamMa)$ Is a mAxwelLiaN.
[^2]: We toOk $n_+\le
n_-\APprox1.5\tImes10^8$ cm$^{-3}$ $(\DelTa_D/10$ Days$)^{-1}$.
[^3]: A pOSsibiliTY ThAt NoVa Muscae lies in froNT oF A large gas cloud Can not Be TOtALLy eXclUded. In this CaSe, parTIcles coULd BE INjeCTed into this clOud, away from THe oBserveR.
[^4]: THe cORrespoNding AnNIhiLation lifetIme $\DElta_d$ (eq. \[\[8.2\]\]–\[\[8.3\]\]) waS obtaiNEd for thERmal plaSma and Is nOt vAlid FOr ThE coLd MEdiUM wHerE PosItrons moStLy AnnihIlatE IN THe boUnd (PosiTroniUm) state.
---
abstraCt: 'WE preSEnt The exAct soLutiOn Of a BrOwnian InchwOrM model of a self-prOpelLed elastiC diMeR whIcH has rECently BeeN prOposed iN \[K. V. KumaR *Et aL*, PHYS. reV. E **77**, 020102(R) (2008)\] as a unifying modeL fOR ThE propulsIon mecHAnIsMS of DNA heLiCasE, polAR Rods oN a viBRaTed surfaCe, crawLInG kEratocyTeS, and MyOsIn Vi.'
auThor:
- 'A. bAule'
- 'k. Vijay kumar'
- SriRam RaMAswamy
title: 'ExaCT solution of a BROwNIAn INchwOrm Model for selF-proPUlsiOn'
---
InTRoDucTIon
============
MeThods Of NOnEQuilibrium statisticAl MechanIcs arE particularly Useful in deSCRIbing traNspoRT pROcesses in bioloGical Systems. In tHIs articlE we inVestigatE a none | 1961, Z. Physik 164, 1
S unyaev R.A., T ita rch uk L.G.,1980, A&A 86,1 21
Sunyaev R. A., [*et al .*]{} ,1 991, Ap J 383 , L49
S un y a evR. A .,[* e tal.*] {}, 1992,ApJ 389, L 75
S vensson R.,1 98 2, ApJ 258 , 3 21
Svensson R. , 1990 ,inP hysic alProce sses i n Hot C osmic Pla sm a s, eds . W. Bri n k ma nn,A.C. Fabian, F. G i ov a nnelli, Kluwer , Dord re c ht , p.3 57
[^1]: whi ch foll o ws from su g g e sti o n $f(\gamma)= f_1(\gamma) g (\g amma)$ ,whe r e $f_1 (\gam ma ) $ i s a Maxwell ian.
[^2]: W e took $n_+\le n_- \appro x1. 5\t imes 1 0^ 8$ cm $^ { -3} $ $ (\D e lta _d/10$ d ay s$ )^{-1 }$.[ ^ 3 ]: A po ssib ility that Nova Mu sca e li e s i n fro nt of a l ar ge ga s clou d can n ot be totally e xclu ded. In t his c ase ,parti c les co uld be inject ed into thi sc l o ud , away from the ob se r v er .
[^4]: The c o rr es p onding a nn ihi lati o n life time $\ Delta_d$ (eq.\ [\ [8 .2\]\]– \[ \[8.3\ ]\ ])was obta i nedfor th ermal pl asmaa nd is not vali d for the cold me d i um wher e p ositrons mo stly anni hila t eint he bo und ( po s it r onium) state.
---ab stract : 'We present theexact solu t i o n of a B rown i an inchworm model of a self-prop e lled ela sticdimer wh ich has r e c ently be enpro pos edi n \ [K. V. Kumar* e t al *, Phys.Rev . E **7 7** , 0 201 02( R) (2008)\] as a un if yi ng m ode l for the prop ul sio nmec hanis m s of D NA he lica se ,p ola r rodso na vibr at ed sur fac e, craw ling ker atocyte s, and My osi n VI. 'au thor:
- 'A. Baule'
- ' K. Vijay K um ar'
- Sri r a m Ramasw amy
title: 'Exact solut i on of a Br ownia n in chworm mo del for s elf - propul sion'---
In tro d u ction = == === == ====
Meth o d s o f non eq uili brium s tatistical mechani c s a re particular lyusef u l i n d e sc r ibi ng tra n s port processesin biologi ca l s ystems. In thi sarticle we inv estig a te a no ne | 1961,_Z. Physik 164,_1
Sunyaev R. A., Titarchuk L. G.,_1980, A&A_86,_121
Sunyaev R. A.,_[*et_al.*]{}, 1991, ApJ_383, L49
Sunyaev R. A.,_[*et al.*]{}, 1992, ApJ_389, L75
Svensson R.,_1982,_ApJ 258, 321
Svensson R., 1990, in Physical Processes in Hot Cosmic Plasmas, eds. W. Brinkmann, A.C. Fabian,_F. Giovannelli,_Kluwer, Dordrecht,_p.357
[^1]:_which_follows from suggestion $f(\gamma)=f_1(\gamma)g(\gamma)$, where_$f_1(\gamma)$ is a Maxwellian.
[^2]: We_took $n_+\le
_ n_-\approx1.5\times10^8$ cm$^{-3}$ $(\Delta_d/10$ days$)^{-1}$.
[^3]: A possibility that Nova_Muscae_lies in front_of a large gas cloud can not be totally_excluded. In this case, particles could_be injected into_this_cloud,_away from the observer.
[^4]:_The corresponding annihilation lifetime $\Delta_d$ (eq. \[\[8.2\]\]–\[\[8.3\]\])_was obtained for thermal plasma and_is not valid for the cold medium_where positrons mostly annihilate in the_bound (positronium) state.
---
abstract: 'We_present the_exact solution of a Brownian_inchworm model of_a self-propelled_elastic dimer which_has recently been proposed in \[K._V. Kumar *et_al*, Phys. Rev. E **77**, 020102(R)_(2008)\]_as a unifying_model_for_the propulsion_mechanisms of DNA_helicase,_polar rods_on_a vibrated surface, crawling keratocytes, and_Myosin_VI.'
author:
- 'A. Baule'
- 'K. Vijay Kumar'
- Sriram_Ramaswamy
title: 'Exact solution of_a_Brownian inchworm model for_self-propulsion'
---
Introduction
============
Methods of nonequilibrium statistical mechanics_are particularly useful in describing transport_processes in_biological systems._In this article we investigate a none |
to orbital phase coverage of eight observing days per orbital period. This phase coverage is a guideline and it should be noted that increasing or decreasing the number, $N_{\rm obs}$, of observations per orbital period will affect the observed RMS by $\sqrt{\rm N_{\rm obs}}$. Noise from a random Gaussian distribution centered at 2 m s$^{-1}$ was added to each measurement to account for measurement uncertainty. For consistency, each comparison data set (three separated by $\Delta t_{\rm opt}$, three consecutive, and three consecutive - no p-modes) has the same number of data points, N = 36.
For each injected planet described above, we then computed a Lomb-Scargle Periodogram to determine the period of each individual data set. We deemed the planet to be recovered if the maximum peak of the LS periodogram was consistent within 1% of the true period of 10.0 days, i.e..
We show in Figure \[fig:trec\] the fraction of planets recovered out of 1000 trials as a function of both semi-amplitude, K, and minimum planet mass in Earth masses. Our method recovers more injected planets than the standard observing strategy for subgiant stars. It should be expected that our method does not recover as many planets as the the simulation where no p-modes were present as this represents the idealized noise floor.
Figure \[fig:massdiff\] is the ratio of the number of planets recovered using our method, three observations optimally sampled (green curve Figure \[fig:trec\]), to the number of planets recovered using the standard observing strategy for subgiant stars (blue curve Figure \[fig:trec\]). Our method recovers a factor of ten more planets than the standard observing strategy, specifically in the range of planet masses between Neptune and Saturn shown as the dashed magenta and black lines respectively.
Quantifying $\Delta$t
=====================
In this section we develop a parameterization of the relationship between $\Delta t$, the optimal time between intra-night observations, and the mass and radius of the star. This relationship can be used to inform future observations of subgiant radial velocity surveys focused on planet detection. In the development of this parameterization, we examined the relationships between $\Delta t$ and other stellar parameters such as luminosity and effective temperature, but determined that mass and radius encapsulate the primary contributions to $\Delta t$.
To develop the numerical model, we generated a population of stars with | to orbital phase coverage of eight observing days per orbital period. This phase coverage is a guideline and it should be note that increase or decreasing the number, $ N_{\rm obs}$, of observations per orbital menstruation will affect the observed RMS by $ \sqrt{\rm N_{\rm obs}}$. randomness from a random Gaussian distribution centered at 2 m s$^{-1}$ was added to each measurement to report for measurement doubt. For consistency, each comparison data set (three separate by $ \Delta t_{\rm opt}$, three consecutive, and three consecutive - no p - mood) has the like number of datum detail, N = 36.
For each injected planet described above, we then calculate a Lomb - Scargle Periodogram to determine the period of each individual data set. We deemed the satellite to be recovered if the maximal peak of the LS periodogram was consistent within 1% of the true period of 10.0 day, i.e..
We show in Figure \[fig: trec\ ] the fraction of planets recovered out of 1000 trials as a function of both semi - amplitude, K, and minimal satellite batch in Earth masses. Our method recovers more injected planets than the standard observing scheme for subgiant stars. It should be expect that our method does not reclaim equally many planet as the the simulation where no p - modes were present as this represents the idealized noise floor.
Figure \[fig: massdiff\ ] is the ratio of the number of planet recovered using our method, three observations optimally sampled (green bend Figure \[fig: trec\ ]), to the number of planets recovered using the standard note scheme for subgiant stars (blue curve Figure \[fig: trec\ ]). Our method recovers a factor of ten more planets than the standard observing scheme, specifically in the range of satellite masses between Neptune and Saturn shown as the dashed magenta and black lines respectively.
quantify $ \Delta$t
= = = = = = = = = = = = = = = = = = = = =
In this section we develop a parameterization of the kinship between $ \Delta t$, the optimal time between intra - night observations, and the batch and radius of the star topology. This relationship can be used to inform future observations of subgiant radial velocity surveys concentrate on planet signal detection. In the development of this parameterization, we examined the relationships between $ \Delta t$ and other leading parameters such as luminosity and effective temperature, but determined that mass and radius encapsulate the chief contributions to $ \Delta t$.
To develop the numerical model, we generated a population of stars with | to orbital phase coverage uf eight observnbg dayv per krbital oeriod. This phase coverage id q guieeline and it should bd noted tjat incrwasiig or decreasing the numnzr, $N_{\rj obs}$, if observationx per orbidal period wiln xfyect the observed RMS by $\sqrt{\rm N_{\rm jbs}}$. Noixe from a random Gauxfian distribution centered at 2 m s$^{-1}$ was zdded tm each measurrment to account for measugemejt uncertainty. For consistenci, easy comparison data set (uhxee separatgd by $\Delta t_{\rm opt}$, three consecjtive, and three cinsfwutive - no '-modes) has the samc numbes of daya points, N = 36.
Fpr xach injected planet desccibed above, we then somputed d Momb-Scargle Perioeotram jo dedermkbe ghe pxrikd of faci individuam data set. Qe deemed the planeu tj be recovered if thq iaximum peak of the LS periodogram was bonsjstent within 1% of the teue period of 10.0 days, i.g..
We show ig Figure \[fig:trec\] the fraction of planets recovereg out uf 1000 tvialr ad a function of both semi-amplitude, K, and miniihm pkanet mass in Carth masses. Our mrtjoc recovers morg injeccsd planets than the dtandarq obswrving stwateby for subgiant stars. It shiuld be expebted that our method dles not recuver as msny planets as the the rimumation wherf no p-modsr were present ar tmis represents the idealized goise floir.
Fiyure \[fig:oassciff\] if the ratil of bve number of planehs reeoverad using okr method, three observations opvmmally samplec (crevn curve Yigure \[fig:trec\]), to tre number of pkanets xecovefed using nhe standerd observind strategy fos subgiant sters (blue survw Fiture \[fie:grec\]). Our methoc recovers a factor of ten more planebs thxh the standard ibstrcing strategy, xpezifycwlky hn the range of olavrt marses beuceek Ndptume and Saturn shown ds tge dashed magenta snq black oines refpectively.
Quamtifying $\Delta$t
=====================
In uhis sxction we drvejop a parameterization of the delationsjip between $\Delea t$, the optimal time between intra-night observations, aid the mass and radius if the star. This rekstionship cen be tsed to itform future observarions of subgiant radial velocity survsys fowused on planet detection. In the development of this parameterization, we examibed thx welationshils brtweet $\Belca t$ anq otiec stellar parametvrs such as luminosity and effecvive tempesacure, but determined that mass avd radius encxpsulate the primary confributioms to $\Delta t$.
To develop the numrrical model, we generatec a popuoatioh of snzrr with | to orbital phase coverage of eight observing orbital This phase is a guideline that or decreasing the $N_{\rm obs}$, of per orbital period will affect the RMS by $\sqrt{\rm N_{\rm obs}}$. Noise from a random Gaussian distribution centered at m s$^{-1}$ was added to each measurement to account for measurement uncertainty. For each data (three by $\Delta t_{\rm opt}$, three consecutive, and three consecutive - no p-modes) has the same number data points, N = 36. For each injected described above, we then a Lomb-Scargle Periodogram to determine period each individual set. deemed planet to be if the maximum peak of the LS periodogram was consistent within 1% of the true period of days, i.e.. in Figure the of recovered out of as a function of both semi-amplitude, planet mass in Earth masses. Our method recovers injected planets the standard observing strategy for subgiant It should be expected that our method does recover as many planets as the the simulation where no p-modes were present as this idealized noise floor. Figure is the ratio the of recovered our method, observations optimally sampled (green curve Figure \[fig:trec\]), to the number of recovered using the standard observing strategy for subgiant stars (blue \[fig:trec\]). method recovers a of ten more planets the observing strategy, specifically in of masses Saturn as dashed magenta and black respectively. Quantifying $\Delta$t ===================== In section we develop a $\Delta t$, the optimal time between intra-night observations, the mass and radius of the star. relationship can be used to inform future observations of subgiant radial velocity focused on In the development of this parameterization, we examined relationships between $\Delta t$ other stellar parameters such as luminosity and effective temperature, determined mass and encapsulate the primary to $\Delta t$. develop the numerical generated a of with | to orbital phase coverage of eIght observIng daYs pEr oRbItal PeriOd. This phase covERage Is a guideline and it shoulD be noTeD That INcReasiNg or decREaSINg tHe NuMbeR, $N_{\RM oBs}$, of oBseRvationS per orbitaL peRiOd will affect THe Observed RMs by $\Sqrt{\rm N_{\rm obs}}$. noiSe from A rAndOM GausSiaN distRibutiON centeRed at 2 m s$^{-1}$ waS aDDed to eACh measuREMeNt to Account for measureMEnT Uncertainty. For ConsisTeNCy, EACh cOmpArison data SeT (threE SeparatED bY $\dELta T_{\Rm opt}$, three conSecutive, and THreE conseCuTivE - No p-modEs) has ThE SamE number of daTa poInts, N = 36.
For eAch injECted plaNEt descrIbed abOve, We tHen cOMpUtEd a loMB-ScARgLe PERioDogram to DeTeRmine The pERIOD of eAch IndiViduaL data set. We deeMed The pLAneT to be RecovEred If The maXimum pEak of ThE LS periodogram wAs coNsistent wIthIn 1% Of tHe True pERiod of 10.0 DayS, i.e..
we show iN Figure \[FIg:tReC\] THE fRaction of planets reCoVEReD out of 1000 trIals as A FuNcTIon of botH sEmi-AmplITUde, K, aNd miNImUm planet Mass in eArTh Masses. OUr Method ReCovErs More iNJectEd planEts than tHe staNDard observing sTRategy for subgIAnT STaRS. It sHouLd be expecteD thaT Our mEthoD DoEs nOT recoVer as MaNY pLAnets as the the simulaTiOn wherE no p-mOdes were preseNt as this rePRESents the IdeaLIzED noise floor.
FigUre \[fiG:massdiff\] iS The ratio Of the Number of Planets reCOVered usiNg oUr mEthOd, tHREe Observations oPTImalLy Sampled (GreEn curve figUre \[Fig:TreC\]), tO the numbeR of planeTs ReCoVeRed Using THe standaRd ObsErVinG straTEgy for SubgiAnt sTaRs (BLue Curve FiGUrE \[FIg:trEc\]). ouR metHod ReCoverS a faCTor Of ten moRe planets ThaN The sTaNdArd obseRving strategy, SpEcifically In The Range oF PLanet masSes between Neptune and SatURn shown As tHe dasHed mAgenta and BlaCk lineS reSPectivEly.
QuaNtifyInG $\DeLTA$t
=====================
In tHIS sEctIoN we develop A PAraMeterIzAtioN of the rElationship between $\dEltA t$, the optimal tIme BetwEEN iNtrA-NiGHt oBsERvaTIOns, and the mass anD radius of tHe STaR. This relatIOnsHiP can be uSed to inForm fUTure obsErvations Of subgianT rAdiaL VEloCity surveyS focused On planet dETectiON. IN the dEveLopmenT oF thIs parAmeterIZatIon, we ExaminEd The relAtionShIps betweEn $\Delta t$ and other stellar ParameTers sUch As luminosIty ANd eFfective tEmpeRature, but dEteRmiNed thAt mASs and RadiUS eNcaPSulatE the PRimary conTRiButIONs To $\Delta t$.
To dEVELop The nuMerICal modEl, we Generated a populatIOn of stars with | to orbital phase coverage of eightobser vin g d ay s pe r or bital period.T hisphase coverage is a gu ideli ne andi tshoul d be no t ed t hat i nc rea si n gor de cre asing t he number, $N _{ \rm obs}$, o f o bservation s p er orbital p eri od wil laff e ct th e o bserv ed RMS by $\s qrt{\rm N _{ \ rm obs } }$. Noi s e f roma random Gaussian di s tribution cent ered a t2 m s $^{ -1} $ was adde dto ea c h measu r em e n t to account for m easurementu nce rtaint y. Fo r consi stenc y, eac h compariso n da ta set (t hree s e parated by $\De lta t_ {\r m o pt}$ , t hr eeco n sec u ti ve, and three c on se cutiv e -n o p -mod es) has thesame number o f d atap oin ts, N = 36 .
F or each injec ted p la net described a bove , we then co mp ute da Lom b -Scarg lePer iodogra m to de t erm in e t he period of each in di v i du al dataset. W e d ee m ed the p la net tob e reco vere d i f the ma ximump ea kof theLS perio do gra m w as co n sist ent wi thin 1%of th e true period o f 10.0 days, i . e. .
W e sho w i n Figure \[ fig: t rec\ ] th e f rac t ion o f pla ne t sr ecovered out of 100 0trials as a function ofboth semi- a m p litude,K, a n dm inimum planetmassin Earth m a sses. Ou r met hod reco vers more i njectedpla net s t han t he standard obs e r ving s trategy fo r subgi ant st ars . I tshould be expecte dth at o urmetho d does no trec ov eras ma n y plan ets a s th eth e si mulatio n w h e re n op- mode s w er e pre sent asthis re presentsthe idea li ze d noise floor.
Figu re \[fig:mas sd iff \] ist h e ratioof the number of planet s recove red usin g ou r method, th ree ob ser v ations optim allysa mpl e d (gre e n c urv eFigure \[f i g :tr ec\]) ,to t he numb er of planets reco v ere d using the s tan dard o bs erv i ng str at e gyf o r subgiant star s (blue cu rv e F igure \[fi g :tr ec \]). Ou r metho d rec o vers afactor of ten more p lane t s th an the sta ndard ob serving s t rateg y ,speci fic ally i nthe rang e of p l ane t mas ses be tw een Ne ptune a nd Satur n shown as the dashed m agenta andbla ck linesres p ect ively.
Q uant ifying $\D elt a$t
==== === = ===== ==== = == =
I n thi s se c tion we d e ve lop a p arameteriza t i o n o f the re l ations hipbetween $\Delta t $ , the optimaltime b etw een intr a- night observat ion s, a nd the m as s and radiu s of the s t ar. T his re lation ship ca n be used t o in for m futureobs er v ationsof s u bgiant rad ia l velo city s u rvey s focused on plane t det e c tion. Inthe d ev elopmen t ofthis param eterization , we e xami ned t he rela ti onship s b et ween $\Del t a t$ andother stella rpara met ers su ch a s lumin osit yand effectiv e te m pe ra t ure , bu t det er mine d that ma s s and ra diu s encaps ul ate t he pri m ar y contributi ons to $ \ D elta t$.
T o de v el o p the numer ical m odel, w e ge ne rated a po p u lation of stars wi t h | to_orbital phase_coverage of eight observing_days per_orbital_period. This_phase_coverage is a_guideline and it_should be noted that_increasing or decreasing_the_number, $N_{\rm obs}$, of observations per orbital period will affect the observed RMS by_$\sqrt{\rm_N_{\rm obs}}$._Noise_from_a random Gaussian distribution centered_at 2 m s$^{-1}$ was added_to each_measurement to account for measurement uncertainty. For consistency,_each_comparison data set_(three separated by $\Delta t_{\rm opt}$, three consecutive, and_three consecutive - no p-modes) has_the same number_of_data_points, N = 36.
For_each injected planet described above, we_then computed a Lomb-Scargle Periodogram to_determine the period of each individual data_set. We deemed the planet to_be recovered if the maximum_peak of_the LS periodogram was consistent_within 1% of_the true_period of 10.0_days, i.e..
We show in Figure \[fig:trec\]_the fraction of_planets recovered out of 1000 trials_as_a function of_both_semi-amplitude,_K, and_minimum planet mass_in_Earth masses._Our_method recovers more injected planets than_the_standard observing strategy for subgiant stars. It_should be expected that_our_method does not recover_as many planets as the_the simulation where no p-modes were_present as_this represents_the idealized noise floor.
Figure \[fig:massdiff\] is the ratio of the number_of planets recovered using our method,_three observations optimally sampled_(green curve_Figure_\[fig:trec\]), to the_number_of planets_recovered using the standard observing strategy for_subgiant stars_(blue curve Figure \[fig:trec\]). Our method_recovers a factor of_ten_more planets than the standard observing_strategy, specifically in the range of_planet masses between Neptune and_Saturn_shown_as the dashed magenta and_black lines respectively.
Quantifying $\Delta$t
=====================
In this section_we develop a_parameterization of the relationship between $\Delta t$,_the_optimal time between intra-night observations, and_the_mass and radius of the star._This_relationship_can be used to inform_future observations of subgiant radial velocity_surveys focused on planet detection. In the development of_this parameterization, we_examined the relationships between $\Delta_t$_and_other stellar parameters such as luminosity and effective temperature, but_determined that_mass and radius_encapsulate the primary contributions to $\Delta t$.
To develop the numerical_model, we generated a population of stars_with |
Studies in Advanced Math., 78, Cambridge University Press, Cambridge, 2002.
G. K. Pedersen, [*Extensions of $C\sp
*$-algebras,*]{} In [*Operator algebras and quantum field theory (Rome, 1996)*]{}, 1–35, Int. Press, Cambridge, MA, 1997.
G. K. Pedersen, Pullback and pushout constructions in $C\sp *$-algebra theory, [*J. Funct. Anal.*]{} [**167**]{} (1999), 243–344.
G. Popescu, Von Neumann inequality for $(B(H)^n)_1$, [*Math. Scand.*]{} [**68**]{} (1991), 292–304.
M. K. Royce, [*Ph. D. dissertation*]{}, University of Houston, in preparation.
---
author:
- 'R. van Lieshout'
- 'M. Min'
- 'C. Dominik'
bibliography:
- 'bib\_ads.bib'
subtitle: 'I. Constraints on the dust composition'
title: Dusty tails of evaporating exoplanets
---
Introduction
============
The recently discovered exoplanet candidates b (hereafter KIC 1255b) and show asymmetric transit profiles, that can be explained by the occultation of stars by comet-like tails of dust emanating from evaporating exoplanets [@2012ApJ...752....1R; @2014ApJ...784...40R]. The plausibility of this scenario has been strengthened by quantitative modelling of the transit light curve, and of the Parker-type thermal wind that can eject dust grains from the planetary atmosphere [@2013MNRAS.433.2294P]. Table \[tbl:sys\] lists the basic properties of the two systems.
If the evaporating-planet scenario is correct, KIC 1255b and may provide a rare chance to probe the interiors of small exoplanets. Since the dust tails are thought to originate in the atmospheres of the planets, which are fed by the evaporating planetary surfaces, knowledge of the dust composition would provide information about the composition of the planets. This, in turn, would shed light on the origin and evolutionary history of the planets, and would therefore comprise a valuable constraint for theories of planet formation and evolution.
In this paper, we demonstrate that measurements of the length of a dust tail can be used to | Studies in Advanced Math. , 78, Cambridge University Press, Cambridge, 2002.
G. K. Pedersen, [ * Extensions of $ C\sp
* $ -algebras, * ] { } In [ * Operator algebras and quantum field hypothesis (Rome, 1996) * ] { }, 1–35, Int. Press, Cambridge, MA, 1997.
G. K. Pedersen, Pullback and pushout construction in $ C\sp * $ -algebra theory, [ * J. Funct. Anal. * ] { } [ * * 167 * * ] { } (1999), 243–344.
G. Popescu, Von Neumann inequality for $ (B(H)^n)_1 $, [ * Math. Scand. * ] { } [ * * 68 * * ] { } (1991), 292–304.
M. K. Royce, [ * Ph. D. dissertation * ] { }, University of Houston, in preparation.
---
writer:
-' R. van Lieshout'
-' M. Min'
-' C. Dominik'
bibliography:
-' bib\_ads.bib'
subtitle:' I. Constraints on the debris composition'
title: Dusty buttocks of evaporating exoplanets
---
Introduction
= = = = = = = = = = = =
The recently detect exoplanet candidates b (hereafter KIC 1255b) and show asymmetrical transit profiles, that can be explained by the occultation of asterisk by comet - like tails of dust emanating from evaporating exoplanets [ @2012ApJ... 752.... 1R; @2014ApJ... 784... 40R ]. The plausibility of this scenario has been strengthened by quantitative modeling of the transit light curve, and of the Parker - character thermal wind that can eject debris grains from the erratic atmosphere [ @2013MNRAS.433.2294P ]. board \[tbl: sys\ ] lists the basic properties of the two systems.
If the evaporate - planet scenario is correct, KIC 1255b and may provide a rare chance to probe the interiors of small exoplanets. Since the dust tails are thought to originate in the atmospheres of the planets, which are fed by the evaporating planetary surfaces, knowledge of the dust constitution would provide information about the composing of the planet. This, in turn, would shed light on the origin and evolutionary history of the planets, and would therefore comprise a valuable restraint for theories of planet formation and evolution.
In this paper, we demonstrate that measurements of the length of a dust tail can be used to | Stkdies in Advanced Math., 78, Gambridge University Prxss, Camgridge, 2002.
G. K. Pedersen, [*Extensions of $C\s'
*$-algwbras,*]{} In [*Operator algebras xnd quantlm field rheocy (Rome, 1996)*]{}, 1–35, Int. Prxas, Cambvndge, JW, 1997.
G. N. Pedersen, Pullnack and puvhout construcdiund in $C\sp *$-algebra theory, [*J. Funct. Anaj.*]{} [**167**]{} (1999), 243–344.
G. Pppfscu, Von Neumagn imqquamptn for $(B(H)^n)_1$, [*Math. Scand.*]{} [**68**]{} (1991), 292–304.
M. K. Royce, [*Ph. D. disvertation*]{}, Unifersity of Houston, in prepwratlon.
---
author:
- 'R. van Lleshout'
- 'M. Mun'
- 'C. Eominik'
bibliugraphy:
- 'bib\_ads.bib'
subtijle: 'I. Constraints on the dust cooposicion'
title: Dosfy jails of eva'oratigg exoplanets
---
Introduwtion
============
Thr recently disgoverxd ezoplanet candidates b (hereafter KIC 1255b) anq show asfmjetric transit prifules, jhat wan cw ebplziied by thf orcultation kf stars by comet-like tails of dtwt emanating rrom eda[orating exoplanets [@2012ApJ...752....1R; @2014ApJ...784...40R]. The plauspbiljty of this scenario haw been strengthened bi quantitaeive modelling of the transit light curve, and of dhe Peryer-uyic thdemwl wind that can eject dust grains from the pjznttagy atmosphere [@2013MNRCS.433.2294P]. Table \[tbl:sys\] kidtx the basic prupertiza kf the two systems.
Lf the gvaporqting-plantt scrnario is correct, KIC 1255b and nay provide c rqre chance to probz the interiurs pf smsll exoplanets. Since thz dust tails are hhought tk originate in thd anmos[heres of the planets, whicr are fed by che evapuratong plwnetary sugfaces, knowledge of the fust eompovition woupd provide information about thx composition ox tve planecs. Thix, in turn, woujd shed light pn the jrigiv and evolltionary iistory of tre planets, ang would therehore compwise a vqluable zonstraint for theories of planet formation and evokutkkn.
In this paper, we demonstrate thst oeafugemxnts jx the length of x djxt takl can be uwdd tp | Studies in Advanced Math., 78, Cambridge University 2002. K. Pedersen, of $C\sp *$-algebras,*]{} field (Rome, 1996)*]{}, 1–35, Press, Cambridge, MA, G. K. Pedersen, Pullback and pushout in $C\sp *$-algebra theory, [*J. Funct. Anal.*]{} [**167**]{} (1999), 243–344. G. Popescu, Von inequality for $(B(H)^n)_1$, [*Math. Scand.*]{} [**68**]{} (1991), 292–304. M. K. Royce, [*Ph. D. University Houston, preparation. author: - 'R. van Lieshout' - 'M. Min' - 'C. Dominik' bibliography: - 'bib\_ads.bib' subtitle: 'I. on the dust composition' title: Dusty tails of exoplanets --- Introduction ============ recently discovered exoplanet candidates b KIC and show transit that be explained by occultation of stars by comet-like tails of dust emanating from evaporating exoplanets [@2012ApJ...752....1R; @2014ApJ...784...40R]. The plausibility of scenario has by quantitative of transit curve, and of thermal wind that can eject dust planetary atmosphere [@2013MNRAS.433.2294P]. Table \[tbl:sys\] lists the basic of the systems. If the evaporating-planet scenario is KIC 1255b and may provide a rare chance probe the interiors of small exoplanets. Since the dust tails are thought to originate in of the planets, which fed by the planetary knowledge the composition would information about the composition of the planets. This, in turn, would light on the origin and evolutionary history of the planets, therefore a valuable constraint theories of planet formation evolution. this paper, we demonstrate of length tail be to | Studies in Advanced Math., 78, CambRidge UniveRsity preSs, CAmBridGe, 2002.
G. K. pedersen, [*ExtensIOns oF $C\sp
*$-algebras,*]{} In [*Operator AlgebRaS And qUAnTum fiEld theoRY (ROME, 1996)*]{}, 1–35, InT. PReSs, CAmBRiDge, MA, 1997.
g. K. PEdersen, pullback anD puShOut constructIOnS in $C\sp *$-algeBra Theory, [*J. Funct. anaL.*]{} [**167**]{} (1999), 243–344.
G. PopeScU, VoN neumaNn iNequaLity foR $(b(H)^n)_1$, [*MatH. Scand.*]{} [**68**]{} (1991), 292–304.
M. K. ROyCE, [*Ph. D. diSSertatiON*]{}, unIverSity of Houston, in prEPaRAtion.
---
author:
- 'R. vaN LieshOuT'
- 'm. MIN'
- 'c. DoMinIk'
bibliogrApHy:
- 'bib\_ADs.bib'
suBTiTLE: 'i. CoNStraints on the Dust composiTIon'
Title: DUsTy tAIls of eVaporAtINg eXoplanets
---
InTrodUction
============
The RecentLY discovERed exopLanet cAndIdaTes b (HErEaFteR Kic 1255b) aND sHow ASymMetric trAnSiT profIles, THAT Can bE exPlaiNed by The occultatioN of StarS By cOmet-lIke taIls oF dUst emAnatinG from EvAporating exoplaNets [@2012apJ...752....1R; @2014ApJ...784...40R]. THe pLaUsiBiLity oF This scEnaRio Has been StrengtHEneD bY QUAnTitative modelling oF tHE TrAnsit ligHt curvE, AnD oF The ParkeR-tYpe TherMAL wind That CAn Eject dusT grainS FrOm The planEtAry atmOsPheRe [@2013MnRAS.433.2294P]. tAble \[Tbl:sys\] Lists the Basic PRoperties of the TWo systems.
If thE EvAPOrATing-PlaNet scenario Is coRRect, kIC 1255b ANd May PRovidE a rarE cHAnCE to probe the interiorS oF small ExoplAnets. Since the Dust tails aRE THought to OrigINaTE in the atmospheRes of The planets, WHich are fEd by tHe evaporAting planETAry surfaCes, KnoWleDge OF ThE dust compositION wouLd Provide InfOrmatioN abOut The ComPoSition of tHe planetS. THiS, iN tUrn, Would SHed light On The OrIgiN and eVOlutioNary hIstoRy Of THe pLanets, aND wOULd thErEfOre cOmpRiSe a vaLuabLE coNstrainT for theorIes OF plaNeT fOrmatioN and evolution.
in This paper, wE dEmoNstratE THat measuRements of the length of a duST tail caN be Used tO | Studies in Advanced Math. , 78, Camb ridge Un ive rs ityPres s, Cambridge,2 002.
G. K. Pedersen, [*Ex tensi on s of$ C\ sp
*$ -algebr a s, * ] {}In [ *Op er a to r alg ebr as andquantum fi eld t heory (Rome, 19 96)*]{}, 1 –35 , Int. Press , C ambrid ge , M A , 199 7.
G. K . Pede r sen, P ullback a nd pushou t constr u c ti onsin $C\sp *$-algeb r at heory, [*J. Fu nct. A na l .* ] { } [ **1 67**]{} (1 99 9), 2 4 3–344.G. P o pes c u, Von Neuman n inequalit y fo r $(B( H) ^n) _ 1$, [* Math. S c and .*]{} [**68 **]{ } (1991), 292–3 0 4.
M.K . Royce , [*Ph . D . d isse r ta ti on* ]{ } , U n iv ers i tyof Houst on ,in pr epar a t i o n.
---
aut hor:- 'R. van Lie sho ut'- 'M . Min '
- ' C. D om inik'
bibli ograp hy :
- 'bib\_ads.b ib'subtitle: 'I .Con st raint s on th e d ust compos ition't itl e: D u st y tails of evapora ti n g e xoplanet s
---In tr o duction== === ==== = = =
Th e re c en tly disc overed ex op lanet c an didate sb ( her eafte r KIC 1255b ) and sh ow as y mmetric transi t profiles, th a tc a nb e ex pla ined by the occ u ltat iono fsta r s bycomet -l i ke tails of dust emana ti ng fro m eva porating exop lanets [@2 0 1 2 ApJ...75 2... . 1R ; @2014ApJ...78 4...4 0R]. The p l ausibili ty of this sc enario ha s been str eng the ned by q ua ntitative mod e l ling o f the t ran sit lig htcur ve, an dof the Pa rker-typ eth er ma l w ind t h at can e je ctdu stgrain s fromthe p lane ta ry atm osphere [@ 2 0 13MN RA S. 433. 229 4P ]. Ta ble\ [tb l:sys\] lists th e b a sicpr op ertiesof the two sy st ems.
If t he ev aporat i n g-planet scenario is correct, K I C 1255b an d may pro vide a ra rechance to probethe in terio rs of s malle x op lan et s. Since t h e du st ta il s ar e thoug ht to originate in the atmospheresofthep l an ets , w h ich a r e f e d by the evapora ting plane ta r ysurfaces,k now le dge ofthe dus t com p osition would pr ovide inf or mati o n ab out the co mpositio n of thep lanet s .This, in turn, w oul d she d ligh t on theorigin a nd evo lutio na ry histo ry of the planets, andwouldthere for e compris e a val uable con stra int for th eor ies of p lan e t for mati o nand evolu tion .
In this pa per , we demonstrat e t hat meas ure m ents o f th e length of a dus t tail can be u sedt o | Studies_in Advanced_Math., 78, Cambridge University_Press, Cambridge,_2002.
G._K. Pedersen,_[*Extensions_of $C\sp
*$-algebras,*]{} In_[*Operator algebras and_quantum field theory (Rome,_1996)*]{}, 1–35, Int._Press,_Cambridge, MA, 1997.
G. K. Pedersen, Pullback and pushout constructions in $C\sp *$-algebra theory, [*J._Funct._Anal.*]{} [**167**]{}_(1999),_243–344.
G._Popescu, Von Neumann inequality for_$(B(H)^n)_1$, [*Math. Scand.*]{} [**68**]{} (1991),_292–304.
M. K._Royce, [*Ph. D. dissertation*]{}, University of Houston, in preparation.
---
author:
-_'R._van Lieshout'
- 'M._Min'
- 'C. Dominik'
bibliography:
- 'bib\_ads.bib'
subtitle: 'I. Constraints on the dust_composition'
title: Dusty tails of evaporating exoplanets
---
Introduction
============
The_recently discovered exoplanet_candidates_b_(hereafter KIC 1255b) and show_asymmetric transit profiles, that can be_explained by the occultation of stars_by comet-like tails of dust emanating from_evaporating exoplanets [@2012ApJ...752....1R; @2014ApJ...784...40R]. The plausibility_of this scenario has been_strengthened by_quantitative modelling of the transit_light curve, and_of the_Parker-type thermal wind_that can eject dust grains from_the planetary atmosphere_[@2013MNRAS.433.2294P]. Table \[tbl:sys\] lists the basic properties_of_the two systems.
If_the_evaporating-planet_scenario is_correct, KIC 1255b and_may_provide a_rare_chance to probe the interiors of_small_exoplanets. Since the dust tails are thought_to originate in the_atmospheres_of the planets, which_are fed by the evaporating_planetary surfaces, knowledge of the dust_composition would_provide information_about the composition of the planets. This, in turn, would shed_light on the origin and evolutionary_history of the planets,_and would_therefore_comprise a valuable_constraint_for theories_of planet formation and evolution.
In this paper,_we demonstrate_that measurements of the length of_a dust tail can_be_used to |
1)$, and, by Remark \[1.1nn\], $${\label}{nn1.4}
\varphi(z,(k+1)^+)=\varphi(z, (k+1)^-),\ \varphi'(z,(k+1)^+)=\alpha_{k+1}\varphi'(z, (k+1)^-),$$ the vector $(\varphi, \varphi')^{\top}(z,(k+1)^+)$ may be recovered using the fundamental matrix corresponding to and the matrix of vertex conditions corresponding to as follows $$\begin{aligned}
&\begin{bmatrix}
\varphi(z, (k+1)^+)\\
\varphi'(z, (k+1)^+)
\end{bmatrix}
=
\begin{bmatrix}
\cosh\frac{\sqrt{-z}}{2}& \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\\
{\alpha_{k+1}} {\sqrt{-z}\sinh\frac{\sqrt{-z}}{2}}&{\alpha_{k+1}}\cosh\frac{\sqrt{-z}}{2}
\end{bmatrix}
\begin{bmatrix}
\varphi(z, k+1/2)\\
\varphi'(z, k+1/2)
\end{bmatrix}.
\end{aligned}$$ Combining this and the induction hypothesis we get $$\begin{aligned}
\begin{split}{\label}{1.4}
&\varphi(z, (k+1)^+)=\cosh\left(\frac{\sqrt{-z}}{2}\right)\varphi\left(z, k+\frac12\right)+ \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\varphi'\left(z, k+\frac12\right)\\
&\quad=\left(\frac{e^{\frac{\sqrt{-z}}2}}{2}\frac{e^{\sqrt{-z}(k+1/2)}}{2\sqrt{-z}}+\frac{e^{\frac{\sqrt{-z}}2}}{2\sqrt{-z}}\frac{e^{\sqrt{-z}(k+1/2)}}{2}\right)c_k(1+o(1))\\
&\quad= \frac{c_ke^{\sqrt{-z}(k+1)}}{2\sqrt{-z}}(1+o(1)).
\ | 1)$, and, by Remark \[1.1nn\ ], $ $ { \label}{nn1.4 }
\varphi(z,(k+1)^+)=\varphi(z, (k+1)^-),\ \varphi'(z,(k+1)^+)=\alpha_{k+1}\varphi'(z, (k+1)^-),$$ the vector $ (\varphi, \varphi')^{\top}(z,(k+1)^+)$ may be recovered using the fundamental matrix corresponding to and the matrix of vertex condition correspond to as follows $ $ \begin{aligned }
& \begin{bmatrix }
\varphi(z, (k+1)^+)\\
\varphi'(z, (k+1)^+)
\end{bmatrix }
=
\begin{bmatrix }
\cosh\frac{\sqrt{-z}}{2 } & \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\\
{ \alpha_{k+1 } } { \sqrt{-z}\sinh\frac{\sqrt{-z}}{2}}&{\alpha_{k+1}}\cosh\frac{\sqrt{-z}}{2 }
\end{bmatrix }
\begin{bmatrix }
\varphi(z, k+1/2)\\
\varphi'(z, k+1/2)
\end{bmatrix }.
\end{aligned}$$ Combining this and the induction guess we get $ $ \begin{aligned }
\begin{split}{\label}{1.4 }
& \varphi(z, (k+1)^+)=\cosh\left(\frac{\sqrt{-z}}{2}\right)\varphi\left(z, k+\frac12\right)+ \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\varphi'\left(z, k+\frac12\right)\\
& \quad=\left(\frac{e^{\frac{\sqrt{-z}}2}}{2}\frac{e^{\sqrt{-z}(k+1/2)}}{2\sqrt{-z}}+\frac{e^{\frac{\sqrt{-z}}2}}{2\sqrt{-z}}\frac{e^{\sqrt{-z}(k+1/2)}}{2}\right)c_k(1+o(1))\\
& \quad= \frac{c_ke^{\sqrt{-z}(k+1)}}{2\sqrt{-z}}(1+o(1) ).
\ | 1)$, anf, by Remark \[1.1nn\], $${\label}{nn1.4}
\vavphi(z,(k+1)^+)=\varphi(z, (k+1)^-),\ \rqrphi'(z,(n+1)^+)=\alpha_{i+1}\varphi'(z, (k+1)^-),$$ the vector $(\varphi, \varphi')^{\tlp}(z,(k+1)^+)$ mat be recovered using tfe fundamvntal mateix rorresponding to and the matris of rectex conditions corresponging to as folnods $$\begin{aligned}
&\begin{bmatrix}
\vsrohi(z, (k+1)^+)\\
\darpny'(z, (k+1)^+)
\end{bmatrix}
=
\gegin{bmetrix}
\coxh\frac{\sqrt{-z}}{2}& \frac{\sinh\frac{\sqgt{-z}}{2}}{\sert{-z}}\\
{\alpha_{k+1}} {\sqrt{-z}\sinh\ftzc{\szet{-z}}{2}}&{\alpha_{k+1}}\cosf\frac{\sqrt{-z}}{2}
\end{bjatrix}
\begin{bmatrix}
\vaxphi(z, k+1/2)\\
\vwtphi'(z, k+1/2)
\enq{bmatrix}.
\end{anigned}$$ Vombining this anv thw induction hypothesiv we get $$\begin{aligged}
\yegin{split}{\label}{1.4}
&\carphh(z, (k+1)^+)=\wosh\uwft(\wrad{\sarf{-z}}{2}\righh)\vacphi\left(z, k+\rrac12\right)+ \feac{\sinh\frac{\sqrt{-z}}{2}}{\sqru{-z}}\vwglhi'\left(z, k+\frzc12\righe)\\
&\quad=\left(\frac{e^{\frac{\sqrt{-z}}2}}{2}\frac{e^{\sqrt{-e}(k+1/2)}}{2\sqrf{-z}}+\frac{e^{\frac{\sqrt{-z}}2}}{2\sqrt{-z}}\frqc{e^{\sqrt{-z}(k+1/2)}}{2}\right)c_k(1+o(1))\\
&\quad= \srac{c_ke^{\sqrt{-z}(k+1)}}{2\sqrt{-z}}(1+o(1)).
\ | 1)$, and, by Remark \[1.1nn\], $${\label}{nn1.4} \varphi(z,(k+1)^+)=\varphi(z, (k+1)^-),$$ vector $(\varphi, may be recovered to the matrix of conditions corresponding to follows $$\begin{aligned} &\begin{bmatrix} \varphi(z, (k+1)^+)\\ \varphi'(z, \end{bmatrix} = \begin{bmatrix} \cosh\frac{\sqrt{-z}}{2}& \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\\ {\alpha_{k+1}} {\sqrt{-z}\sinh\frac{\sqrt{-z}}{2}}&{\alpha_{k+1}}\cosh\frac{\sqrt{-z}}{2} \end{bmatrix} \begin{bmatrix} \varphi(z, k+1/2)\\ \varphi'(z, k+1/2) \end{aligned}$$ Combining this and the induction hypothesis we get $$\begin{aligned} \begin{split}{\label}{1.4} &\varphi(z, (k+1)^+)=\cosh\left(\frac{\sqrt{-z}}{2}\right)\varphi\left(z, \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\varphi'\left(z, &\quad=\left(\frac{e^{\frac{\sqrt{-z}}2}}{2}\frac{e^{\sqrt{-z}(k+1/2)}}{2\sqrt{-z}}+\frac{e^{\frac{\sqrt{-z}}2}}{2\sqrt{-z}}\frac{e^{\sqrt{-z}(k+1/2)}}{2}\right)c_k(1+o(1))\\ \frac{c_ke^{\sqrt{-z}(k+1)}}{2\sqrt{-z}}(1+o(1)). | 1)$, and, by Remark \[1.1nn\], $${\label}{nn1.4}
\varphI(z,(k+1)^+)=\varphi(z, (K+1)^-),\ \varpHi'(z,(K+1)^+)=\alPhA_{k+1}\vaRphi'(Z, (k+1)^-),$$ the vector $(\varPHi, \vaRphi')^{\top}(z,(k+1)^+)$ may be recovereD usinG tHE funDAmEntal Matrix cORrESPonDiNg To aNd THe MatriX of Vertex cOnditions cOrrEsPonding to as fOLlOws $$\begin{alIgnEd}
&\begin{bmatrIx}
\vArphi(z, (K+1)^+)\\
\vArpHI'(z, (k+1)^+)
\enD{bmAtrix}
=
\Begin{bMAtrix}
\cOsh\frac{\sqRt{-Z}}{2}& \Frac{\siNH\frac{\sqRT{-Z}}{2}}{\sQrt{-z}}\\
{\Alpha_{k+1}} {\sqrt{-z}\sinh\frAC{\sQRt{-z}}{2}}&{\alpha_{k+1}}\cosh\fRac{\sqrT{-z}}{2}
\ENd{BMAtrIx}
\bEgin{bmatriX}
\vArphi(Z, K+1/2)\\
\varphi'(Z, K+1/2)
\eND{BMatRIx}.
\end{aligned}$$ COmbining thiS And The indUcTioN HypothEsis wE gET $$\beGin{aligned}
\bEgin{Split}{\labeL}{1.4}
&\varphI(Z, (k+1)^+)=\cosh\lEFt(\frac{\sQrt{-z}}{2}\riGht)\VarPhi\lEFt(Z, k+\FraC12\rIGht)+ \FRaC{\siNH\frAc{\sqrt{-z}}{2}}{\sQrT{-z}}\VarphI'\lefT(Z, K+\FRac12\rIghT)\\
&\quaD=\left(\Frac{e^{\frac{\sqrt{-Z}}2}}{2}\frAc{e^{\sQRt{-z}(K+1/2)}}{2\sqrt{-Z}}+\frac{E^{\fraC{\sQrt{-z}}2}}{2\sQrt{-z}}\frAc{e^{\sqRt{-Z}(k+1/2)}}{2}\right)c_k(1+o(1))\\
&\quad= \fRac{c_Ke^{\sqrt{-z}(k+1)}}{2\sQrt{-Z}}(1+o(1)).
\ | 1)$, and, by Remark \[1.1n n\], $${\l abel} {nn 1.4 }\var phi( z,(k+1)^+)=\va r phi( z, (k+1)^-),\ \varphi' (z,(k +1 ) ^+)= \ al pha_{ k+1}\va r ph i ' (z, ( k+ 1)^ -) , $$ thevec tor $(\ varphi, \v arp hi ')^{\top}(z, ( k+ 1)^+)$ may be recovered u sin g thefu nda m ental ma trixcorres p onding to and t he matrix of vert e x c ondi tions correspondi n gt o as follows $ $\begi n{ a li g n ed}
&\be gi n{bma t rix}
\va r phi(z, (k+1)^ +)\\
\v arphi' (z , ( k +1)^+)
\e nd{bmatrix}
=
\ begin{b m atrix} \ cos h\fr a c{ \s qrt {- z }}{ 2 }& \f r ac{ \sinh\fr ac {\ sqrt{ -z}} { 2 } } {\sq rt{ -z}} \\
{\alpha _{k +1}} {\s qrt{- z}\si nh\f ra c{\sq rt{-z} }{2}} &{ \alpha_{k+1}}\c osh\ frac{\sqr t{- z} }{2 } \end {bm atr ix}
\b e gin {b m a t ri x}
\varphi (z , k+ 1/2)\\
\v ar p hi'(z, k +1 /2)
\e nd{b m at rix}.
\ en d{ aligned }$ $ Comb in ing th is an d the induc tion hyp othes i s we get $$\be g in{aligned}
\ begi n{s plit}{\labe l}{1 . 4}
&\v a rphi( z, (k +1 ) ^+ ) =\cosh\left(\frac{\ sq rt{-z} }{2}\ right)\varphi \left(z, k + \ f rac12\ri ght) + \ f rac{\sinh\frac {\sqr t{-z}}{2}} { \sqrt{-z }}\va rphi'\le ft(z, k+\ f r ac12\rig ht) \\ & \quad=\left(\ f r ac{e ^{ \frac{\ sqr t{-z}}2 }}{ 2}\ fra c{e ^{ \sqrt{-z} (k+1/2)} }{ 2\ sq rt {-z }}+\f r ac{e^{\f ra c{\ sq rt{ -z}}2 } }{2\sq rt{-z }}\f ra c{ e ^{\ sqrt{-z } (k + 1 /2)} }{ 2} \rig ht) c_ k(1+o (1)) \ \
&\quad= \ fra c {c_k e^ {\ sqrt{-z }(k+1)}}{2\sq rt {-z}}(1+o( 1) ). \ | 1)$, and,_by Remark_\[1.1nn\], $${\label}{nn1.4}
\varphi(z,(k+1)^+)=\varphi(z, (k+1)^-),\ \varphi'(z,(k+1)^+)=\alpha_{k+1}\varphi'(z,_(k+1)^-),$$ the_vector_$(\varphi, \varphi')^{\top}(z,(k+1)^+)$_may_be recovered using_the fundamental matrix_corresponding to and the_matrix of vertex_conditions_corresponding to as follows $$\begin{aligned}
&\begin{bmatrix}
__ ___ \varphi(z, (k+1)^+)\\
_ _\varphi'(z, (k+1)^+)
_ \end{bmatrix}
__ _ =
_ \begin{bmatrix}
_ __\cosh\frac{\sqrt{-z}}{2}&_\frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\\
_ {\alpha_{k+1}} {\sqrt{-z}\sinh\frac{\sqrt{-z}}{2}}&{\alpha_{k+1}}\cosh\frac{\sqrt{-z}}{2}
_ _ \end{bmatrix}
_ \begin{bmatrix}
_ \varphi(z,_k+1/2)\\
_ _ \varphi'(z, k+1/2)
_ _ _ \end{bmatrix}.
_ _ \end{aligned}$$ Combining this and the_induction_hypothesis we get_$$\begin{aligned}
__ _ __\begin{split}{\label}{1.4}
__ &\varphi(z,_(k+1)^+)=\cosh\left(\frac{\sqrt{-z}}{2}\right)\varphi\left(z,_k+\frac12\right)+ \frac{\sinh\frac{\sqrt{-z}}{2}}{\sqrt{-z}}\varphi'\left(z, k+\frac12\right)\\
_ _&\quad=\left(\frac{e^{\frac{\sqrt{-z}}2}}{2}\frac{e^{\sqrt{-z}(k+1/2)}}{2\sqrt{-z}}+\frac{e^{\frac{\sqrt{-z}}2}}{2\sqrt{-z}}\frac{e^{\sqrt{-z}(k+1/2)}}{2}\right)c_k(1+o(1))\\
_ _ &\quad= \frac{c_ke^{\sqrt{-z}(k+1)}}{2\sqrt{-z}}(1+o(1)).
_ _ \ |
required to be less than 25%$\rts$ if one lepton is identified and less than 20%$\rts$ if more than one lepton is identified. Additionally, the cut on the leading lepton isolation is reinforced: the energy in a $30^{\circ}$ cone around the direction of the electron or muon momentum must be smaller than 2.7 times the electron or muon energy.
### Selection efficiency and background {#selection-efficiency-and-background-2.unnumbered}
The combination of the two selections is chosen according to the $\bar{N}_{95}$ procedure. For $\deltm$ $<$ 11 $\gev$, the logical OR of the two selections is used: the high $\deltm$ selection helps to recover efficiency while leaving the background level unchanged. For $\deltm$ $\geq$ 11 $\gev$, only the high $\deltm$ selection is used. This is in contrast to the situation for the 130–172 GeV data; because the background was low for both the low and the high $\deltm$ selections, the OR of the two selections was optimal for all $\deltm$.
The $\stop \to \rm{b} \ell \snu$ selection efficiencies are given in Figure \[steff\]a. The effect of switching from the OR of the two selections to the high $\deltm$ selection by itself can be seen at $\deltm$ = 11 $\gev$. The background to the low $\deltm$ selection is dominated by $\ggqq$ and has a total expectation of 0.8 events. For the high $\deltm$ selection, the very low expected background (0.1 events expected or ${\sim} 2 \,{\rm fb}$) is dominated by $\ww$ events.
Systematic uncertainties
------------------------
The systematic uncertainties on the $\stop$ and $\sbot$ selection efficiencies originating from the physical processes in the Monte Carlo simulation as well as those related to detector effects are evaluated following the procedure described in Reference [@ALEPH_stop]. The relative uncertainty on the selection efficiency in the case of $\stop \to \rm{c} \chi$ is 13% for low $\Delta m$ and 6% for high $\Delta m$; in the case of $\stop \to \rm{b} \ell \snu$ it is 16% for low $\Delta m$ and 6% for high $\Delta | required to be less than 25%$\rts$ if one lepton is identified and less than 20%$\rts$ if more than one lepton is identified. Additionally, the cut on the lead lepton isolation is reinforce: the energy in a $ 30^{\circ}$ cone around the direction of the electron or muon momentum must be smaller than 2.7 time the electron or muon energy.
# # # Selection efficiency and backdrop { # selection-efficiency-and-background-2.unnumbered }
The combination of the two survival is chosen harmonize to the $ \bar{N}_{95}$ procedure. For $ \deltm$ $ < $ 11 $ \gev$, the logical operating room of the two selections is used: the gamey $ \deltm$ survival helps to recover efficiency while leaving the background level unaltered. For $ \deltm$ $ \geq$ 11 $ \gev$, only the high $ \deltm$ selection is use. This is in contrast to the situation for the 130–172 GeV datum; because the background was abject for both the low and the high $ \deltm$ selections, the OR of the two selections was optimum for all $ \deltm$.
The $ \stop \to \rm{b } \ell \snu$ selection efficiencies are given in Figure \[steff\]a. The effect of switching from the OR of the two survival to the gamey $ \deltm$ selection by itself can be seen at $ \deltm$ = 11 $ \gev$. The background to the low $ \deltm$ selection is dominated by $ \ggqq$ and has a total anticipation of 0.8 events. For the high $ \deltm$ excerpt, the very depleted expected background (0.1 events ask or $ { \sim } 2 \,{\rm fb}$) is predominate by $ \ww$ events.
Systematic uncertainties
------------------------
The taxonomic uncertainties on the $ \stop$ and $ \sbot$ selection efficiencies originating from the physical processes in the Monte Carlo pretense as well as those related to detector effects are evaluated play along the procedure describe in Reference [ @ALEPH_stop ]. The relative uncertainty on the selection efficiency in the case of $ \stop \to \rm{c } \chi$ is 13% for low $ \Delta m$ and 6% for eminent $ \Delta m$; in the case of $ \stop \to \rm{b } \ell \snu$ it is 16% for low $ \Delta m$ and 6% for high $ \Delta | reeuired to be less than 25%$\rus$ if one lepton nw idenvified znd less than 20%$\rts$ if more than one lxptob is udentified. Additionallh, the cut on the oeadmng lepton isolavjon is vzinfodged: tke energy in a $30^{\clrc}$ cone armund the direcdiun of the electron or muon momentum mtst be xmwller than 2.7 tiies uhe elednrin or muon energy.
### Selection erficienby and background {#selection-efficiency-and-bacngroknd-2.unnumbered}
The clmbination if trw two selectkons is chosen accordihg to the $\bar{N}_{95}$ procedure. For $\deutm$ $<$ 11 $\yev$, the logucql LT of the two selebtions is used: the hich $\deltk$ selection hekps to recover efficiency wiile leaving the bachground lavzl unchanged. For $\deltn$ $\teq$ 11 $\ggv$, onny tfw hkgh $\dxltj$ seleftikn is used. This is in contrast to the siuuaeppn for the 130–172 FeV daea; because the background was low for botv tge low and the high $\delrm$ selections, the OR lf the twj selections was optimal for all $\deltm$.
The $\stop \to \rm{b} \xlu \sum$ rwlfction efficiencies are given in Figure \[steff\]a. Fht enfect of switchikg from the OR of yhf yro selections to thz hjgh $\deltm$ selectioj by itfelf xan be setn at $\deltm$ = 11 $\gev$. The background ro the low $\dvltm$ selection is domiuated by $\ggqd$ anc has a total expectation of 0.8 evehts. For the high $\delfo$ selection, the xerj lof expectta background (0.1 evegts expecved ox ${\sim} 2 \,{\ro fb}$) is doiinated by $\ww$ eyants.
Systematic uncfrtaiuties
------------------------
Dhe systemwtic uncertainties on the $\stop$ eid $\sbot$ selecjiot exficiencnes orlginating from ehe physical ptocesses nn the Monte Carmo simunation as wqll as those sglated to detxctor effqcts are evaluagdd following tne procedlrt describee in Reference [@ALEPM_stop]. Fhe relative unetrtqinty on the sekecgiog vffmciensf in the casa of $\stul \to \fm{c} \chi$ is 13% wor kow $\Delta m$ and 6% for higg $\Delta m$; in the cssc of $\stop \to \rm{b} \qll \snu$ it is 16% for low $\Delta m$ wnd 6% hor hijh $\Delya | required to be less than 25%$\rts$ if is and less 20%$\rts$ if more Additionally, cut on the lepton isolation is the energy in a $30^{\circ}$ cone the direction of the electron or muon momentum must be smaller than 2.7 the electron or muon energy. ### Selection efficiency and background {#selection-efficiency-and-background-2.unnumbered} The combination the selections chosen to the $\bar{N}_{95}$ procedure. For $\deltm$ $<$ 11 $\gev$, the logical OR of the two selections used: the high $\deltm$ selection helps to recover while leaving the background unchanged. For $\deltm$ $\geq$ 11 only high $\deltm$ is This in contrast to situation for the 130–172 GeV data; because the background was low for both the low and the $\deltm$ selections, of the selections optimal all $\deltm$. The \rm{b} \ell \snu$ selection efficiencies are \[steff\]a. The effect of switching from the OR the two to the high $\deltm$ selection by can be seen at $\deltm$ = 11 $\gev$. background to the low $\deltm$ selection is dominated by $\ggqq$ and has a total expectation events. For the high selection, the very expected (0.1 expected ${\sim} 2 fb}$) is dominated by $\ww$ events. Systematic uncertainties ------------------------ The systematic on the $\stop$ and $\sbot$ selection efficiencies originating from the in Monte Carlo simulation well as those related detector are evaluated following the in [@ALEPH_stop]. on selection in the case of \to \rm{c} \chi$ is 13% low $\Delta m$ and in the case of $\stop \to \rm{b} \ell it is 16% for low $\Delta m$ 6% for high $\Delta | required to be less than 25%$\rts$ if One lepton iS idenTifIed AnD lesS thaN 20%$\rts$ if more than ONe lePton is identified. AdditiOnallY, tHE cut ON tHe leaDing lepTOn ISOlaTiOn Is rEiNFoRced: tHe eNergy in A $30^{\circ}$ cone aRouNd The direction OF tHe electron Or mUon momentum mUst Be smalLeR thAN 2.7 timeS thE elecTron or MUon eneRgy.
### SelectIoN EfficiENcy and bACKgRounD {#selection-efficieNCy-ANd-background-2.unNumberEd}
tHe COMbiNatIon of the twO sElectIOns is chOSeN ACCorDIng to the $\bar{N}_{95}$ pRocedure. For $\DEltM$ $<$ 11 $\gev$, thE lOgiCAl OR of The twO sELecTions is used: The hIgh $\deltm$ sElectiON helps tO Recover EfficiEncY whIle lEAvInG thE bACkgROuNd lEVel UnchangeD. FOr $\Deltm$ $\Geq$ 11 $\gEV$, ONLy thE hiGh $\deLtm$ seLection is used. thiS is iN ConTrast To the SituAtIon foR the 130–172 Gev data; BeCause the backgroUnd wAs low for bOth ThE loW aNd the HIgh $\delTm$ sEleCtions, tHe OR of tHE twO sELECtIons was optimal for aLl $\DELtM$.
The $\stop \To \rm{b} \eLL \sNu$ SElection EfFicIencIES are gIven IN FIgure \[steFf\]a. The EFfEcT of switChIng froM tHe Or of The twO SeleCtions To the higH $\deltM$ Selection by itsELf can be seen at $\DElTM$ = 11 $\GeV$. the bAckGround to the Low $\dELtm$ sElecTIoN is DOminaTed by $\GgQQ$ aND has a total expectatiOn Of 0.8 evenTs. For The high $\deltm$ sElection, thE VERy low expEcteD BaCKground (0.1 events eXpectEd or ${\sim} 2 \,{\rm fB}$) Is dominaTed by $\Ww$ events.
systematiC UNcertainTieS
------------------------
ThE sySteMATiC uncertaintieS ON the $\StOp$ and $\sbOt$ sElectioN efFicIenCieS oRiginatinG from the PhYsIcAl ProCesseS In the MonTe carLo SimUlatiON as welL as thOse rElAtED to DetectoR EfFECts aRe EvAluaTed FoLlowiNg thE ProCedure dEscribed iN ReFErenCe [@aLePH_stop]. the relative unCeRtainty on tHe SelEction EFFiciency In the case of $\stop \to \rm{c} \chi$ IS 13% for low $\delTa m$ anD 6% for High $\Delta M$; in The casE of $\STop \to \rM{b} \ell \sNu$ it iS 16% fOr lOW $\delta M$ ANd 6% For HiGh $\Delta | required to be less than25%$\rts$if on e l ept on isiden tified and les s tha n 20%$\rts$ if more th an on el epto n i s ide ntified . A d d iti on al ly, t h ecut o n t he lead ing lepton is ol ation is rei n fo rced: theene rgy in a $30 ^{\ circ}$ c one aroun d t he di rectio n of th e electro no r muon momentu m mu st b e smaller than 2. 7 t i mes the electr on ormu o ne n erg y.
### Selec ti on ef f iciency an d b ack g round {#selec tion-effici e ncy -and-b ac kgr o und-2. unnum be r ed}
The combi nati on of the two s e lection s is cho sen ac cor din g to th e$\b ar { N}_ { 95 }$p roc edure. F or $ \delt m$ $ < $ 1 1 $\ gev $, t he lo gical OR of t hetwos ele ction s isused :the h igh $\ deltm $selection helps torecover e ffi ci enc ywhile leavin g t hebackgro und lev e l u nc h a n ge d. For $\deltm$ $\ ge q $ 1 1 $\gev$ , only th eh igh $\de lt m$sele c t ion i s us e d. This is in co n tr as t to th esituat io n f orthe 1 3 0–17 2 GeVdata; be cause the background was low for b o th t he lowand the high $ \del t m$ s elec t io ns, the O R ofth e t w o selections was op ti mal fo r all $\deltm$.
T he $\stop\ t o \rm{b}\ell \s n u$ selection e ffici encies are given in Figu re \[ste ff\]a. Th e effect o f s wit chi ngf r om the OR of th e twose lection s t o the h igh $\ del tm$ s electionby itsel fca nbe se en at $\deltm$ = 11 $ \ge v$. T h e back groun d to t he low $\delt m $s e lect io nis d omi na ted b y $\ g gqq $ and h as a tota l e x pect at io n of 0. 8 events. For t he high $\ de ltm $ sele c t ion, the very low expected back g round ( 0.1 even ts e xpected o r $ {\sim} 2\ ,{\rmfb}$)is do mi nat e d by $ \ w w$ ev en ts.
Syste m a tic unce rt aint ies
--- ------------------ - --
The systemat icunce r t ai nti e so n t he $\s t o p$ and $\sbot$selectionef f ic iencies or i gin at ing fro m the p hysic a l proce sses in t he MonteCa rlos i mul ation as w ell as t hose rela t ed to de tecto r e ffects a reevalu ated f o llo wingthe pr oc eduredescr ib ed in Re ference [@ALEPH_stop].The re lativ e u ncertaint y o n th e selecti on e fficiencyinthe case of $\sto p \t o \ rm{ c } \ch i$ i s 13% forl ow $\ D e lt a m$ and 6% f o r h igh $ \De l ta m$; inthe case of $\sto p \to \rm{b} \e ll \ s n u$iti s 16 %for low $\Delt a m $a n d 6% for h igh $\Delta | required_to be_less than 25%$\rts$ if_one lepton_is_identified and_less_than 20%$\rts$ if_more than one_lepton is identified. Additionally,_the cut on_the_leading lepton isolation is reinforced: the energy in a $30^{\circ}$ cone around the direction_of_the electron_or_muon_momentum must be smaller than_2.7 times the electron or_muon energy.
###_Selection efficiency and background {#selection-efficiency-and-background-2.unnumbered}
The combination of the_two_selections is chosen_according to the $\bar{N}_{95}$ procedure. For $\deltm$ $<$ 11 $\gev$,_the logical OR of the two_selections is used:_the_high_$\deltm$ selection helps to_recover efficiency while leaving the background_level unchanged. For $\deltm$ $\geq$ 11 $\gev$,_only the high $\deltm$ selection is used._This is in contrast to the_situation for the 130–172 GeV_data; because_the background was low for_both the low_and the_high $\deltm$ selections,_the OR of the two selections_was optimal for_all $\deltm$.
The $\stop \to \rm{b} \ell_\snu$_selection efficiencies are_given_in_Figure \[steff\]a. The_effect of switching_from_the OR_of_the two selections to the high_$\deltm$_selection by itself can be seen at_$\deltm$ = 11 $\gev$. The_background_to the low $\deltm$_selection is dominated by $\ggqq$_and has a total expectation of_0.8 events. For_the high_$\deltm$ selection, the very low expected background (0.1 events expected or_${\sim} 2 \,{\rm fb}$) is dominated_by $\ww$ events.
Systematic uncertainties
------------------------
The_systematic uncertainties_on_the $\stop$ and_$\sbot$_selection efficiencies_originating from the physical processes in the_Monte Carlo_simulation as well as those related_to detector effects are_evaluated_following the procedure described in Reference [@ALEPH_stop]._The relative uncertainty on the selection_efficiency in the case of_$\stop_\to_\rm{c} \chi$ is 13% for_low $\Delta m$ and 6% for_high $\Delta m$;_in the case of $\stop \to \rm{b}_\ell_\snu$ it is 16% for low_$\Delta_m$ and 6% for high $\Delta |
accurate forecasting on large highway networks a challenging task. Recently, diffusion convolutional recurrent neural networks (DCRNNs) have achieved state-of-the-art results in traffic forecasting by capturing the spatiotemporal dynamics of the traffic. Despite the promising results, adopting DCRNN for large highway networks still remains elusive because of computational and memory bottlenecks. We present an approach to apply DCRNN for a large highway network. We use a graph-partitioning approach to decompose a large highway network into smaller networks and train them simultaneously on a cluster with graphics processing units (GPU). For the first time, we forecast the traffic of the entire California highway network with 11,160 traffic sensor locations simultaneously. We show that our approach can be trained within 3 hours of wall-clock time using 64 GPUs to forecast speed with high accuracy. Further improvements in the accuracy are attained by including overlapping sensor locations from nearby partitions and finding high-performing hyperparameter configurations for the DCRNN using DeepHyper, a hyperparameter tuning package. We demonstrate that a single DCRNN model can be used to train and forecast the speed and flow simultaneously and the results preserve fundamental traffic flow dynamics. We expect our approach for modeling a large highway network in short wall-clock time as a potential core capability in advanced highway traffic monitoring systems, where forecasts can be used to adjust traffic management strategies proactively given anticipated future conditions.
Introduction
============
In the United States alone, the estimated loss in economic value due to traffic congestion reaches into the tens or hundreds of billions of dollars, impacting not only the productivity lost due to additional travel time but also the additional inefficiencies and energy required for vehicle operation. To address these issues, Intelligent Transportation Systems (ITS) [@bishop2005intelligent] seek to better manage and mitigate congestion and other traffic-related issues via a range of data-informed strategies and highway traffic monitoring systems. Near-term traffic forecasting is a foundational component of these strategies; and accurate forecasting across a range of normal, elevated, and extreme levels of congestion is critical for improved traffic control, routing optimization, probability of incident prediction, and identification of other approaches for handling emerging patterns of congestion [@teklu2007genetic; @tang2005traffic]. Furthermore, these predictions and the related machine learning configurations and weights associated with a highly accurate model can be used to delve more deeply into the dynamics of a particular transportation network in order to identify additional areas of improvement above and beyond | accurate forecasting on large highway networks a challenging undertaking. Recently, dissemination convolutional recurrent neural networks (DCRNNs) have achieved country - of - the - art results in dealings forecasting by capturing the spatiotemporal dynamics of the dealings. Despite the bright results, adopting DCRNN for big highway networks still remains baffling because of computational and memory bottlenecks. We present an overture to apply DCRNN for a large highway net. We use a graph - partitioning approach to disintegrate a large highway network into smaller networks and train them simultaneously on a bunch with graphics processing unit (GPU). For the first time, we forecast the dealings of the entire California highway network with 11,160 traffic sensor locations simultaneously. We show that our overture can be trained within 3 hours of wall - clock time using 64 GPUs to forecast speed with high accuracy. Further improvements in the accuracy are attained by include overlapping sensor location from nearby partition and finding high - performing hyperparameter configurations for the DCRNN use DeepHyper, a hyperparameter tuning software. We demonstrate that a single DCRNN model can be use to train and forecast the speed and flow simultaneously and the results conserve fundamental traffic flow dynamics. We ask our approach for model a large highway network in short wall - clock time as a likely core capability in advanced highway traffic monitoring systems, where forecasts can be used to adjust traffic management strategies proactively give anticipated future condition.
Introduction
= = = = = = = = = = = =
In the United States alone, the estimated personnel casualty in economic value due to traffic congestion strive into the tens or hundred of billions of dollars, impacting not only the productivity lose due to additional travel time but also the additional inefficiencies and energy required for fomite operation. To address these issue, Intelligent Transportation Systems (ITS) [ @bishop2005intelligent ] seek to better manage and mitigate congestion and other traffic - related offspring via a range of data - informed strategies and highway traffic monitoring systems. cheeseparing - condition dealings prediction is a foundational component of these strategies; and accurate prediction across a range of normal, elevated, and extreme levels of congestion is critical for improved dealings control, routing optimization, probability of incident prediction, and identification of early approaches for handle emerging patterns of congestion [ @teklu2007genetic; @tang2005traffic ]. Furthermore, these prediction and the related machine learning shape and weights associated with a highly accurate exemplar can be used to delve more deeply into the dynamics of a particular transportation net in club to identify additional areas of improvement above and beyond | acfurate forecasting on lavge highway networks a rhallenfing tasy. Recently, diffusion convoluvionql rexurrent neural networkr (DCRNNs) jave achuevev state-of-the-art cssults lu trarnic fmcecasting by cakturing the vpatiotemporal dhncmics of the traffic. Despite the proiising tedults, adopting DCRMG fod large highway networks still remzins ellsive because of vomputational and memory blttlfnecks. We present wn approach to wpply DCRNN fur a large highway netsork. We use a graph-partitioning apprpach to dexonpodg a large hijhway getwork into smaller networls and train tmem smmulraneously on a clustec with graphics procgssing unids (GPU). For the firsr rime, fe fmrecxwt ghe tcafric of thx entire Camifornia hithway network with 11,160 twqffic sensor mocatijnf simultaneously. We show that our approdch can be trained within 3 hours of wall-clock tlme using 64 GPUs to forecast speed with high accuracy. Furthar im'ruvenekts kb hhe accuracy are attained by including overla[limg sensor locatijns from nesrhy kartitions and findiuf gigh-performing hypfrparamgter cinfiguratyons for the DCRNN using DeepHyper, a hyperpcraneter tuning packaye. We demonscrate jhat a single DCRNN model can be hsed to traln and foddcast the speed xnd fnow simultaneously and the results 'resexve fundxmenjal trasfic flow fynamlws. We expect our aoproaeh fos modeling a large highway network in shocv wall-clock toma av a poteutial gore capability in advanced hnghway txaffic monitorinf systeks, where fowecasts can ba used to adjnst traffyc mqnagwment sgfategies proacyively gineu anticipqted future conditlons.
Ivfroduction
============
In thz Obited States akond, tre extykated loss it ecunoooc vauue due to brawfic congestion reaches hnto the tens or hundrrdf of biloions of dollars, impavting not only the prodnctivivy losy doe to additional travel time buf also thf aqditional inqffigiensies and euergy required for vehicle operation. To eddress these issues, Inrelligent Transportctlon Systems (MTS) [@bifhop2005intelnigent] seek to bettee manage and mitinate congestion and otger trdffic-gelated issues via a range of data-informed strategies and highway traffic monitoring systema. Nesr-terk crayfic fowecadving is a foundatponal component of these strategmes; and acwuxate forecasting across a ranbe of normal, elgvated, and extreme levels kf congextion is critical for improved yraffic control, routing pptimizarion, lrobabpmivy of incident predivtion, cnd idwntificqtiok of other appfoavhcs for hanflinj ejerging patterns of congestion [@twklu2007gebehic; @tang2005traffic]. Hurthermore, bhtse presictiogs and the relatsd mzfhine learning congigurations anc weights associated wyth a highly axburate model van bg used to delve more deeplf into the dynamics of a particular transportation hetwork ih order bo identify additolnal areas of improvemcnt afove and beyond | accurate forecasting on large highway networks a Recently, convolutional recurrent networks (DCRNNs) have forecasting capturing the spatiotemporal of the traffic. the promising results, adopting DCRNN for highway networks still remains elusive because of computational and memory bottlenecks. We present approach to apply DCRNN for a large highway network. We use a graph-partitioning to a highway into smaller networks and train them simultaneously on a cluster with graphics processing units (GPU). For first time, we forecast the traffic of the California highway network with traffic sensor locations simultaneously. We that approach can trained 3 of wall-clock time 64 GPUs to forecast speed with high accuracy. Further improvements in the accuracy are attained by including sensor locations partitions and high-performing configurations the DCRNN using hyperparameter tuning package. We demonstrate that model can be used to train and forecast speed and simultaneously and the results preserve fundamental flow dynamics. We expect our approach for modeling large highway network in short wall-clock time as a potential core capability in advanced highway systems, where forecasts can used to adjust management proactively anticipated conditions. Introduction In the United States alone, the estimated loss in economic value to traffic congestion reaches into the tens or hundreds of dollars, not only the lost due to additional time also the additional inefficiencies required vehicle these Intelligent Systems (ITS) [@bishop2005intelligent] seek better manage and mitigate congestion other traffic-related issues via and highway traffic monitoring systems. Near-term traffic forecasting a foundational component of these strategies; and forecasting across a range of normal, elevated, and extreme levels of congestion critical for control, routing optimization, probability of incident prediction, and of other approaches for emerging patterns of congestion [@teklu2007genetic; @tang2005traffic]. Furthermore, these predictions the machine learning and weights associated a highly accurate can be used more deeply the of network in order to identify additional of above and beyond | accurate forecasting on largE highway neTworkS a cHalLeNginG tasK. Recently, diffuSIon cOnvolutional recurrent nEural NeTWorkS (dCrNNs) hAve achiEVeD STatE-oF-tHe-aRt REsUlts iN trAffic foRecasting bY caPtUring the spatIOtEmporal dynAmiCs of the traffIc. DEspite ThE prOMisinG reSults, AdoptiNG DCRNN For large hIgHWay netWOrks stiLL ReMainS elusive because of COmPUtational and meMory boTtLEnECKs. WE prEsent an appRoAch to APply DCRnn fOR A LarGE highway netwoRk. We use a graPH-paRtitioNiNg aPProach To decOmPOse A large highwAy neTwork into SmalleR NetworkS And traiN them sImuLtaNeouSLy On A clUsTEr wITh GraPHicS processInG uNits (GpU). FoR THE FirsT tiMe, we ForecAst the traffic Of tHe enTIre califOrnia HighWaY netwOrk witH 11,160 trafFiC sensor locationS simUltaneousLy. WE sHow ThAt our APproacH caN be Trained Within 3 hOUrs Of WALL-cLock time using 64 GPUs tO fOREcAst speed With hiGH aCcURacy. FurtHeR imProvEMEnts iN the ACcUracy are AttainED bY iNcludinG oVerlapPiNg sEnsOr locATionS from nEarby parTitioNS and finding higH-Performing hypERpARAmETer cOnfIgurations fOr thE dCRNn usiNG DEephYper, a HyperPaRAmETer tuning package. We dEmOnstraTe thaT a single DCRNN Model can be USED to train And fOReCAst the speed and Flow sImultaneouSLy and the ResulTs preserVe fundameNTAl traffiC flOw dYnaMicS. wE eXpect our approACH for MoDeling a LarGe highwAy nEtwOrk In sHoRt wall-cloCk time as A pOtEnTiAl cOre caPAbility iN aDvaNcEd hIghwaY TraffiC moniToriNg SySTemS, where fOReCASts cAn Be Used To aDjUst trAffiC ManAgement StrategieS prOActiVeLy Given anTicipated futuRe Conditions.
inTroDuctioN
============
iN the UnitEd States alone, the estimatED loss in EcoNomic ValuE due to traFfiC congeStiON reachEs into The teNs Or hUNDreds OF BiLliOnS of dollars, IMPacTing nOt Only The prodUctivity lost due to aDDitIonal travel tiMe bUt alSO ThE adDItIOnaL iNEffICIencies and energY required fOr VEhIcle operatIOn. TO aDdress tHese issUes, InTElligenT TransporTation SysTeMs (ITs) [@BIshOp2005intelligEnt] seek tO better maNAge anD MiTigatE coNgestiOn And Other TraffiC-RelAted iSsues vIa A range Of datA-iNformed sTrategies and highway trafFic monItoriNg sYstems. NeaR-teRM trAffic foreCastIng is a founDatIonAl comPonENt of tHese STrAteGIes; anD accURate forecAStIng ACRoSs a range of nORMAl, eLevatEd, aND extreMe leVels of congestion iS Critical for impRoveD TRafFic COntrOl, Routing optimizAtiOn, PRObabilitY oF incident prEdiction, AnD IdentIficatIon of oTher appROAcHEs for hAndlIng Emerging pAttErNS of congEsTiON [@teklu2007GeneTiC; @tang2005tRaffic]. fUrthERMore, these predictIons aND The reLAteD machInE learniNG conFigurationS and weights AssociAted With a Highly aCcUrate mOdeL cAn be used to DElve more dEeply Into the DyNamiCs oF a partIculAR TransPortAtIon Network in ORDeR To IdENtiFy adDitioNaL areAs of improVEment aboVe aND beyond | accurate forecasting on l arge highw ay ne two rks a cha llen ging task. Rec e ntly , diffusion convolutio nal r ec u rren t n eural networ k s( D CRN Ns )hav ea ch ieved st ate-of- the-art re sul ts in trafficf or ecasting b y c apturing the sp atiote mp ora l dyna mic s ofthe tr a ffic.Despite t he promis i ng resu l t s, ado pting DCRNN for l a rg e highway netwo rks st il l r e m ain s e lusive bec au se of computa t io n a l an d memory bottl enecks. Wep res ent an a ppr o ach to appl yD CRN N for a lar ge h ighway ne twork. We usea graph- partit ion ing app r oa ch to d e com p os e a lar ge highw ay n etwor k in t o s mall ernetw orksand train the m s imul t ane ously on a clu st er wi th gra phics p rocessing units (GP U). For t hefi rst t ime,w e fore cas t t he traf fic oft heen t i r eCalifornia highway n e t wo rk with11,160 tr af f ic senso rloc atio n s simu ltan e ou sly. Weshow t h at o ur appr oa ch can b e t rai ned w i thin 3 hou rs of wa ll-cl o ck time using6 4 GPUs to for e ca s t s p eedwit h high accu racy . Fur ther im pro v ement s inth e a c curacy are attained b y incl uding overlappingsensor loc a t i ons from nea r by partitions and find ing high-p e rforming hype rparamet er config u r ations f orthe DC RNN u si ng DeepHyper, a hyp er paramet ertuningpac kag e.Wede monstrate that asi ng le D CRN N mod e l can be u sed t o t raina nd for ecast the s pe e d a nd flow si m u ltan eo us ly a ndth e res ults pre serve f undamenta l t r affi cfl ow dyna mics. We expe ct our appro ac h f or mod e l ing a la rge highway network ins hort wa ll- clock tim e as a po ten tial c ore capabi lity i n adv an ced h ighwa y tr aff ic monitorin g sys tems, w here foreca sts can be used to adj ust traffic m ana geme n t s tra t eg i espr o act i v ely given antic ipated fut ur e c onditions.
In tr oductio n
===== ===== = =
In t he United States a lo ne,t h e e stimated l oss in e conomic v a lue d u eto tr aff ic con ge sti on re achesi nto thetens o rhundre ds of b illionsof dollars, impacting n ot onl y the pr oductivit y l o stdue to ad diti onal trave l t ime butals o theaddi t io nal ineff icie n cies ande ne rgy r eq uired for v e h i cle oper ati o n. Toaddr ess these issues, Intelligent Tr ansp o r tat ion Syst em s (ITS) [@bish op2 00 5 i ntellige nt ] seek to b etter ma na g e and mitig ate co ngestio n an d other tra ffi c-related is su e s via a r an g e of d ata- in formed strat e gies a nd highway traff ic mo n i torin g sy stems .Near-te r m tr affic fore casting isa foun dati onalcompone nt of th ese s trategies; and accur ate f orecast in g ac ros s a ra ngeo f norm al,el eva ted, ande x tr e me l e vel s of cong es tion is criti c al for i mpr o ved tra ff icc o ntrol, ro u t ing optimi zat ion,p r obabilityo f in c id e nt pr edicti on, an d ident i fic at ion ofoth e r approach es for ha n dli ng eme rging pa tt erns o f c on gestion [@teklu 2007gene tic; @ta n g 2005tr aff i c].F urther mo re, these pr edict io n s and t h e re lated m a chi nelea rnin g c onfi g urations and weights ass ocia ted w i tha hig hly a cc urate mo d el c an be us e d to delve mo re deeply intot hed ynam ics of a particul art ransp orta tio n ne two rk i n orde r toi de ntif y addi tion al a reas of i mp rovemen t abo v e a nd b eyond | accurate_forecasting on_large highway networks a_challenging task._Recently,_diffusion convolutional_recurrent_neural networks (DCRNNs)_have achieved state-of-the-art_results in traffic forecasting_by capturing the_spatiotemporal_dynamics of the traffic. Despite the promising results, adopting DCRNN for large highway networks_still_remains elusive_because_of_computational and memory bottlenecks. We_present an approach to apply_DCRNN for_a large highway network. We use a graph-partitioning_approach_to decompose a_large highway network into smaller networks and train them_simultaneously on a cluster with graphics_processing units (GPU)._For_the_first time, we forecast_the traffic of the entire California_highway network with 11,160 traffic sensor_locations simultaneously. We show that our approach_can be trained within 3 hours_of wall-clock time using 64_GPUs to_forecast speed with high accuracy._Further improvements in_the accuracy_are attained by_including overlapping sensor locations from nearby_partitions and finding_high-performing hyperparameter configurations for the DCRNN_using_DeepHyper, a hyperparameter_tuning_package._We demonstrate_that a single_DCRNN_model can_be_used to train and forecast the_speed_and flow simultaneously and the results preserve_fundamental traffic flow dynamics._We_expect our approach for_modeling a large highway network_in short wall-clock time as a_potential core_capability in_advanced highway traffic monitoring systems, where forecasts can be used to_adjust traffic management strategies proactively given_anticipated future conditions.
Introduction
============
In the_United States_alone,_the estimated loss_in_economic value_due to traffic congestion reaches into the_tens or_hundreds of billions of dollars, impacting_not only the productivity_lost_due to additional travel time but_also the additional inefficiencies and energy_required for vehicle operation. To_address_these_issues, Intelligent Transportation Systems (ITS)_[@bishop2005intelligent] seek to better manage and_mitigate congestion and_other traffic-related issues via a range of_data-informed_strategies and highway traffic monitoring systems._Near-term_traffic forecasting is a foundational component_of_these_strategies; and accurate forecasting across_a range of normal, elevated, and_extreme levels of congestion is critical for improved traffic_control, routing optimization,_probability of incident prediction, and_identification_of_other approaches for handling emerging patterns of congestion [@teklu2007genetic; @tang2005traffic]._Furthermore, these_predictions and the_related machine learning configurations and weights associated with a highly_accurate model can be used to delve_more deeply into the dynamics of a particular transportation network in_order to identify additional areas of improvement above_and beyond |
) $ with $\langle f,\,f\rangle =\omega_{+}(f) \omega_{-}(f)$ such that for even $m$ and odd $\ell$ with $1\leq m,\ell\leq k-1$,
1. $$\frac{\Lambda(f,m)}{w_{+}(f)}, \frac{\Lambda(f,\ell)}{w_{-}(f)} \in \mathbb{Q}(f),$$
2. for each $\sigma\in\text{Gal}(\bar{\mathbb{Q}}/\mathbb{Q})$, $$\left(\frac{\Lambda(f,m)}{\omega_+(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,m)}{\omega_+(f^\sigma)},\quad \left(\frac{\Lambda(f,\ell)}{\omega_-(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,\ell)}{\omega_-(f^\sigma)}.$$
<!-- -->
1. The above theorem is an analogous result of that for elliptic modular forms proved in [@KZ] (Theorem in page 202). We can also extend the rationality easily to arbitrary L-values as did in Theorem 8.3 of [@CD2].
2. The above theorem is a special case of Shimura theorem (Theorem 4.3 in [@Sh]) by taking $n=2$, $\psi=1$, and $k_1=k_2=k.$
Proofs
======
We need the following multi variable Lipschitz summation formula.
\[lip\] ([**multi-variable Lipschitz summation formula**]{}) Assume that $\im (s)>2$. For $z\in \mathbb{H}^2$, $$\sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}
=\frac{(2\pi)^{2s}}{e^{\pi is}\Gamma(s)^{2}D^{1/2}}\sum_{\xi\in\mathfrak d^{-1}_+}N(\xi)^{s-1}\exp(2\pi i\textrm{Tr}(\xi z)),$$
By the multi-index notation, $$\sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}= \sum_{x\in\mathcal O}
N(z+x)^{-s}=\sum_{x\in\mathcal O}
(z_1+x)^{-s}(z_2+x')^{-s | ) $ with $ \langle f,\,f\rangle = \omega_{+}(f) \omega_{-}(f)$ such that for even $ m$ and odd $ \ell$ with $ 1\leq m,\ell\leq k-1 $,
1. $ $ \frac{\Lambda(f, m)}{w_{+}(f) }, \frac{\Lambda(f,\ell)}{w_{-}(f) } \in \mathbb{Q}(f),$$
2. for each $ \sigma\in\text{Gal}(\bar{\mathbb{Q}}/\mathbb{Q})$, $ $ \left(\frac{\Lambda(f, m)}{\omega_+(f)}\right)^\sigma=\frac{\Lambda(f^\sigma, m)}{\omega_+(f^\sigma)},\quad \left(\frac{\Lambda(f,\ell)}{\omega_-(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,\ell)}{\omega_-(f^\sigma)}.$$
<! -- -- >
1. The above theorem is an analogous result of that for elliptic modular forms testify in [ @KZ ] (Theorem in page 202). We can besides extend the rationality easily to arbitrary lambert - value as did in Theorem 8.3 of [ @CD2 ].
2. The above theorem is a special character of Shimura theorem (Theorem 4.3 in [ @Sh ]) by fill $ n=2 $, $ \psi=1 $, and $ k_1 = k_2 = k.$
Proofs
= = = = = =
We necessitate the following multi variable Lipschitz sum formula.
\[lip\ ] ([ * * multi - variable Lipschitz summation formula * * ] { }) wear that $ \im (s)>2$. For $ z\in \mathbb{H}^2 $, $ $ \sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1 } }
= \frac{(2\pi)^{2s}}{e^{\pi is}\Gamma(s)^{2}D^{1/2}}\sum_{\xi\in\mathfrak d^{-1}_+}N(\xi)^{s-1}\exp(2\pi i\textrm{Tr}(\xi z)),$$
By the multi - index notation, $ $ \sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}= \sum_{x\in\mathcal O }
N(z+x)^{-s}=\sum_{x\in\mathcal O }
(z_1+x)^{-s}(z_2+x')^{-s | ) $ wlth $\langle f,\,f\rangle =\omegx_{+}(f) \omega_{-}(f)$ such jhqt for even $j$ and oda $\ell$ with $1\leq m,\ell\leq k-1$,
1. $$\frec{\Lanbda(f,n)}{w_{+}(f)}, \frac{\Lambda(f,\ell)}{w_{-}(f)} \kn \mathbb{E}(f),$$
2. for wach $\wigma\in\texv{Fal}(\bar{\mathbb{Q}}/\jwthby{Q})$, $$\left(\frac{\Lambds(f,m)}{\omega_+(f)}\rhght)^\sigma=\frac{\Ldmcdc(f^\sigma,m)}{\omega_+(f^\sigma)},\quad \left(\frac{\Lambqa(f,\ell)}{\okeha_-(f)}\right)^\sigma=\ftac{\Lakfda(f^\apgna,\ell)}{\omega_-(f^\sigma)}.$$
<!-- -->
1. The above fheorem is an analogpus result of that for elllptif modular forms prlved in [@KZ] (Jgeowwm in page 202). De can also extend the rationality easily to arbitrarh L-vakues as die un Hveorem 8.3 of [@RD2].
2. Thv above theorck is a speciak case of Shimmra tieorwm (Theorem 4.3 in [@Sh]) by vaking $n=2$, $\psi=1$, and $k_1=k_2=k.$
Kroofs
======
We naeb the following multi vqriabne Lhpscfutz sujmetikn forlule.
\[lip\] ([**multi-vzriable Lipwchitz summation fotmtoa**]{}) Assume thaf $\im (s)>2$. Fjr $z\in \mathbb{H}^2$, $$\sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}
=\xrad{(2\pi)^{2s}}{e^{\pi is}\Gamma(s)^{2}D^{1/2}}\sum_{\xi\ib\mathfrak d^{-1}_+}N(\xi)^{s-1}\exp(2\pi l\textrm{Tr}(\vi z)),$$
By the multi-index notation, $$\sum_{x\in\mathcal O}(z+x)^{-s\kathbh{1}}= \rum_{r\lk\matfxap O}
N(z+x)^{-s}=\sum_{x\in\mathcal O}
(z_1+x)^{-s}(z_2+x')^{-s | ) $ with $\langle f,\,f\rangle =\omega_{+}(f) \omega_{-}(f)$ for $m$ and $\ell$ with $1\leq \in 2. for each $$\left(\frac{\Lambda(f,m)}{\omega_+(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,m)}{\omega_+(f^\sigma)},\quad \left(\frac{\Lambda(f,\ell)}{\omega_-(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,\ell)}{\omega_-(f^\sigma)}.$$ <!-- 1. The above theorem is an result of that for elliptic modular forms proved in [@KZ] (Theorem in page We can also extend the rationality easily to arbitrary L-values as did in 8.3 [@CD2]. The theorem is a special case of Shimura theorem (Theorem 4.3 in [@Sh]) by taking $n=2$, $\psi=1$, $k_1=k_2=k.$ Proofs ====== We need the following multi Lipschitz summation formula. \[lip\] Lipschitz summation formula**]{}) Assume that (s)>2$. $z\in \mathbb{H}^2$, O}(z+x)^{-s\mathbf{1}} is}\Gamma(s)^{2}D^{1/2}}\sum_{\xi\in\mathfrak i\textrm{Tr}(\xi z)),$$ By multi-index notation, $$\sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}= \sum_{x\in\mathcal O} N(z+x)^{-s}=\sum_{x\in\mathcal O} (z_1+x)^{-s}(z_2+x')^{-s | ) $ with $\langle f,\,f\rangle =\omega_{+}(f) \oMega_{-}(f)$ such tHat foR evEn $m$ AnD odd $\Ell$ wIth $1\leq m,\ell\leq k-1$,
1. $$\FRac{\LAmbda(f,m)}{w_{+}(f)}, \frac{\Lambda(f,\elL)}{w_{-}(f)} \in \MaTHbb{Q}(F),$$
2. FoR each $\Sigma\in\TExT{gAl}(\bAr{\MaThbB{Q}}/\MAtHbb{Q})$, $$\lEft(\Frac{\LamBda(f,m)}{\omega_+(F)}\riGhT)^\sigma=\frac{\LaMBdA(f^\sigma,m)}{\omEga_+(F^\sigma)},\quad \leFt(\fRac{\LamBdA(f,\eLL)}{\omegA_-(f)}\rIght)^\sIgma=\frAC{\LambdA(f^\sigma,\elL)}{\oMEga_-(f^\siGMa)}.$$
<!-- -->
1. The abOVE tHeorEm is an analogous reSUlT Of that for ellipTic modUlAR fORMs pRovEd in [@KZ] (TheoReM in paGE 202). We can aLSo EXTEnd THe rationality Easily to arbITraRy L-valUeS as DId in ThEorem 8.3 Of [@cd2].
2. ThE above theorEm is A special cAse of SHImura thEOrem (TheOrem 4.3 in [@sh]) bY taKing $N=2$, $\PsI=1$, aNd $k_1=K_2=k.$
pRooFS
======
WE neED thE followiNg MuLti vaRiabLE lIPschItz SummAtion Formula.
\[lip\] ([**mulTi-vAriaBLe LIpschItz suMmatIoN formUla**]{}) AssUme thAt $\Im (s)>2$. For $z\in \mathbb{h}^2$, $$\sum_{X\in\mathcaL O}(z+X)^{-s\MatHbF{1}}
=\frac{(2\PI)^{2s}}{e^{\pi iS}\GaMma(S)^{2}D^{1/2}}\sum_{\xi\In\mathfRAk d^{-1}_+}n(\xI)^{S-1}\EXp(2\Pi i\textrm{Tr}(\xi z)),$$
By thE mULTi-Index notAtion, $$\sUM_{x\In\MAthcal O}(z+X)^{-s\MatHbf{1}}= \sUM_{X\in\maThcaL o}
N(Z+x)^{-s}=\sum_{x\iN\mathcAL O}
(Z_1+x)^{-S}(z_2+x')^{-s | ) $ with $\langle f,\,f\ra ngle =\ome ga_{+ }(f ) \ om ega_ {-}( f)$ such thatf or e ven $m$ and odd $\ell$ with $ 1 \leq m, \ell\ leq k-1 $ ,1 . $$ \f rac {\ L am bda(f ,m) }{w_{+} (f)}, \fra c{\ La mbda(f,\ell) } {w _{-}(f)} \in \mathbb{Q}( f), $$
2. for each$\s igma\ in\tex t {Gal}( \bar{\mat hb b {Q}}/\ m athbb{Q } ) $, $$\ left(\frac{\Lambd a (f , m)}{\omega_+(f )}\rig ht ) ^\ s i gma =\f rac{\Lambd a( f^\si g ma,m)}{ \ om e g a _+( f ^\sigma)},\qu ad \left(\f r ac{ \Lambd a( f,\ e ll)}{\ omega _- ( f)} \right)^\si gma= \frac{\La mbda(f ^ \sigma, \ ell)}{\ omega_ -(f ^\s igma ) }. $$
< !- - -- >
1.The above t he or em is ana n a l ogou s r esul t ofthat for elli pti c mo d ula r for ms pr oved i n [@K Z] (Th eorem i n page 202). We can also ext end t hera tiona l ity ea sil y t o arbit rary L- v alu es a s d id in Theorem 8.3of [ @C D2].
2. Thea bo ve theoremis aspec i a l cas e of Sh imura th eorem( Th eo rem 4.3 i n [@Sh ]) by ta king$ n=2$ , $\ps i=1$, an d $k_ 1 =k_2=k.$
Proo f s
======
Wen ee d th e fol low ing multi v aria b le L ipsc h it z s u mmati on fo rm u la .
\[lip\] ([**multi -v ariabl e Lip schitz summat ion formul a * * ]{}) Ass umet ha t $\im (s)>2$.For $ z\in \math b b{H}^2$, $$\s um_{x\in \mathcalO } (z+x)^{- s\m ath bf{ 1}} = \f rac{(2\pi)^{2 s } }{e^ {\ pi is}\ Gam ma(s)^{ 2}D ^{1 /2} }\s um _{\xi\in\ mathfrak d ^{ -1 }_ +}N (\xi) ^ {s-1}\ex p( 2\p ii\t extrm { Tr}(\x i z)) ,$$
B yt hemulti-i n de x nota ti on , $$ \su m_ {x\in \mat h cal O}(z+x )^{-s\mat hbf { 1}}= \ su m_{x\in \mathcal O}
N (z +x)^{-s}=\ su m_{ x\in\m a t hcal O}(z_1+x)^{-s}(z_2+x')^{- s | ) $_with $\langle_f,\,f\rangle =\omega_{+}(f) \omega_{-}(f)$ such_that for_even_$m$ and_odd_$\ell$ with $1\leq_m,\ell\leq k-1$,
1. _$$\frac{\Lambda(f,m)}{w_{+}(f)}, \frac{\Lambda(f,\ell)}{w_{-}(f)} \in_\mathbb{Q}(f),$$
2. for_each_$\sigma\in\text{Gal}(\bar{\mathbb{Q}}/\mathbb{Q})$, $$\left(\frac{\Lambda(f,m)}{\omega_+(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,m)}{\omega_+(f^\sigma)},\quad \left(\frac{\Lambda(f,\ell)}{\omega_-(f)}\right)^\sigma=\frac{\Lambda(f^\sigma,\ell)}{\omega_-(f^\sigma)}.$$
<!-- -->
1. The above theorem is an analogous result of that_for_elliptic modular_forms_proved_in [@KZ] (Theorem in page_202). We can also extend_the rationality_easily to arbitrary L-values as did in Theorem_8.3_of [@CD2].
2. _The above theorem is a special case of Shimura_theorem (Theorem 4.3 in [@Sh]) by_taking $n=2$, $\psi=1$,_and_$k_1=k_2=k.$
Proofs
======
We_need the following multi_variable Lipschitz summation formula.
\[lip\] ([**multi-variable Lipschitz_summation formula**]{}) Assume that $\im (s)>2$._For $z\in \mathbb{H}^2$, $$\sum_{x\in\mathcal O}(z+x)^{-s\mathbf{1}}
=\frac{(2\pi)^{2s}}{e^{\pi is}\Gamma(s)^{2}D^{1/2}}\sum_{\xi\in\mathfrak d^{-1}_+}N(\xi)^{s-1}\exp(2\pi_i\textrm{Tr}(\xi z)),$$
By the multi-index notation, $$\sum_{x\in\mathcal_O}(z+x)^{-s\mathbf{1}}= \sum_{x\in\mathcal O}
N(z+x)^{-s}=\sum_{x\in\mathcal O}
(z_1+x)^{-s}(z_2+x')^{-s |
17 / 0.54
26 $CH_2C^{+}$? - 0.87 / 0.27 0.68 / 0.23 0.65 / 0.45 0.84 / 0.33 - 0.77 / 0.58
28 $CO^+$ 1.98 / 0.14 2.72 / 0.30 2.26 / 0.26 2.13 / 0.26 2.69 / 0.26 3.96 / 0.10 3.36 / 0.42
29 $COH^+$; $HCO^+$ 6.13 / 0.16 6.95 / 0.25 6.32 / 0.21 6.42 / 0.21 6.66 / 0.29 7.15 / 0.17 7.06 / 0.40
30 $CH_2O^+$; $CH3COOH^{++}$? 1.33 / 0.23 - 0.27 / 0.24 0.55 / 0.05 0.22 / 0.45 - 0.21 / 0.16
31 $CH_3O^{+}$? - 1.27 / 0.23 1.42 / 0.33 1.41 / 0.27 1.29 / 0.23 1.43 / 0.07 0.92 / 0.27
40 $CCO^+$ 0.37 / 0.01 0.56 / 0.18 0.54 / 0.21 0.52 / 0.22 0.64 / 0.15 1.15 / 0.18 0.71 / 0.18
41 $CHCO^+$ 1.17 / 0.12 1.67 / 0.12 1.41 / 0.14 1.39 / 0.18 1.49 / 0.12 1.78 / 0.05 1.59 / 0.24
42 $CH_2CO^+$ 3.47 / 0.09 3.69 / 0.14 3.70 / 0.14 3.63 / 0.11 3.56 / 0.14 3.21 / 0.07 | 17 / 0.54
26 $ CH_2C^{+}$? - 0.87 / 0.27 0.68 / 0.23 0.65 / 0.45 0.84 / 0.33 - 0.77 / 0.58
28 $ CO^+$ 1.98 / 0.14 2.72 / 0.30 2.26 / 0.26 2.13 / 0.26 2.69 / 0.26 3.96 / 0.10 3.36 / 0.42
29 $ COH^+$; $ HCO^+$ 6.13 / 0.16 6.95 / 0.25 6.32 / 0.21 6.42 / 0.21 6.66 / 0.29 7.15 / 0.17 7.06 / 0.40
30 $ CH_2O^+$; $ CH3COOH^{++}$? 1.33 / 0.23 - 0.27 / 0.24 0.55 / 0.05 0.22 / 0.45 - 0.21 / 0.16
31 $ CH_3O^{+}$? - 1.27 / 0.23 1.42 / 0.33 1.41 / 0.27 1.29 / 0.23 1.43 / 0.07 0.92 / 0.27
40 $ CCO^+$ 0.37 / 0.01 0.56 / 0.18 0.54 / 0.21 0.52 / 0.22 0.64 / 0.15 1.15 / 0.18 0.71 / 0.18
41 $ CHCO^+$ 1.17 / 0.12 1.67 / 0.12 1.41 / 0.14 1.39 / 0.18 1.49 / 0.12 1.78 / 0.05 1.59 / 0.24
42 $ CH_2CO^+$ 3.47 / 0.09 3.69 / 0.14 3.70 / 0.14 3.63 / 0.11 3.56 / 0.14 3.21 / 0.07 | 17 / 0.54
26 $CH_2C^{+}$? - 0.87 / 0.27 0.68 / 0.23 0.65 / 0.45 0.84 / 0.33 - 0.77 / 0.58
28 $RO^+$ 1.98 / 0.14 2.72 / 0.30 2.26 / 0.26 2.13 / 0.26 2.69 / 0.26 3.96 / 0.10 3.36 / 0.42
29 $GJH^+$; $HDL^+$ 6.13 / 0.16 6.95 / 0.25 6.32 / 0.21 6.42 / 0.21 6.66 / 0.29 7.15 / 0.17 7.06 / 0.40
30 $CH_2O^+$; $CH3COOH^{++}$? 1.33 / 0.23 - 0.27 / 0.24 0.55 / 0.05 0.22 / 0.45 - 0.21 / 0.16
31 $CH_3O^{+}$? - 1.27 / 0.23 1.42 / 0.33 1.41 / 0.27 1.29 / 0.23 1.43 / 0.07 0.92 / 0.27
40 $CCO^+$ 0.37 / 0.01 0.56 / 0.18 0.54 / 0.21 0.52 / 0.22 0.64 / 0.15 1.15 / 0.18 0.71 / 0.18
41 $CHCO^+$ 1.17 / 0.12 1.67 / 0.12 1.41 / 0.14 1.39 / 0.18 1.49 / 0.12 1.78 / 0.05 1.59 / 0.24
42 $CH_2CO^+$ 3.47 / 0.09 3.69 / 0.14 3.70 / 0.14 3.63 / 0.11 3.56 / 0.14 3.21 / 0.07 | 17 / 0.54 26 $CH_2C^{+}$? - 0.87 0.68 0.23 0.65 0.45 0.84 / 28 1.98 / 0.14 / 0.30 2.26 0.26 2.13 / 0.26 2.69 / 3.96 / 0.10 3.36 / 0.42 29 $COH^+$; $HCO^+$ 6.13 / 0.16 6.95 0.25 6.32 / 0.21 6.42 / 0.21 6.66 / 0.29 7.15 / 0.17 / 30 $CH3COOH^{++}$? / 0.23 - 0.27 / 0.24 0.55 / 0.05 0.22 / 0.45 - 0.21 / 0.16 $CH_3O^{+}$? - 1.27 / 0.23 1.42 / 0.33 / 0.27 1.29 / 1.43 / 0.07 0.92 / 40 0.37 / 0.56 0.18 / 0.21 0.52 0.22 0.64 / 0.15 1.15 / 0.18 0.71 / 0.18 41 $CHCO^+$ 1.17 / 0.12 1.67 / 1.41 / / 0.18 / 1.78 0.05 1.59 / $CH_2CO^+$ 3.47 / 0.09 3.69 / 0.14 3.63 / 0.11 3.56 / 0.14 3.21 0.07 | 17 / 0.54
26 $CH_2C^{+}$? - 0.87 / 0.27 0.68 / 0.23 0.65 / 0.45 0.84 / 0.33 - 0.77 / 0.58
28 $CO^+$ 1.98 / 0.14 2.72 / 0.30 2.26 / 0.26 2.13 / 0.26 2.69 / 0.26 3.96 / 0.10 3.36 / 0.42
29 $COH^+$; $HCO^+$ 6.13 / 0.16 6.95 / 0.25 6.32 / 0.21 6.42 / 0.21 6.66 / 0.29 7.15 / 0.17 7.06 / 0.40
30 $CH_2O^+$; $CH3COOH^{++}$? 1.33 / 0.23 - 0.27 / 0.24 0.55 / 0.05 0.22 / 0.45 - 0.21 / 0.16
31 $CH_3O^{+}$? - 1.27 / 0.23 1.42 / 0.33 1.41 / 0.27 1.29 / 0.23 1.43 / 0.07 0.92 / 0.27
40 $CCo^+$ 0.37 / 0.01 0.56 / 0.18 0.54 / 0.21 0.52 / 0.22 0.64 / 0.15 1.15 / 0.18 0.71 / 0.18
41 $CHCO^+$ 1.17 / 0.12 1.67 / 0.12 1.41 / 0.14 1.39 / 0.18 1.49 / 0.12 1.78 / 0.05 1.59 / 0.24
42 $CH_2CO^+$ 3.47 / 0.09 3.69 / 0.14 3.70 / 0.14 3.63 / 0.11 3.56 / 0.14 3.21 / 0.07 | 17 / 0.54
26 $CH_2C ^{+}$? - 0.8 7 / 0 .27 0.68 / 0.23 0. 65 /0. 4 5 0 .8 4 / 0 .33 - 0. 77 /0.58 2 8 $CO^+$ 1.98/ 0 .14 2.72 / 0. 30 2 .2 6 / 0.26 2 .13 / 0.26 2.69/ 0.26 3. 9 6 / 0. 1 0 3.3 6 /0.42
29 $COH^+ $ ;$ HCO^+$ 6 . 13/ 0 .16 6.95 / 0.25 6.32/ 0 . 2 1 6 .42 / 0.21 6.66 / 0.29 7 .15 /0. 17 7.06/ 0.4 0 30 $CH_2 O^+$ ; $CH3COO H^{++} $ ? 1.33 / 0.23 - 0. 27 / 0 . 24 0 .55 / 0.05 0. 22 /0.45 - 0.2 1 / 0.16
31 $ C H_3 O^{+} $? - 1.27 / 0.2 3 1.42/ 0 .3 3 1 .41 / 0.27 1. 29/ 0.23 1.43/ 0. 07 0. 92 / 0.27
40 $ C CO ^+$ 0. 37 / 0 .01 0.5 6 / 0.18 0.54 / 0. 21 0.52 / 0.22 0.6 4 / 0.15 1. 15 / 0 .18 0. 71 /0 .18
41 $ CHCO^+$ 1.17/ 0. 1 2 1.67 /0.1 2 1. 41 /0. 1 4 1.39 / 0.18 1.49 / 0.12 1.7 8 / 0.05 1. 59 / 0.244 2 $ CH_2 C O^ + $ 3.4 7 / 0.09 3.6 9 / 0.14 3.70 / 0 .14 3. 63/ 0 .11 3 . 56 / 0.14 3.2 1 / 0. 07 | 17 /_0.54
_26 _ _$CH_2C^{+}$?_ __ _ _ _ __ _-_ _0.87_/_0.27 0.68 /_0.23 0.65 /_0.45 _ 0.84 / 0.33 __ _ - 0.77 /_0.58
28 _ $CO^+$___ _ _ _ _ 1.98 / 0.14_ 2.72 / 0.30_ _2.26 / 0.26 _2.13 / 0.26_ _2.69 / 0.26_ 3.96 / 0.10 _ 3.36 /_0.42
29 __ $COH^+$; $HCO^+$___ _ __ __ __6.13 / 0.16 6.95 /_0.25 6.32_/_0.21 6.42_/ 0.21 6.66_/ 0.29 7.15 /_0.17 _ 7.06_/ 0.40
30 $CH_2O^+$; $CH3COOH^{++}$?_ _ 1.33 / 0.23_ __ __ _ - 0.27_/ 0.24_ 0.55 / 0.05 _ 0.22 / 0.45__ _ -_ 0.21 / 0.16
__31_ _$CH_3O^{+}$? _ _ __ __ ___- 1.27 /_0.23 1.42 / 0.33_ 1.41 / 0.27 1.29 /_0.23 _1.43 / 0.07 _0.92_/_0.27
40 $CCO^+$ _ _ _ _ _ 0.37 / 0.01 0.56 /_0.18 0.54 / 0.21 _0.52 / 0.22 0.64 /_0.15 1.15_/_0.18 0.71 / 0.18
_41___ _$CHCO^+$ _ _ _ _ _ __ 1.17 / 0.12 1.67 / 0.12_ 1.41 / 0.14 _ 1.39 / 0.18 __1.49 / 0.12_ 1.78 /_0.05 1.59 / 0.24
42_ $CH_2CO^+$ __ _ _ _ _ _ _ _ _ 3.47 / 0.09 3.69_/_0.14 3.70 / 0.14 3.63_/_0.11__ 3.56 /_0.14 _3.21 / 0.07_ |
-sur-Yvette, France'
- 'KU Leuven, Instituut voor Kern- en Stralingsfysica, 3001 Leuven, Belgium'
author:
- 'P. Arthuis'
- 'A. Tichai'
- 'J. Ripoche'
- 'T. Duguet'
bibliography:
- 'Bibliographie.bib'
title: |
`ADG`: Automated generation and evaluation of many-body diagrams\
II. Particle-number projected Bogoliubov many-body perturbation theory
---
many-body theory,[*ab initio*]{},perturbation theory,diagrammatic expansion 21.60.De
[**NEW VERSION PROGRAM SUMMARY**]{}
[*Program Title:*]{} `ADG`\
[*Licensing provisions:*]{} GPLv3\
[*Programming language:*]{} Python3\
[*Supplementary material:*]{} Yes.\
[*Journal reference of previous version:*]{} P. Arthuis, T. Duguet, A. Tichai, R.-D. Lasseri and J.-P. Ebran, “ADG: Automated generation and evaluation of many-body diagrams I. Bogoliubov many-body perturbation theory”, Computer Physics Communications 240 (2019), pp. 202-227.\
[*Does the new version supersede the previous version?:*]{} Yes.\
[*Reasons for the new version:*]{} Incorporation of a new formalism into the program.\
[*Summary of revisions:*]{} Addition of off-diagonal BMBPT to the formalisms for which diagrams can be generated, fix of a wrong symmetry factor, move of the codebase from Python2 to Python3 while maintaining support for Python2, various optimizations to reduce the time and memory necessary to the program.\
[*Nature of problem:*]{} As formal and numerical developments in many-body-perturbation-theory-based [*ab initio*]{} methods make higher orders reachable, manually producing and evaluating all the diagrams becomes rapidly untractable as both their number and complexity grow quickly, making it prone to mistakes and oversights.\
[*Solution method:*]{} Diagonal BMBPT diagrams are encoded as square matrices known as oriented adjacency matrices in graph theory, and then turned into graph objects using the *NetworkX* package. Off-diagonal BMBPT diagrams can | -sur - Yvette, France'
-' KU Leuven, Instituut voor Kern- en Stralingsfysica, 3001 Leuven, Belgium'
author:
-' P. Arthuis'
-' A. Tichai'
-' J. Ripoche'
-' T. Duguet'
bibliography:
-' Bibliographie.bib'
title: |
'ADG `: Automated generation and evaluation of many - body diagrams\
II. atom - numeral projected Bogoliubov many - body disturbance hypothesis
---
many - body theory,[*ab initio*]{},perturbation theory, diagrammatic expansion 21.60.De
[ * * NEW VERSION plan SUMMARY * * ] { }
[ * Program Title :* ] { } 'ADG`\
[ * Licensing provisions :* ] { } GPLv3\
[ * Programming language :* ] { } Python3\
[ * auxiliary fabric :* ] { } Yes.\
[ * Journal reference of previous interpretation :* ] { } P. Arthuis, T. Duguet, A. Tichai, R.-D. Lasseri and J.-P. Ebran, “ ADG: Automated generation and evaluation of many - soundbox diagrams I. Bogoliubov many - body perturbation theory ”, Computer Physics Communications 240 (2019), pp. 202 - 227.\
[ * Does the new adaptation supersede the previous version? :* ] { } Yes.\
[ * Reasons for the modern version :* ] { } Incorporation of a new formalism into the program.\
[ * Summary of revisions :* ] { } Addition of away - diagonal BMBPT to the formalism for which diagrams can be generated, fix of a improper symmetry factor, move of the codebase from Python2 to Python3 while maintaining support for Python2, diverse optimizations to reduce the time and memory necessary to the program.\
[ * Nature of problem :* ] { } As formal and numerical developments in many - body - perturbation - theory - establish [ * ab initio * ] { } methods make high orders reachable, manually producing and evaluating all the diagram becomes quickly untractable as both their number and complexity grow quickly, lay down it prone to mistakes and oversights.\
[ * Solution method :* ] { } Diagonal BMBPT diagrams are encoded as square matrices known as oriented adjacency matrices in graph theory, and then turned into graph aim practice the * NetworkX * package. Off - diagonal BMBPT diagrams can | -sur-Jvette, France'
- 'KU Leuven, Lnstituut voor Kgrb- en Svralingafysica, 3001 Leuven, Belgium'
author:
- 'P. Arthums'
- 'A. Richau'
- 'J. Ripoche'
- 'T. Duguet'
biblkography:
- 'Hibliogrqphit.bib'
title: |
`ADG`: Automatcb gensvatiou end evaluation pf many-bodf diagrams\
HI. Pcrticle-number projected Bogoliubov mwny-body pfrturbation thgory
---
msgy-bosj uheory,[*ab initio*]{},perturbation theorg,diagrakmatic expansoon 21.60.De
[**NEW VERSION PROGRAM DUMMWRY**]{}
[*Program Title:*]{} `AFG`\
[*Licensing produsions:*]{} GPLv3\
[*Pfogramming language:*]{} Pyjhon3\
[*Supplementary material:*]{} Yes.\
[*Jojrnal reference od pggvious versiin:*]{} P. Wrthuis, T. Dunlet, A. Thchai, R.-C. Lasseri and M.-P. Ebcan, “QDG: Automated generatmon and evaluation os many-bodf biagrams I. Bogoliubov mqny-bogy partufvatkon tieody”, Comoutxr Physics Dommunicatiins 240 (2019), pp. 202-227.\
[*Does the ntw dvtsion superssde thq [revious version?:*]{} Yes.\
[*Reasons for the new vedsion:*]{} Incorporation of q new formalism into jhe prograi.\
[*Summary of revisions:*]{} Addition of off-diagonal BMB[T to ghe fjfnapisms for which diagrams can be generated, fix kf a wrong symmetrn factor, move of tne cpqebase from Pithon2 to Pgthon3 while maintalning sopport for Pythjn2, vsrious optimizations to redyce the time qnd memory necessaxy to the prugrak.\
[*Natute of problem:*]{} As formal and numerical fevelopmehgs in many-body-peftugbathon-theory-based [*ab initio*]{} mqthods maje hngher oraers reachwble, manuaply pvmducing and evaluahing cll tve diagramd becomes rapidly untractable as both their nombar dnd compjexitn grow quickly, iaking it prong to mistckes avd oversiggts.\
[*Soluvion method:*]{} Qiagonal BMBPD diagrams arx encoded as wquaee matrkzes known as otiented ablacency mqtrices in graph tmeory, znd then turned inuo graph objects uskng tje *NqdworkX* packace. Owf-dksgonau BMBPT dianraos csn | -sur-Yvette, France' - 'KU Leuven, Instituut voor Stralingsfysica, Leuven, Belgium' - 'P. Arthuis' Ripoche' 'T. Duguet' bibliography: 'Bibliographie.bib' title: | Automated generation and evaluation of many-body II. Particle-number projected Bogoliubov many-body perturbation theory --- many-body theory,[*ab initio*]{},perturbation theory,diagrammatic expansion [**NEW VERSION PROGRAM SUMMARY**]{} [*Program Title:*]{} `ADG`\ [*Licensing provisions:*]{} GPLv3\ [*Programming language:*]{} Python3\ material:*]{} [*Journal of version:*]{} P. Arthuis, T. Duguet, A. Tichai, R.-D. Lasseri and J.-P. Ebran, “ADG: Automated generation and of many-body diagrams I. Bogoliubov many-body perturbation theory”, Physics Communications 240 (2019), 202-227.\ [*Does the new version the version?:*]{} Yes.\ for new Incorporation of a formalism into the program.\ [*Summary of revisions:*]{} Addition of off-diagonal BMBPT to the formalisms for which diagrams be generated, a wrong factor, of codebase from Python2 while maintaining support for Python2, various the time and memory necessary to the program.\ of problem:*]{} formal and numerical developments in many-body-perturbation-theory-based initio*]{} methods make higher orders reachable, manually producing evaluating all the diagrams becomes rapidly untractable as both their number and complexity grow quickly, prone to mistakes and [*Solution method:*]{} Diagonal diagrams encoded square known as adjacency matrices in graph theory, and then turned into graph objects the *NetworkX* package. Off-diagonal BMBPT diagrams can | -sur-Yvette, France'
- 'KU Leuven, InStituut vooR Kern- En STraLiNgsfYsicA, 3001 Leuven, Belgium'
AUthoR:
- 'P. Arthuis'
- 'A. Tichai'
- 'J. RipocHe'
- 'T. DuGuET'
bibLIoGraphY:
- 'BiblioGRaPHIe.bIb'
TiTle: |
`aDg`: auTomatEd gEneratiOn and evaluAtiOn Of many-body diAGrAms\
II. PartiCle-Number projecTed bogoliUbOv mANy-bodY peRturbAtion tHEory
---
maNy-body theOrY,[*Ab initIO*]{},perturBATiOn thEory,diagrammatic eXPaNSion 21.60.De
[**NEW VERSIoN PROGrAm sUmmaRY**]{}
[*proGram Title:*]{} `AdG`\
[*licenSIng provISiONS:*]{} gPLV3\
[*programming laNguage:*]{} PythoN3\
[*supPlemenTaRy mATerial:*]{} yes.\
[*JoUrNAl rEference of pReviOus versioN:*]{} P. ArthUIs, T. DuguET, A. TichaI, R.-D. LasSerI anD J.-P. EBRaN, “AdG: AUtOMatED gEneRAtiOn and evaLuAtIon of Many-BODY DiagRamS I. BoGoliuBov many-body peRtuRbatIOn tHeory”, compuTer PHySics COmmuniCatioNs 240 (2019), Pp. 202-227.\
[*Does the new verSion Supersede The PrEviOuS versIOn?:*]{} Yes.\
[*REasOns For the nEw versiON:*]{} InCoRPORaTion of a new formalisM iNTO tHe prograM.\
[*SummaRY oF rEVisions:*]{} ADdItiOn of OFF-diagOnal bmBpT to the fOrmaliSMs FoR which dIaGrams cAn Be gEneRated, FIx of A wrong Symmetry FactoR, Move of the codebASe from Python2 tO pyTHOn3 WHile MaiNtaining supPort FOr PyThon2, VArIouS OptimIzatiOnS To REduce the time and memoRy NecessAry to The program.\
[*NatUre of problEM:*]{} aS formal aNd nuMErICal developmentS in maNy-body-pertURbation-tHeory-Based [*ab iNitio*]{} methODS make higHer OrdErs ReaCHAbLe, manually proDUCing AnD evaluaTinG all the DiaGraMs bEcoMeS rapidly uNtractabLe As BoTh TheIr numBEr and comPlExiTy GroW quicKLy, makiNg it pRone To MiSTakEs and ovERsIGHts.\
[*SOlUtIon mEthOd:*]{} diagoNal BmbPT DiagramS are encodEd aS SquaRe MaTrices kNown as orienteD aDjacency maTrIceS in graPH Theory, anD then turned into graph objECts usinG thE *NetwOrkX* Package. OfF-diAgonal bMBpt diagrAms can | -sur-Yvette, France'
- 'KU Leuven, I nstit uut vo or Ker n- e n Stralingsfys i ca,3001 Leuven, Belgium'autho r: - 'P . A rthui s'
- 'A . T i c hai '-'J. R i po che'- ' T. Dugu et'
biblio gra ph y:
- 'Biblio g ra phie.bib'tit le: |
`A DG` : Auto ma ted gener ati on an d eval u ationof many-b od y diagr a ms\
I I. Par ticle-number proj e ct e d Bogoliubov m any-bo dy pe r t urb ati on theory-- -
ma n y-bodyt he o r y ,[* a b initio*]{}, perturbatio n th eory,d ia gra m maticexpan si o n 2 1.60.De
[* *NEW VERSIONPROGRA M SUMMAR Y **]{}
[*Prog ram Ti tle: * ]{ }`AD G` \
[* L ic ens i ngprovisio ns :* ]{} G PLv3 \ [ * Prog ram ming lang uage:*]{} Pyt hon 3\
[ * Sup pleme ntary mat er ial:* ]{} Ye s.\
[ *J ournal referenc e of previous ve rs ion :* ]{} P . Arthu is, T. Duguet , A. Ti c hai ,R . - D. Lasseri and J.-P. E b r an , “ADG:Automa t ed g e neration a ndeval u a tionof m a ny -body di agrams I. B ogoliub ov many- bo dyper turba t iontheory ”, Compu ter P h ysics Communic a tions 240 (20 1 9) , pp . 202 -22 7.\
[*Doesthen ew v ersi o nsup e rsede thepr e vi o us version?:*]{} Ye s. \
[*Re asons for the newversion:*] { } Incorpor atio n o f a new formali sm in to the pro g ram.\
[* Summa ry of re visions:* ] { } Additi onofoff -di a g on al BMBPT to t h e for ma lisms f orwhich d iag ram s c anbe generate d, fix o fawr on g s ymmet r y factor ,mov eofthe c o debase from Pyt ho n2 toPython3 wh i l e ma in ta inin g s up portforP yth on2, va rious opt imi z atio ns t o reduc e the time an dmemory nec es sar y to t h e program .\
[*Nature of problem: * ]{} Asfor mal a nd n umericaldev elopme nts in man y-body -pert ur bat i o n-the o r y- bas ed [*ab init i o *]{ } met ho ds m ake hig her orders reachab l e,manually prod uci ng a n d e val u at i ngal l th e diagrams become s rapidlyun t ra ctable asb oth t heir nu mber an d com p lexitygrow quic kly, maki ng itp r one to mistak es and o versights . \
[*S o lu tionmet hod:*] {} Di agona l BMBP T di agram s areen codedas sq ua re matri ces known as oriented a djacen cy ma tri ces in gr aph the ory, andthen turned in togra ph ob jec t s usi ng t h e*Ne t workX * pa c kage. Off - di ago n a lBMBPT diagr a m s ca n | -sur-Yvette, France'
-_'KU Leuven,_Instituut voor Kern- en_Stralingsfysica, 3001_Leuven,_Belgium'
author:
- 'P. Arthuis'
-_'A. Tichai'
-_'J. Ripoche'
- 'T. Duguet'
bibliography:
- 'Bibliographie.bib'
title:_|
_ `ADG`: Automated generation_and evaluation of_many-body_diagrams\
II. Particle-number projected Bogoliubov many-body perturbation theory
---
many-body theory,[*ab initio*]{},perturbation theory,diagrammatic_expansion_21.60.De
[**NEW VERSION_PROGRAM_SUMMARY**]{}
[*Program_Title:*]{} `ADG`\
[*Licensing provisions:*]{} GPLv3\
[*Programming language:*]{}_Python3\
[*Supplementary material:*]{} Yes.\
[*Journal reference of_previous version:*]{}_P. Arthuis, T. Duguet, A. Tichai, R.-D. Lasseri_and_J.-P. Ebran, “ADG:_Automated generation and evaluation of many-body diagrams I. Bogoliubov_many-body perturbation theory”, Computer Physics Communications_240 (2019), pp._202-227.\
[*Does_the_new version supersede the_previous version?:*]{} Yes.\
[*Reasons for the new_version:*]{} Incorporation of a new formalism_into the program.\
[*Summary of revisions:*]{} Addition of_off-diagonal BMBPT to the formalisms for_which diagrams can be generated,_fix of_a wrong symmetry factor, move_of the codebase_from Python2_to Python3 while_maintaining support for Python2, various optimizations_to reduce the_time and memory necessary to the_program.\
[*Nature_of problem:*]{} As_formal_and_numerical developments_in many-body-perturbation-theory-based [*ab_initio*]{}_methods make_higher_orders reachable, manually producing and evaluating_all_the diagrams becomes rapidly untractable as both_their number and complexity_grow_quickly, making it prone_to mistakes and oversights.\
[*Solution method:*]{}_Diagonal BMBPT diagrams are encoded as_square matrices_known as_oriented adjacency matrices in graph theory, and then turned into graph_objects using the *NetworkX* package. Off-diagonal_BMBPT diagrams can |
classification and unsupervised learning. On the contrary, possible statistical models of the other measures need to take into account the exact form of the objective function (or the loss function) and therefore, they cannot be generally applicable. Lastly, for small sample sizes, as also noted by Archer and Kimes in [@Archer2008], subsampling or bootstrapping of the samples during training might not be desirable, which limits the use of permutation importance.
A Model of False Positive Rates for Selection Frequency {#sec:model}
=======================================================
The main question we want to answer in this section is [*“What is the probability that a feature $f$ is selected $k$ times in a Random Forest under the null hypothesis?”*]{} with the null hypothesis being [*“There is no statistical relationship between features (covariates) and labels (response variables).”*]{} Answering this question will allow us to estimate false positive rates for a given threshold on selection frequency. Consequently, it will provide us a principled way to determine thresholds to split relevant and non-relevant features.
Our strategy is to build an approximate probabilistic model of selection frequency that takes into account the details of the training procedure, such as the feature subsampling scheme, number of trees and the size of the random feature subset each node optimizes over[^3]. Once this model is built we can compute the expected rate of false positives for a given threshold on selection frequency and consequently determine thresholds by limiting this rate. We take a bottom-up approach and start by modeling the probability of a feature $f$ being selected at a given node and build-up the model from there.
The randomness in the feature selection process is due to the random nature of forest training. As mentioned in Section \[sec:RF\] there are two components that provide this random nature: feature subsampling and sample subsampling or bootstrapping. In our approach we explicitly formulate the feature subsampling and this is the source of the probabilistic nature of our model. We do not explicitly model the sample subsampling scheme and its effect on the selection frequency. There are two main reasons for this. First, the effect of sample subsampling on feature selection is much more difficult to explicitly formulate. Second, for small datasets bootstrapping or subsampling of the data might not be used at all when training a random forest. Feature subsampling on the other hand is used all the time. However, although we do not model sample | classification and unsupervised learning. On the contrary, possible statistical model of the early measures need to bring into report the exact form of the objective routine (or the passing function) and consequently, they cannot be by and large applicable. Lastly, for little sample sizes, as besides noted by Archer and Kimes in [ @Archer2008 ], subsampling or bootstrapping of the samples during training might not be desirable, which limits the use of permutation importance.
A Model of False Positive Rates for Selection Frequency { # sec: model }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
The chief question we want to answer in this incision is [ * “ What is the probability that a feature $ f$ is selected $ k$ times in a Random Forest under the null guess? ” * ] { } with the null hypothesis being [ * “ There exist no statistical relationship between features (covariates) and labels (reception variables). ” * ] { } Answering this question will allow us to estimate false positive rates for a given threshold on selection frequency. Consequently, it will provide us a principled way to determine thresholds to split relevant and non - relevant features.
Our scheme is to construct an approximate probabilistic model of excerpt frequency that remove into account the details of the training procedure, such as the feature subsampling scheme, number of trees and the size of the random feature of speech subset each node optimizes over[^3 ]. Once this model is built we can compute the expected rate of false positives for a given threshold on selection frequency and consequently determine threshold by restrict this rate. We take a bottom - up approach and start by modeling the probability of a feature of speech $ f$ being selected at a given node and build - up the model from there.
The randomness in the feature selection process is due to the random nature of forest education. As mentioned in Section \[sec: RF\ ] there are two components that provide this random nature: feature subsampling and sample distribution subsampling or bootstrapping. In our approach we explicitly formulate the feature subsampling and this is the source of the probabilistic nature of our exemplar. We do not explicitly model the sample subsampling outline and its effect on the selection frequency. There are two main reasons for this. foremost, the effect of sample subsampling on feature choice is much more difficult to explicitly formulate. Second, for small datasets bootstrapping or subsampling of the data might not be use at all when training a random forest. feature of speech subsampling on the other hand is used all the time. However, although we do not model sample | clwssification and unsuperyised learning. Ou the cmntrarg, possibue statistical models of the orher neasures need to take knto accolnt the ezact dorm of thx objectlre fuhgtion (ir the loss fukction) and dherefore, they cxnuot be generally applicable. Lastly, fjr smalk dample sizes, af alxj nofvd by Archer and Kimes in [@Arched2008], subsakpling or booystrapping of the samples furijg training might jot be desitzblq, which limitr the use of permutatikn importance.
A Model of False Pusitire Rates fot Aepgction Frequxncy {#svc:model}
=======================================================
The malm quesdion we want to answev in vhis section is [*“What is tie probability that w feature $f$ is selected $k$ tinew in d Ratdom Dordst uided the julm hypothesjs?”*]{} with the null hypothesis beond [*“There is no atatiseisal relationship between features (covarpatea) and labels (response vqriables).”*]{} Answering thls questijn will allow us to estimate false positive rates for e eivtn threryopd on selection frequency. Consequently, it wilj ptonide us a principjed way to cehetiine thresholas to spljt relevant and noj-relevagt feqtures.
Our strstegy is to build an approxumate probabplisric model of seleccion frequeney thaj takex into account the detanls of the trainijg procedhfe, such as the fdatlre vubsampling scheme, number jf trees end tke size uf tne ranqom featurf subset each node optimlzes lvar[^3]. Once thls model is built we can computx the expectec satv of falsz posibives for a givqn threshold ou selectnon frdquency ans consexuently detqrmine threshmpds by limitmng this wate. We rake a cuttom-up approavh and stcxt by modwling the probabillty ow a feature $f$ benug selected at a bivdn gofe agg build-up tha moael grom ghere.
Tht rcnaomnrss in the feature salecfion process is dur bo the rabdom nattre of forest training. As mentilned mn Secvion \[sev:RF\] there are two components that provide hhif random nattre: neatore subsampking and sample subsampling or bootstra'ping. In our approach ww explicitly formulctt the feature subswmpling atd this is the sourcw of the probabillstic nature of our mosel. We do nlt explicitly model the sample subsampling scheme and its effect on the swlection frequency. Thete arv tco kain rqasois for this. First, nhe effect of sample subsampling on featgrz selection is much more diffocjlt to explicktly formulate. Second, fod small catasets bootstrapping or subsakpling of the data might not be ysed zt all whei training a random gorest. Featyre subwampping on the oghet mand is usfd akl the time. However, although we do not midfl sample | classification and unsupervised learning. On the contrary, models the other need to take of objective function (or loss function) and they cannot be generally applicable. Lastly, small sample sizes, as also noted by Archer and Kimes in [@Archer2008], subsampling bootstrapping of the samples during training might not be desirable, which limits the of importance. Model False Positive Rates for Selection Frequency {#sec:model} ======================================================= The main question we want to answer in section is [*“What is the probability that a $f$ is selected $k$ in a Random Forest under null with the hypothesis [*“There no statistical relationship features (covariates) and labels (response variables).”*]{} Answering this question will allow us to estimate false positive rates a given selection frequency. it provide a principled way thresholds to split relevant and non-relevant is to build an approximate probabilistic model of frequency that into account the details of the procedure, such as the feature subsampling scheme, number trees and the size of the random feature subset each node optimizes over[^3]. Once this built we can compute expected rate of positives a threshold selection frequency consequently determine thresholds by limiting this rate. We take a bottom-up and start by modeling the probability of a feature $f$ at given node and the model from there. randomness the feature selection process to random training. mentioned Section \[sec:RF\] there are components that provide this random feature subsampling and sample approach we explicitly formulate the feature subsampling and is the source of the probabilistic nature our model. We do not explicitly model the sample subsampling scheme and effect on frequency. There are two main reasons for this. the effect of sample on feature selection is much more difficult to explicitly Second, small datasets or subsampling of data might not used at all a random Feature on is used all the time. However, we not model sample | classification and unsupervIsed learniNg. On tHe cOntRaRy, poSsibLe statistical mODels Of the other measures need To takE iNTo acCOuNt the Exact foRM oF THe oBjEcTivE fUNcTion (oR thE loss fuNction) and tHerEfOre, they cannoT Be Generally aPplIcable. Lastly, For Small sAmPle SIzes, aS alSo notEd by ArCHer and kimes in [@ArChER2008], subsaMPling or BOOtStraPping of the samples DUrINg training mighT not be DeSIrABLe, wHicH limits the UsE of peRMutatioN ImPORTanCE.
A Model of FalsE Positive RaTEs fOr SeleCtIon fRequenCy {#sec:MoDEl}
=======================================================
THe main questIon wE want to anSwer in THis sectIOn is [*“WhaT is the ProBabIlitY ThAt A feAtURe $f$ IS sEleCTed $K$ times in A RAnDom FoRest UNDER the NulL hypOthesIs?”*]{} with the null HypOtheSIs bEing [*“THere iS no sTaTistiCal relAtionShIp between featurEs (coVariates) aNd lAbEls (ReSponsE VariabLes).”*]{} ansWering tHis quesTIon WiLL ALlOw us to estimate falsE pOSItIve rates For a giVEn ThREshold on SeLecTion FREquenCy. CoNSeQuently, iT will pROvIdE us a priNcIpled wAy To dEteRmine THresHolds tO split reLevanT And non-relevant FEatures.
Our strATeGY Is TO buiLd aN approximatE proBAbilIstiC MoDel OF seleCtion FrEQuENcy that takes into accOuNt the dEtailS of the traininG procedure, SUCH as the feAturE SuBSampling scheme, NumbeR of trees anD The size oF the rAndom feaTure subseT EAch node oPtiMizEs oVer[^3]. oNCe This model is buILT we cAn Compute The ExpecteD raTe oF faLse PoSitives foR a given tHrEsHoLd On sElectIOn frequeNcY anD cOnsEquenTLy deteRmine ThreShOlDS by LimitinG ThIS Rate. we TaKe a bOttOm-Up appRoacH And Start by Modeling tHe pRObabIlItY of a feaTure $f$ being selEcTed at a giveN nOde And buiLD-Up the modEl from there.
The randomnesS In the feAtuRe selEctiOn process Is dUe to thE raNDom natUre of fOrest TrAinING. As meNTIoNed In section \[sec:rf\] TheRe are TwO comPonents That provide this ranDOm nAture: feature sUbsAmplING aNd sAMpLE suBsAMplING or bootstrappinG. In our apprOaCH wE explicitlY ForMuLate the Feature SubsaMPling anD this is thE source of ThE proBABilIstic naturE of our moDel. We do noT ExpliCItLy modEl tHe sampLe SubSamplIng schEMe aNd its Effect On The selEctioN fRequency. there are two main reasons fOr this. first, The Effect of sAmpLE suBsampling On feAture selecTioN is Much mOre DIfficUlt tO ExPliCItly fOrmuLAte. Second, FOr SmaLL DaTasets bootsTRAPpiNg or sUbsAMpling Of thE data might not be usED at all when traiNing A RAndOm fORest. feAture subsampliNg oN tHE Other hanD iS used all the Time. HoweVeR, AlthoUgh we dO not moDel sampLE | classification and unsupe rvised lea rning . O n t he con trar y, possible st a tist ical models of the oth er me as u resn ee d totake in t oa c cou nt t heex a ct form of the ob jective fu nct io n (or the lo s sfunction)and therefore,the y cann ot be gener all y app licabl e . Last ly, for s ma l l samp l e sizes , as als o noted by Archer an d Kimes in [@Ar cher20 08 ] ,s u bsa mpl ing or boo ts trapp i ng of t h es a m ple s during train ing might n o t b e desi ra ble , which limi ts the use of per muta tion impo rtance .
A Mod e l of Fa lse Po sit ive Rat e sfo r S el e cti o nFre q uen cy {#sec :m od el}
= ==== = = = = ==== === ==== ===== ============= === ==== = === ===== =
Th e ma in ques tion w e wan tto answer in th is s ection is [* “W hat i s the probab ili tythat afeature $f$ i s s el ected $k$ times in a R an dom Fore st und e rth e null hy po the sis? ” * ]{} w itht he null hy pothes i sbe ing [*“ Th ere is n o s tat istic a l re lation ship bet weenf eatures (covar i ates) and lab e ls ( re s pons e v ariables).” *]{} Answ erin g t his quest ion w il l a l low us to estimatefa lse po sitiv e rates for a given thr e s h old on s elec t io n frequency. Co nsequ ently, itw ill prov ide u s a prin cipled wa y to deter min e t hre sho l d sto split rele v a nt a nd non-re lev ant fea tur es.
O urst rategy is to buil dan a pp rox imate probabil is tic m ode l ofs electi on fr eque nc yt hat takesi nt o acco un tthedet ai ls of the tra ining p rocedure, su c h as t he featur e subsampling s cheme, num be r o f tree s and thesize of the random feat u re subs eteachnode optimize s o ver[^3 ].O nce th is mod el is b uil t we ca n co mpu te the expec t e d r ate o ffals e posit ives for a given t h res hold on selec tio n fr e q ue ncy an d co ns e que n t ly determine th resholds b yl im iting this rat e. We tak e a bot tom-u p approa ch and st art by mo de ling t heprobabilit y of a f eature $f $ bein g s elect edat a g iv ennodeand bu i ld- up th e mode lfrom t here.
The rand omness in the feature s electi on pr oce ss is due to the random n atur e of fores t t rai ning. As menti oned in Se c tion\[se c :RF\] the r eare t wo components t h atprovi det his ra ndom nature: features ubsampling and sam p l e s ubs a mpli ng or bootstrapp ing .I n our app ro ach we expl icitly f or m ulate the f eature subsam p l in g and t hisisthe sourc e o ft he prob ab il i stic n atur eof our model . Wed o not explicitlymodel t he sa m ple subs am pling s c heme and its e ffect on th e sele ctio n fre quency. T here a retw o main rea s ons for t his.First,th e ef fec t of s ampl e subsa mpli ng on features e le c ti on ismuch more d iffi cult to e x plicitly fo r mulate. S eco n d , fors ma l l datasetsboo tstra p p ing or sub s ampl i ng of th e data might not be use dat allwhe n traininga randomf ore st . Fe ature su bs ampl in g o nthe other handis usedall thet i me. Ho wev e r, a l though w e do not mod el sa mp l e | classification_and unsupervised_learning. On the contrary,_possible statistical_models_of the_other_measures need to_take into account_the exact form of_the objective function_(or_the loss function) and therefore, they cannot be generally applicable. Lastly, for small sample_sizes,_as also_noted_by_Archer and Kimes in [@Archer2008],_subsampling or bootstrapping of the_samples during_training might not be desirable, which limits the_use_of permutation importance.
A_Model of False Positive Rates for Selection Frequency {#sec:model}
=======================================================
The_main question we want to answer_in this section_is_[*“What_is the probability that_a feature $f$ is selected $k$_times in a Random Forest under_the null hypothesis?”*]{} with the null hypothesis_being [*“There is no statistical relationship_between features (covariates) and labels_(response variables).”*]{}_Answering this question will allow_us to estimate_false positive_rates for a_given threshold on selection frequency. Consequently,_it will provide_us a principled way to determine_thresholds_to split relevant_and_non-relevant_features.
Our strategy_is to build_an_approximate probabilistic_model_of selection frequency that takes into_account_the details of the training procedure, such_as the feature subsampling_scheme,_number of trees and_the size of the random_feature subset each node optimizes over[^3]._Once this_model is_built we can compute the expected rate of false positives for_a given threshold on selection frequency_and consequently determine thresholds_by limiting_this_rate. We take_a_bottom-up approach_and start by modeling the probability of_a feature_$f$ being selected at a given_node and build-up the_model_from there.
The randomness in the feature_selection process is due to the_random nature of forest training._As_mentioned_in Section \[sec:RF\] there are two_components that provide this random nature:_feature subsampling and_sample subsampling or bootstrapping. In our approach_we_explicitly formulate the feature subsampling and_this_is the source of the probabilistic_nature_of_our model. We do not_explicitly model the sample subsampling scheme_and its effect on the selection frequency. There are_two main reasons_for this. First, the effect_of_sample_subsampling on feature selection is much more difficult to explicitly_formulate. Second,_for small datasets_bootstrapping or subsampling of the data might not be used_at all when training a random forest._Feature subsampling on the other hand is used all the time._However, although we do not model sample |
\exp\left[ -{3M_{\rm Pl}^4 \over 8}\left({1\over V(\phi_t)} -
{1\over V(\phi_f)}\right) \right]$$
The continued nucleation and expansion of bubbles of one vacuum within the other will result in a spacetime that is a rather inhomogeneous mixture of the two vacua. There is an intriguing thermal interpretation of this mixture if $V(\phi_f) - V(\phi_t)
\ll (V(\phi_f) + V(\phi_t))/2 \equiv \bar V$, so that the geometry of space is approximately the same in the regions of either vacua, with a Hubble parameter $$\bar H \approx \sqrt{8\pi \bar V \over 3 M_{\rm Pl}^2}$$
It seems plausible that the fraction of space contained in each of the vacua might tend to a constant, with the nucleation of true vacuum bubbles in false vacuum regions being just balanced by the nucleation of false vacuum bubbles in true vacuum regions. For such an equilibrium to hold, the volumes $\Omega_f$ and $\Omega_t$ of false and true vacuum must satisfy $${ \Omega_f \over \Omega_t} =
{\Gamma_{t\rightarrow f} \over \Gamma_{f\rightarrow t}} \approx
e^{-\Omega_{\rm hor} [V(\phi_f) - V(\phi_t)]/T_H }
\label{rateratio}$$ where the horizon volume $\Omega_{\rm hor}= (4\pi/3) \bar H^{-3}$ and the Hawking temperature $T_H = \bar H/2\pi$. If we view the de Sitter space as being somewhat analogous to an ensemble of quasi-independent horizon volumes in a thermal bath, then this leading contribution to the volume ratio is essentially a Boltzmann factor.
The zero mode corrections to the nucleation rate are consistent with this thermodynamic picture. Their effect is to multiply the ratio in Eq. (\[rateratio\]) by $${{\cal V}(G/H_f) \over {\cal V}(G/H_t)} \, {J_f^{G/H_f} \over
J_t^{G/H_t} }
= \left ( {\bar H\over \sqrt{3}\,\pi} \right)^{N_t | \exp\left [ -{3M_{\rm Pl}^4 \over 8}\left({1\over V(\phi_t) } -
{ 1\over V(\phi_f)}\right) \right]$$
The continued nucleation and expansion of bubbles of one vacuum within the other will leave in a spacetime that is a preferably inhomogeneous mixture of the two vacua. There be an challenging thermal interpretation of this mixture if $ V(\phi_f) - V(\phi_t)
\ll (V(\phi_f) + V(\phi_t))/2 \equiv \bar V$, so that the geometry of space is approximately the like in the region of either vacua, with a Hubble argument $ $ \bar planck's constant \approx \sqrt{8\pi \bar V \over 3 M_{\rm Pl}^2}$$
It seem plausible that the fraction of space contained in each of the void might tend to a changeless, with the nucleation of true vacuum bubbles in false vacuum region being just balanced by the nucleation of false void bubbles in true void regions. For such an balance to hold, the volume $ \Omega_f$ and $ \Omega_t$ of false and true vacuum must meet $ $ { \Omega_f \over \Omega_t } =
{ \Gamma_{t\rightarrow f } \over \Gamma_{f\rightarrow t } } \approx
e^{-\Omega_{\rm hor } [ V(\phi_f) - V(\phi_t)]/T_H }
\label{rateratio}$$ where the horizon volume $ \Omega_{\rm hor}= (4\pi/3) \bar H^{-3}$ and the Hawking temperature $ T_H = \bar H/2\pi$. If we view the de Sitter space as being somewhat analogous to an ensemble of quasi - independent horizon volume in a thermal bathroom, then this leading contribution to the volume ratio is essentially a Boltzmann factor.
The zero mode corrections to the nucleation rate are reproducible with this thermodynamic picture. Their effect is to reproduce the proportion in Eq. (\[rateratio\ ]) by $ $ { { \cal V}(G / H_f) \over { \cal V}(G / H_t) } \, { J_f^{G / H_f } \over
J_t^{G / H_t } }
= \left ({ \bar H\over \sqrt{3}\,\pi } \right)^{N_t | \exo\left[ -{3M_{\rm Pl}^4 \over 8}\left({1\ovtr V(\phi_t)} -
{1\ovgr V(\phi_f)}\cight) \rjght]$$
The zontinued nucleation and expensiin of bubbles of one vacuum within tje other wilo result in a spacebnme tgwt iv a rather inhokogeneous kixture of the tdo vacua. There is an intriguing thermwl intetpgetation of thys mpxeure pf $V(\phi_f) - V(\phi_t)
\ll (V(\phi_f) + V(\phi_t))/2 \squiv \ber V$, so that thr geometry of space is appgoxilately the same in the regionw of wither vacua, with a Hubble parametgr $$\bar H \approx \sqrt{8\pi \bar V \ovdr 3 M_{\xm Pl}^2}$$
It seens plwosible that vhe frwction of space contdined im each of the yacua mitht tend to a constanv, with the nucleatiog of true vccuum bubbles in falsw cacuuk recionr bekng jnst balanfed by the nudleation of false vacuum bubblts yb true vacuum regiogs. For such an equilibrium to hold, the vonumss $\Omega_f$ and $\Omega_t$ of false and true vacuul must saeisfy $${ \Omega_f \over \Omega_t} =
{\Gamma_{t\rightarrow f} \mver \Jaoma_{y\vlghtxerlw t}} \approx
e^{-\Omega_{\rm hor} [V(\phi_f) - V(\phi_t)]/T_H }
\jzbtl{rsteratio}$$ where the horizom gokome $\Omega_{\rm hot}= (4\pi/3) \bax H^{-3}$ and the Hawking tfmperatore $T_H = \bar H/2\pi$. If ee view the de Sitter space as being soiwwhat analogous to an ensemblz of qoasi-incependent horizon volumzs in z thermal bwth, then ffis leading contfibltiot to the volume ratio is efsentiallb a Bpltzmanv favtor.
Thq zero modf corvactions to the nucpeatiln rate are fonsistent with this thermodynamic picture. Thgir efxect is co mulbiply the ratio in Eq. (\[rateratip\]) by $${{\caj V}(G/H_w) \over {\cal V}(G/H_t)} \, {O_f^{G/H_f} \over
J_t^{D/H_t} }
= \ngft ( {\bar H\ovec \sqrt{3}\,\pi} \wighr)^{N_t | \exp\left[ -{3M_{\rm Pl}^4 \over 8}\left({1\over V(\phi_t)} - \right]$$ continued nucleation expansion of bubbles other result in a that is a inhomogeneous mixture of the two vacua. is an intriguing thermal interpretation of this mixture if $V(\phi_f) - V(\phi_t) \ll + V(\phi_t))/2 \equiv \bar V$, so that the geometry of space is approximately same the of vacua, with a Hubble parameter $$\bar H \approx \sqrt{8\pi \bar V \over 3 M_{\rm Pl}^2}$$ It plausible that the fraction of space contained in of the vacua might to a constant, with the of vacuum bubbles false regions just balanced by nucleation of false vacuum bubbles in true vacuum regions. For such an equilibrium to hold, the volumes and $\Omega_t$ and true must $${ \over \Omega_t} = \over \Gamma_{f\rightarrow t}} \approx e^{-\Omega_{\rm hor} } \label{rateratio}$$ where the horizon volume $\Omega_{\rm hor}= \bar H^{-3}$ the Hawking temperature $T_H = \bar If we view the de Sitter space as somewhat analogous to an ensemble of quasi-independent horizon volumes in a thermal bath, then this to the volume ratio essentially a Boltzmann The mode to nucleation rate consistent with this thermodynamic picture. Their effect is to multiply the in Eq. (\[rateratio\]) by $${{\cal V}(G/H_f) \over {\cal V}(G/H_t)} \, J_t^{G/H_t} = \left ( H\over \sqrt{3}\,\pi} \right)^{N_t | \exp\left[ -{3M_{\rm Pl}^4 \over 8}\left({1\over V(\Phi_t)} -
{1\over V(\pHi_f)}\riGht) \RigHt]$$
the cOntiNued nucleation ANd exPansion of bubbles of one vAcuum WiTHin tHE oTher wIll resuLT iN A SpaCeTiMe tHaT Is A rathEr iNhomogeNeous mixtuRe oF tHe two vacua. ThERe Is an intrigUinG thermal inteRprEtatioN oF thIS mixtUre If $V(\phI_f) - V(\phi_T)
\Ll (V(\phi_F) + V(\phi_t))/2 \equIv \BAr V$, so tHAt the geOMEtRy of Space is approximatELy THe same in the regIons of EiTHeR VAcuA, wiTh a Hubble pArAmeteR $$\Bar H \appROx \SQRT{8\pi \BAr V \over 3 M_{\rm Pl}^2}$$
IT seems plausIBle That thE fRacTIon of sPace cOnTAinEd in each of tHe vaCua might tEnd to a COnstant, WIth the nUcleatIon Of tRue vACuUm BubBlES in FAlSe vACuuM regions BeInG just BalaNCED By thE nuCleaTion oF false vacuum bUbbLes iN TruE vacuUm regIons. foR such An equiLibriUm To hold, the volumeS $\OmeGa_f$ and $\OmeGa_t$ Of FalSe And trUE vacuuM muSt sAtisfy $${ \OMega_f \ovER \OmEgA_T} =
{\gAmMa_{t\rightarrow f} \over \gaMMA_{f\RightarrOw t}} \appROx
E^{-\OMEga_{\rm hor} [v(\pHi_f) - v(\phi_T)]/t_h }
\labeL{ratERaTio}$$ where The horIZoN vOlume $\OmEgA_{\rm hor}= (4\Pi/3) \Bar h^{-3}$ anD the HAWkinG tempeRature $T_H = \Bar H/2\pI$. if we view the de SITter space as beINg SOMeWHat aNalOgous to an enSembLE of qUasi-INdEpeNDent hOrizoN vOLuMEs in a thermal bath, theN tHis leaDing cOntribution to The volume rATIO is essenTialLY a bOltzmann factor.
the zeRo mode corrECtions to The nuCleation Rate are coNSIstent wiTh tHis TheRmoDYNaMic picture. TheIR EffeCt Is to mulTipLy the raTio In EQ. (\[raTerAtIo\]) by $${{\cal V}(G/h_f) \over {\caL V}(g/H_T)} \, {J_F^{G/h_f} \oVer
J_t^{g/h_t} }
= \left ( {\baR H\OveR \sQrt{3}\,\Pi} \rigHT)^{N_t | \exp\left[ -{3M_{\rm Pl}^ 4 \over 8} \left ({1 \ov er V(\ phi_ t)} -
{1 \ over V(\phi_f)}\right) \ri ght]$ $Thec on tinue d nucle a ti o n an dex pan si o nof bu bbl es of o ne vacuumwit hi n the otherw il l result i n a spacetime t hat is ara the r inho mog eneou s mixt u re ofthe two v ac u a. The r e is an i nt rigu ing thermal inter p re t ation of thismixtur ei f$ V (\p hi_ f) - V(\ph i_ t)
\l l (V(\ph i _f ) + V( \ phi_t))/2 \eq uiv \bar V$ , so thatth e g e ometry of s pa c e i s approxima tely the same in th e region s of eit her va cua , w itha H ub ble p a ram e te r $ $ \ba r H \app ro x \sqr t{8\ p i \ barV \ over 3 M_ {\rm Pl}^2}$$
I t se e msplaus iblethat t he fr action of s pa ce contained in eac h of thevac ua mi gh t ten d to acon sta nt, wit h the n u cle at i o n o f true vacuum bubb le s in false v acuumr eg io n s beingju stbala n c ed by the nu cleation of fa l se v acuum b ub bles i ntru e v acuum regi ons. F or suchan eq u ilibrium to ho l d, the volume s $ \ O me g a_f$ an d $\Omega_t $ of fals e an d t rue vacuu m mus ts at i sfy $${ \Omega_f \o ve r \Ome ga_t} =
{\Gamm a_{t\right a r r ow f} \o ver\ Ga m ma_{f\rightarr ow t} } \approx e^{ -\Ome ga_{\rmhor} [V(\ p h i_f) - V (\p hi_ t)] /T_ H }\label{ratera t i o}$$ w here th e h orizonvol ume $\ Ome ga _{\rm hor }= (4\pi /3 )\b ar H^ {-3}$ and theHa wki ng te mpera t ure $T _H =\bar H /2 \ pi$ . If we vi e w the d eSitt ersp ace a s be i ngsomewha t analogo ust o an e ns emble o f quasi-indep en dent horiz on vo lumesi n a therm al bath, then this lead i ng cont rib ution tothe volum e r atio i s e s sentia lly aBoltz ma nnf a ctor. Th e z er o mode cor r e cti ons t othenucleat ion rate are consi s ten t with this t her mody n a mi c p i ct u re. T h eir e ffect is to mul tiply thera t io in Eq. (\ [ rat er atio\]) by $${ {\cal V}(G/H_ f) \over{\cal V}( G/ H_t) } \,{J_f^{G/H_ f} \over
J_t^{G/H _ t} }
= \ left ( { \ba r H\o ver \s q rt{ 3}\,\ pi} \r ig ht)^{N _t | \exp\left[_-{3M_{\rm Pl}^4_\over 8}\left({1\over V(\phi_t)} -
_ __ _{1\over_V(\phi_f)}\right) \right]$$
The continued_nucleation and expansion_of bubbles of one_vacuum within the_other_will result in a spacetime that is a rather inhomogeneous mixture of the two_vacua._There is_an_intriguing_thermal interpretation of this mixture_if $V(\phi_f) - V(\phi_t)
\ll (V(\phi_f)_+ V(\phi_t))/2_\equiv \bar V$, so that the geometry of_space_is approximately the_same in the regions of either vacua, with a_Hubble parameter $$\bar H \approx _\sqrt{8\pi \bar V_\over_3_M_{\rm Pl}^2}$$
It seems plausible_that the fraction of space contained_in each of the vacua might_tend to a constant, with the nucleation_of true vacuum bubbles in false_vacuum regions being just balanced_by the_nucleation of false vacuum bubbles_in true vacuum_regions. For_such an equilibrium_to hold, the volumes $\Omega_f$ and_$\Omega_t$ of false_and true vacuum must satisfy $${_\Omega_f_\over \Omega_t} =_
__ {\Gamma_{t\rightarrow_f} \over \Gamma_{f\rightarrow_t}}_\approx
__ e^{-\Omega_{\rm hor} [V(\phi_f)_-_V(\phi_t)]/T_H }
\label{rateratio}$$ where the horizon volume $\Omega_{\rm_hor}= (4\pi/3) \bar H^{-3}$_and_the Hawking temperature $T_H_= \bar H/2\pi$. If we_view the de Sitter space as_being somewhat_analogous to_an ensemble of quasi-independent horizon volumes in a thermal bath, then_this leading contribution to the volume_ratio is essentially a_Boltzmann factor.
The_zero_mode corrections to_the_nucleation rate_are consistent with this thermodynamic picture. Their_effect is_to multiply the ratio in Eq. (\[rateratio\])_by $${{\cal V}(G/H_f) \over_{\cal_V}(G/H_t)} \, {J_f^{G/H_f} \over
J_t^{G/H_t} }_
=_ \left ( {\bar H\over_\sqrt{3}\,\pi}_\right)^{N_t |
_*$-isomorphisms, stable fibrations and cofibrations form a model structure on the category of orthogonal $G$-spectra, the [*stable model structure*]{}.
2. Every cofibration of orthogonal $G$-spectra is an h-cofibration.
3. Let $i:A{\longrightarrow}B$ be a cofibration of orthogonal $G$-spectra and $j:K{\longrightarrow}L$ a $G$-cofibration of based $G$-spaces. Then the pushout-product morphism $$i\boxempty j \ = \ (i{\wedge}L)\cup(B{\wedge}j)\ : \
A{\wedge}L\cup_{A{\wedge}K}B{\wedge}K \ {\longrightarrow}\ B{\wedge}L$$ is a cofibration of orthogonal $G$-spectra. If moreover $i$ is a $\pi_*$-isomorphism or $j$ is a ${\mathcal Com}$-equivalence, then $i \boxempty j$ is a $\pi_*$-isomorphism.
4. The stable model structure is proper, stable, topological and cofibrantly generated.
5. The fibrant objects in the stable model structure are the $G$-$\Omega$-spectra.
\(i) We start by establishing a level model structure for orthogonal $G$-spectra. Orthogonal $G$-spectra are continuous based functors from ${\mathbf O}$ to the category $G{\mathbf{T}}_*$, so we can employ the machinery from [@schwede:global Prop.C.23] that produces level model structures on enriched functor categories. Here the base category is ${\mathcal V}=G{\mathbf{T}}_*$, with smash product as monoidal structure. The index category is ${\mathcal D}={\mathbf O}$, where the group $G$ acts trivially on all morphism spaces. The category ${\mathcal D}^*$ of enriched functors from ${\mathbf O}$ to $G{\mathbf{T}}_*$ then becomes $\operatorname{Sp}_G$. The dimension function on ${\mathbf O}$ is the vector space dimension, and then the abstract skeleta of [@schwede:global Con.C.13] specialize to the skeleta above.
To apply [@schwede:global Prop.C.23] we need to specify a model structure on the category of $O(m)$- | _ * $ -isomorphisms, stable fibrations and cofibrations form a model structure on the class of extraneous $ G$-spectra, the [ * static model structure * ] { }.
2. Every cofibration of orthogonal $ G$-spectra is an henry - cofibration.
3. Let $ i: A{\longrightarrow}B$ be a cofibration of orthogonal $ G$-spectra and $ joule: K{\longrightarrow}L$ a $ G$-cofibration of based $ G$-spaces. Then the pushout - product morphism $ $ i\boxempty j \ = \ (i{\wedge}L)\cup(B{\wedge}j)\: \
A{\wedge}L\cup_{A{\wedge}K}B{\wedge}K \ { \longrightarrow}\ B{\wedge}L$$ is a cofibration of extraneous $ G$-spectra. If furthermore $ i$ is a $ \pi_*$-isomorphism or $ j$ is a $ { \mathcal Com}$-equivalence, then $ i \boxempty j$ is a $ \pi_*$-isomorphism.
4. The stable model social organization is proper, stable, topological and cofibrantly generated.
5. The fibrant object in the stable model structure are the $ G$-$\Omega$-spectra.
\(i) We start by establishing a level exemplar structure for orthogonal $ G$-spectra. Orthogonal $ G$-spectra are continuous based functors from $ { \mathbf O}$ to the class $ G{\mathbf{T}}_*$, so we can employ the machinery from [ @schwede: global Prop. C.23 ] that grow level exemplar structures on enriched functor categories. Here the base category is $ { \mathcal V}=G{\mathbf{T}}_*$, with smash product as monoidal social organization. The index category is $ { \mathcal D}={\mathbf O}$, where the group $ G$ acts trivially on all morphism spaces. The category $ { \mathcal D}^*$ of enriched functors from $ { \mathbf O}$ to $ G{\mathbf{T}}_*$ then becomes $ \operatorname{Sp}_G$. The property function on $ { \mathbf O}$ is the vector space dimension, and then the abstract skeleta of [ @schwede: global Con. C.13 ] specialize to the skeleta above.
To apply [ @schwede: global Prop. C.23 ] we need to specify a model social organization on the class of $ O(m)$- | _*$-isolorphisms, stable fibratioks and cofibrations fork a mosel struzture on the category of ortiogobal $G$-wpectra, the [*stable moddl structlre*]{}.
2. Evert cohibration of ortikgonal $N$-fpecfva is en h-cofibration.
3. Let $i:A{\lotgrightarrow}B$ te a cofibration of orthogonal $G$-spectra and $j:K{\kojgrightarrow}L$ w $G$-cpsibrzniin of based $G$-spaces. Then the lushout-kroduct morphism $$i\noxempty j \ = \ (i{\wedge}L)\cup(B{\wfdge}u)\ : \
A{\wedge}P\cup_{A{\wedge}K}V{\weddw}K \ {\longrighgarrow}\ B{\wedge}L$$ is a coribration of orthogonal $G$-spectrx. If koreover $i$ iw a $\ki_*$-isomorphisn or $t$ is a ${\mathcal Com}$-eqgivalenve, then $i \boxekptb j$ us a $\pi_*$-isomorphism.
4. Tie stable model strusture is [rkper, stable, topoligucal dnd woficeangly gxnedated.
5. Thx fibrant ogjects in tye stable model stricelte are the $G$-$\Kmega$-s[estra.
\(i) We start by establishing a level kodsl structure for orthoginal $G$-spectra. Orthogojal $G$-specera are continuous based functors from ${\mathbf O}$ tm the zattgiry $G{\oqtjbf{T}}_*$, so we can employ the machinery from [@schwqse:blpbal Prop.C.23] thab produces level mpdfl ftructures on enricksd functor categoried. Here jhe bawe categowy ix ${\mathcal V}=G{\mathbf{T}}_*$, with smqsh product cs nonoidal structure. The index eategoty is ${\kathcal D}={\mathbf O}$, where the group $G$ achs triviamuy on all morphirm xpdces. The category ${\mathcal Q}^*$ of enrirhed yunctors frok ${\mathff O}$ to $G{\mwthbf{B}}_*$ then becomes $\operwtorncme{Sp}_C$. The dimejsion function on ${\mathbf O}$ is tix vector spacg dhmetsion, anb then the abstract skeleta of [@sckwede:gloyal Cov.C.13] specialpze to thx skeleta abjve.
To apply [@swjwede:global 'rop.C.23] we geed to wpecify x model structire on thv eategory if $O(m)$- | _*$-isomorphisms, stable fibrations and cofibrations form a on category of $G$-spectra, the [*stable of $G$-spectra is an 3. Let $i:A{\longrightarrow}B$ a cofibration of orthogonal $G$-spectra and a $G$-cofibration of based $G$-spaces. Then the pushout-product morphism $$i\boxempty j \ = (i{\wedge}L)\cup(B{\wedge}j)\ : \ A{\wedge}L\cup_{A{\wedge}K}B{\wedge}K \ {\longrightarrow}\ B{\wedge}L$$ is a cofibration of orthogonal $G$-spectra. moreover is $\pi_*$-isomorphism $j$ is a ${\mathcal Com}$-equivalence, then $i \boxempty j$ is a $\pi_*$-isomorphism. 4. The stable model is proper, stable, topological and cofibrantly generated. 5. fibrant objects in the model structure are the $G$-$\Omega$-spectra. We by establishing level structure orthogonal $G$-spectra. Orthogonal are continuous based functors from ${\mathbf O}$ to the category $G{\mathbf{T}}_*$, so we can employ the machinery [@schwede:global Prop.C.23] level model on functor Here the base ${\mathcal V}=G{\mathbf{T}}_*$, with smash product as index category is ${\mathcal D}={\mathbf O}$, where the $G$ acts on all morphism spaces. The category D}^*$ of enriched functors from ${\mathbf O}$ to then becomes $\operatorname{Sp}_G$. The dimension function on ${\mathbf O}$ is the vector space dimension, and abstract skeleta of [@schwede:global specialize to the above. apply Prop.C.23] need to a model structure on the category of $O(m)$- | _*$-isomorphisms, stable fibratiOns and cofiBratiOns ForM a ModeL strUcture on the catEGory Of orthogonal $G$-spectra, thE [*stabLe MOdel STrUcturE*]{}.
2. Every cOFiBRAtiOn Of OrtHoGOnAl $G$-spEctRa is an h-CofibratioN.
3. LeT $i:a{\longrightarROw}b$ be a cofibrAtiOn of orthogonAl $G$-SpectrA aNd $j:k{\LongrIghTarroW}L$ a $G$-coFIbratiOn of based $g$-sPAces. ThEN the pusHOUt-ProdUct morphism $$i\boxemPTy J \ = \ (I{\wedge}L)\cup(B{\wedGe}j)\ : \
A{\weDgE}l\cUP_{a{\weDge}k}B{\wedge}K \ {\loNgRightARrow}\ B{\weDGe}l$$ IS A coFIbration of ortHogonal $G$-speCTra. if moreOvEr $i$ IS a $\pi_*$-isOmorpHiSM or $J$ is a ${\mathcal com}$-eQuivalencE, then $i \BOxempty J$ Is a $\pi_*$-isOmorphIsm.
4. the StabLE mOdEl sTrUCtuRE iS prOPer, Stable, toPoLoGical And cOFIBRantLy gEnerAted.
5. THe fibrant objeCts In thE StaBle moDel stRuctUrE are tHe $G$-$\OmeGa$-speCtRa.
\(i) We start by estAbliShing a levEl mOdEl sTrUcturE For ortHogOnaL $G$-spectRa. OrthoGOnaL $G$-SPECtRa are continuous basEd FUNcTors from ${\Mathbf o}$ To ThE Category $g{\mAthBf{T}}_*$, sO WE can eMploY ThE machineRy from [@SChWeDe:globaL PRop.C.23] thAt ProDucEs levEL modEl struCtures on EnricHEd functor categORies. Here the baSE cATEgORy is ${\MatHcal V}=G{\mathbF{T}}_*$, wiTH smaSh prODuCt aS MonoiDal stRuCTuRE. The index category is ${\MaThcal D}={\MathbF O}$, where the groUp $G$ acts triVIALly on all MorpHIsM Spaces. The categOry ${\maThcal D}^*$ of enRIched funCtors From ${\mathBf O}$ to $G{\matHBF{T}}_*$ then beComEs $\oPerAtoRNAmE{Sp}_G$. The dimensION funCtIon on ${\maThbF O}$ is the VecTor SpaCe dImEnsion, and Then the aBsTrAcT sKelEta of [@SChwede:glObAl COn.c.13] spEcialIZe to thE skelEta aBoVe.
tO apPly [@schwEDe:GLObal prOp.c.23] we nEed To SpeciFy a mODel StructuRe on the caTegORy of $o(m)$- | _*$-isomorphisms, stable f ibrationsand c ofi bra ti onsform a model struc t ureon the category of ort hogon al $G$- s pe ctra, the [* s ta b l e m od el st ru c tu re*]{ }.
2. Ev ery cofibr ati on of orthogon a l$G$-spectr a i s an h-cofib rat ion.
3. L e t $i: A{\ longr ightar r ow}B$be a cofi br a tion o f orthog o n al $G$ -spectra and $j:K { \l o ngrightarrow}L $ a $G $- c of i b rat ion of based$G $-spa c es. The n t h e pus h out-product m orphism $$i \ box emptyj\ = \ (i{\ wedge }L ) \cu p(B{\wedge} j)\: \
A{ \ wedge}L \ cup_{A{ \wedge }K} B{\ wedg e }K \ {\ lo n gri g ht arr o w}\ B{\wedg e} L$ $ isa co f i b r atio n o f or thogo nal $G$-spect ra. Ifm ore over$i$ i s a$\ pi_*$ -isomo rphis mor $j$ is a ${\ math cal Com}$ -eq ui val en ce, t h en $i\bo xem pty j$is a $\ p i_* $- i s o mo rphism.
4. The s ta b l emodel st ructur e i sp roper, s ta ble , to p o logic al a n dcofibran tly ge n er at ed.
5. The fi br ant ob jects in t he sta ble mode l str u cture are the$ G$-$\Omega$-s p ec t r a.
\(i ) W e start byesta b lish inga l eve l mode l str uc t ur e for orthogonal $G$ -s pectra . Ort hogonal $G$-s pectra are c o ntinuous bas e df unctors from $ {\mat hbf O}$ to the cate gory$G{\math bf{T}}_*$ , so we ca n e mpl oythe m ac hinery from [ @ s chwe de :global Pr op.C.23 ] t hat pr odu ce s level m odel str uc tu re sonenric h ed funct or ca te gor ies.H ere th e bas e ca te go r y i s ${\ma t hc a l V}= G{ \m athb f{T }} _*$,with sma sh prod uct as mo noi d al s tr uc ture. T he index cate go ry is ${\m at hca l D}={ \ m athbf O} $, where the group $G$a cts tri via lly o n al l morphis m s paces. Th e categ ory ${ \math ca l D } ^ *$ of e nr ich ed functorsf r om${\ma th bf O }$ to $ G{\mathbf{T}}_*$ t h enbecomes $\ope rat orna m e {S p}_ G $. The d i men s i on function on${\mathbfO} $ i s the vect o r s pa ce dime nsion,and t h en theabstractskeleta o f[@sc h w ede :global Co n.C.13]specializ e to t h eskele taabove.
Toapply [@sch w ede :glob al Pro p. C.23]we ne ed to spec ify a model structure o n thecateg ory of $O(m) $- | _*$-isomorphisms, stable fibrations_and cofibrations_form a model structure_on the_category_of orthogonal_$G$-spectra,_the [*stable model_structure*]{}.
2. Every_cofibration of orthogonal $G$-spectra_is an h-cofibration.
3.__Let $i:A{\longrightarrow}B$ be a cofibration of orthogonal $G$-spectra and $j:K{\longrightarrow}L$ a $G$-cofibration of based_$G$-spaces._Then the_pushout-product_morphism_$$i\boxempty j \ = \_(i{\wedge}L)\cup(B{\wedge}j)\ : \
_ _ A{\wedge}L\cup_{A{\wedge}K}B{\wedge}K \ {\longrightarrow}\ B{\wedge}L$$_is_a cofibration of_orthogonal $G$-spectra. If moreover $i$ is a $\pi_*$-isomorphism or_$j$ is a ${\mathcal Com}$-equivalence, then_$i \boxempty j$_is_a_$\pi_*$-isomorphism.
4. The stable_model structure is proper, stable, topological_and cofibrantly generated.
5. The fibrant_objects in the stable model structure are_the $G$-$\Omega$-spectra.
\(i) We start by establishing_a level model structure for_orthogonal $G$-spectra._Orthogonal $G$-spectra are continuous based_functors from ${\mathbf_O}$ to_the category $G{\mathbf{T}}_*$,_so we can employ the machinery_from [@schwede:global Prop.C.23]_that produces level model structures on_enriched_functor categories. Here_the_base_category is_${\mathcal V}=G{\mathbf{T}}_*$, with_smash_product as_monoidal_structure. The index category is ${\mathcal_D}={\mathbf_O}$, where the group $G$ acts trivially_on all morphism spaces._The_category ${\mathcal D}^*$ of_enriched functors from ${\mathbf O}$_to $G{\mathbf{T}}_*$ then becomes $\operatorname{Sp}_G$. The_dimension function_on ${\mathbf_O}$ is the vector space dimension, and then the abstract skeleta_of [@schwede:global Con.C.13] specialize to the_skeleta above.
To apply [@schwede:global_Prop.C.23] we_need_to specify a_model_structure on_the category of $O(m)$- |
to the maximum value ($\sim 1$) for high-speed boosts.
Otherwise, a maximally entangled spin-spin state in $\mathcal{S}$ can be constructed through an anti-symmetric superposition between positive helicities $$\label{state2B}
\vert \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\bb{\bm{q}} \rangle^{B} - \vert u_1\bb{\bm{q}} \rangle^{A} \otimes \vert u_1\bb{\bm{p}} \rangle^{B}}{\sqrt{2}},$$ which, according to the correspondence from (\[hel\]), indeed can be recast as $$\label{state2}
\vert \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\bb{\bm{q}} \rangle^{B} - \vert u_2\bb{\bm{p}} \rangle^{A} \otimes \vert u_2\bb{\bm{q}} \rangle^{B}}{\sqrt{2}},$$ which corresponds to a much more realistic configuration, for which particles in the subspace $A$ and $B$ have well defined momenta, $\bm{p}$ and $\bm{q}$, respectively, in agreement with the construction from the previous section. Fig. \[Graph02\] depicts the variation of the global and the spin-spin entanglement of $\vert \psi_2 \rangle$ as function of the boost rapidity $\omega$. In this case, the variation of entanglement is independent of the boost angle and, as for the state from Eq. (\[state1\]), the global entanglement increases under Lorentz boosts. On the other hand, spin-spin entanglement is degraded by the boost transformation and for high speed boosts the spin-spin state is completely separable.
A third anti-symmetric configuration is given by $$\label{state3}
\vert \psi_3 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_2\bb{\bm{p}} \rangle^{B} -\vert u_2\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\bb{\bm{p}} \rangle^{B}}{\sqrt{ | to the maximum value ($ \sim 1 $) for high - speed boosts.
differently, a maximally entangle spin - spin department of state in $ \mathcal{S}$ can be manufacture through an anti - symmetric superposition between positivist helicities $ $ \label{state2B }
\vert \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p } } \rangle^{A } \otimes \vert u_1\bb{\bm{q } } \rangle^{B } - \vert u_1\bb{\bm{q } } \rangle^{A } \otimes \vert u_1\bb{\bm{p } } \rangle^{B}}{\sqrt{2}},$$ which, accord to the correspondence from (\[hel\ ]), indeed can be recast as $ $ \label{state2 }
\vert \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p } } \rangle^{A } \otimes \vert u_1\bb{\bm{q } } \rangle^{B } - \vert u_2\bb{\bm{p } } \rangle^{A } \otimes \vert u_2\bb{\bm{q } } \rangle^{B}}{\sqrt{2}},$$ which match to a much more realistic shape, for which particles in the subspace $ A$ and $ B$ have well defined momenta, $ \bm{p}$ and $ \bm{q}$, respectively, in agreement with the construction from the previous incision. Fig. \[Graph02\ ] depicts the variation of the global and the spin - spin web of $ \vert \psi_2 \rangle$ as function of the boost rapidity $ \omega$. In this lawsuit, the variation of entanglement is autonomous of the boost slant and, as for the state from Eq. (\[state1\ ]), the global entanglement increases under Lorentz boosts. On the other bridge player, spin - spin entanglement is degraded by the boost transformation and for high speed boosts the spin - spin state is completely separable.
A third anti - symmetric configuration is given by $ $ \label{state3 }
\vert \psi_3 \rangle = \frac{\vert u_1\bb{\bm{p } } \rangle^{A } \otimes \vert u_2\bb{\bm{p } } \rangle^{B } -\vert u_2\bb{\bm{p } } \rangle^{A } \otimes \vert u_1\bb{\bm{p } } \rangle^{B}}{\sqrt { | to the maximum value ($\sim 1$) nor high-speed boosts.
Othxrwise, z maximauly entangled spin-spin state ib $\matycal{S}$ can be constructdd througj an antu-symnwtric supeclositiok betscen pmwitive helicitles $$\label{stdte2B}
\vert \psi_2 \rdnelz = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vewt u_1\bb{\bk{q}} \rangle^{B} - \vert u_1\bb{\ni{q}} \rzngle^{A} \otimes \vert u_1\bb{\bm{p}} \rangle^{B}}{\ssrt{2}},$$ whibh, according to tne correspondence from (\[hel\]), indfed can be recast ws $$\label{stajs2}
\vewr \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangme^{A} \otimes \vert u_1\bb{\bm{q}} \rangle^{B} - \vert u_2\bb{\bm{p}} \rabgoe^{A} \mtimes \vert u_2\bb{\bi{q}} \rangle^{B}}{\sqvn{2}},$$ which correslonds to a mucm morx reqlistic configuration, for which particlgs in the vuyspace $A$ and $B$ have wwlo defhned momdbta, $\bm{k}$ aid $\gm{q}$, redpertively, in zgreement wuth the constructiom sgpm the previkus sestyon. Fig. \[Graph02\] depicts the variation of tve flobal and the spin-spin entanglement of $\vert \psi_2 \rangje$ as function of the boost rapidity $\omega$. In thiv casx, ghe vwfuahion of entanglement is independent of the bojat akgle and, as for bhe state from Eq. (\[syahe1\]), jhe global entxnglemzht increases under Llrentz foostw. On the jthet hand, spin-spin entanglemenr is degradeb bt the boost transflrmation anb for nigh xpeed boosts the spin-spnn stafe is complftely sepzfable.
A third antk-sykmatric configuration is givqn by $$\labxl{stace3}
\vert \pri_3 \rsngle = \frac{\vert k_1\bb{\bm{i}} \rangle^{A} \otimes \vegt u_2\by{\bm{p}} \sangle^{B} -\vegt u_2\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\uu{\bm{p}} \rangle^{B}}{\sard{ | to the maximum value ($\sim 1$) for Otherwise, maximally entangled state in $\mathcal{S}$ anti-symmetric between positive helicities \vert \psi_2 \rangle \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\bb{\bm{q}} - \vert u_1\bb{\bm{q}} \rangle^{A} \otimes \vert u_1\bb{\bm{p}} \rangle^{B}}{\sqrt{2}},$$ which, according to the correspondence (\[hel\]), indeed can be recast as $$\label{state2} \vert \psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \otimes u_1\bb{\bm{q}} - u_2\bb{\bm{p}} \rangle^{A} \otimes \vert u_2\bb{\bm{q}} \rangle^{B}}{\sqrt{2}},$$ which corresponds to a much more realistic configuration, for which in the subspace $A$ and $B$ have well momenta, $\bm{p}$ and $\bm{q}$, in agreement with the construction the section. Fig. depicts variation the global and spin-spin entanglement of $\vert \psi_2 \rangle$ as function of the boost rapidity $\omega$. In this case, the of entanglement of the angle as the state from the global entanglement increases under Lorentz other hand, spin-spin entanglement is degraded by the transformation and high speed boosts the spin-spin state completely separable. A third anti-symmetric configuration is given $$\label{state3} \vert \psi_3 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_2\bb{\bm{p}} \rangle^{B} -\vert u_2\bb{\bm{p}} \rangle^{A} u_1\bb{\bm{p}} \rangle^{B}}{\sqrt{ | to the maximum value ($\sim 1$) for hiGh-speed booSts.
OtHerWisE, a MaxiMallY entangled spin-SPin sTate in $\mathcal{S}$ can be conStrucTeD ThroUGh An antI-symmetRIc SUPerPoSiTioN bETwEen poSitIve heliCities $$\labeL{stAtE2B}
\vert \psi_2 \ranGLe = \Frac{\vert u_1\bB{\bm{P}} \rangle^{A} \otimEs \vErt u_1\bb{\Bm{Q}} \raNGle^{B} - \vErt U_1\bb{\bm{Q}} \ranglE^{a} \otimeS \vert u_1\bb{\bM{p}} \RAngle^{B}}{\SQrt{2}},$$ whicH, ACcOrdiNg to the correspondENcE From (\[hel\]), indeed cAn be reCaST aS $$\LAbeL{stAte2}
\vert \psi_2 \RaNgle = \fRAc{\vert u_1\BB{\bM{P}} \RAngLE^{A} \otimes \vert u_1\Bb{\bm{q}} \rangle^{b} - \VerT u_2\bb{\bm{P}} \rAngLE^{A} \otimEs \verT u_2\BB{\bm{Q}} \rangle^{B}}{\sqrT{2}},$$ whiCh correspOnds to A Much morE RealistIc confIguRatIon, fOR wHiCh pArTIclES iN thE SubSpace $A$ anD $B$ HaVe welL defINED MomeNta, $\Bm{p}$ aNd $\bm{q}$, Respectively, iN agReemENt wIth thE consTrucTiOn froM the prEviouS sEction. Fig. \[Graph02\] dEpicTs the variAtiOn Of tHe GlobaL And the SpiN-spIn entanGlement OF $\veRt \PSI_2 \RaNgle$ as function of thE bOOSt Rapidity $\Omega$. IN ThIs CAse, the vaRiAtiOn of ENTanglEmenT Is IndependEnt of tHE bOoSt angle AnD, as for ThE stAte From EQ. (\[StatE1\]), the glObal entaNglemENt increases undER Lorentz boostS. on THE oTHer hAnd, Spin-spin entAnglEMent Is deGRaDed BY the bOost tRaNSfORmation and for high spEeD boostS the sPin-spin state iS completelY SEParable.
A ThirD AnTI-symmetric confIguraTion is giveN By $$\label{sTate3}
\vErt \psi_3 \raNgle = \frac{\vERT u_1\bb{\bm{p}} \rAngLe^{A} \OtiMes \VERt U_2\bb{\bm{p}} \rangle^{B} -\VERt u_2\bB{\bM{p}} \ranglE^{A} \oTimes \veRt u_1\Bb{\bM{p}} \rAngLe^{b}}{\sqrt{ | to the maximum value ($\s im 1$) for high -sp eed b oost s.
Otherwise, a m a xima lly entangled spin-spi n sta te in $ \ ma thcal {S}$ ca n b e con st ru cte dt hr oughananti-sy mmetric su per po sition betwe e npositive h eli cities $$\la bel {state 2B }
\ v ert \ psi _2 \r angle= \frac {\vert u_ 1\ b b{\bm{ p }} \ran g l e^ {A}\otimes \vert u_1 \ bb { \bm{q}} \rangl e^{B}-\ ve r t u_ 1\b b{\bm{q}}\r angle ^ {A} \ot i me s \ ver t u_1\bb{\bm{p }} \rangle^ { B}} {\sqrt {2 }}, $ $ whic h, ac co r din g to the co rres pondencefrom ( \ [hel\]) , indeed can b e r eca st a s $ $\ lab el { sta t e2 }
\ v ert \psi_2\r an gle = \fr a c { \ vert u_ 1\bb {\bm{ p}} \rangle^{ A}\oti m es\vert u_1\ bb{\ bm {q}}\rangl e^{B} - \vert u_2\bb{\ bm{p }} \rangl e^{ A} \o ti mes \ v ert u_ 2\b b{\ bm{q}}\rangle ^ {B} }{ \ s q rt {2}},$$ which corr es p o nd s to a m uch mo r ere a listic c on fig urat i o n, fo r wh i ch particl es int he s ubspace $ A$ and $ B$hav e wel l def ined m omenta,$\bm{ p }$ and $\bm{q} $ , respectivel y ,i n a g reem ent with the c onst r ucti on f r om th e prev iousse c ti o n. Fig. \[Graph02\] d epicts thevariation ofthe global a n d the sp in-s p in entanglement o f $\v ert \psi_2 \rangle$ as f unctionof the bo o s t rapidi ty$\o meg a$. I nthis case, th e vari at ion ofent angleme ntisind epe nd ent of th e boostan gl ean d,as fo r the sta te fr om Eq . (\[ s tate1\ ]), t he g lo ba l en tanglem e nt i ncre as es und erLo rentz boo s ts. On the other ha nd, spin -s pi n entan glement is de gr aded by th eboo st tra n s formatio n and for high speed bo o sts the sp in-sp in s tate is c omp letely se p arable .
A t hirdan ti- s y mmetr i c c onf ig uration is g ive n by$$ \lab el{stat e3}
\vert \psi_3 \ r ang le = \frac{\v ert u_1 \ b b{ \bm { p} } \r an g le^ { A } \otimes \vert u_2\bb{\b m{ p }} \rangle^{ B } - \v ert u_2 \bb{\bm {p}}\ rangle^ {A} \otim es \vertu_ 1\bb { \ bm{ p}} \rangl e^{B}}{\ sqrt{ | to_the maximum_value ($\sim 1$) for_high-speed boosts.
Otherwise,_a_maximally entangled_spin-spin_state in $\mathcal{S}$_can be constructed_through an anti-symmetric superposition_between positive helicities_$$\label{state2B}
\vert_\psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert u_1\bb{\bm{q}} \rangle^{B} - \vert u_1\bb{\bm{q}} \rangle^{A}_\otimes_\vert u_1\bb{\bm{p}}_\rangle^{B}}{\sqrt{2}},$$_which,_according to the correspondence from_(\[hel\]), indeed can be recast_as $$\label{state2}
\vert_\psi_2 \rangle = \frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert_u_1\bb{\bm{q}}_\rangle^{B} - \vert_u_2\bb{\bm{p}} \rangle^{A} \otimes \vert u_2\bb{\bm{q}} \rangle^{B}}{\sqrt{2}},$$ which corresponds to_a much more realistic configuration, for_which particles in_the_subspace_$A$ and $B$ have_well defined momenta, $\bm{p}$ and $\bm{q}$,_respectively, in agreement with the construction_from the previous section. Fig. \[Graph02\] depicts the_variation of the global and the_spin-spin entanglement of $\vert \psi_2_\rangle$ as_function of the boost rapidity_$\omega$. In this_case, the_variation of entanglement_is independent of the boost angle_and, as for_the state from Eq. (\[state1\]), the global_entanglement_increases under Lorentz_boosts._On_the other_hand, spin-spin entanglement_is_degraded by_the_boost transformation and for high speed_boosts_the spin-spin state is completely separable.
A third_anti-symmetric configuration is given_by_$$\label{state3}
\vert \psi_3 \rangle =_\frac{\vert u_1\bb{\bm{p}} \rangle^{A} \otimes \vert_u_2\bb{\bm{p}} \rangle^{B} -\vert u_2\bb{\bm{p}} \rangle^{A} \otimes_\vert u_1\bb{\bm{p}}_\rangle^{B}}{\sqrt{ |
choose 2}} \ge \frac{t^2}2\left( \frac{|A\cap I|^2}{k}+
\frac{|A\setminus I|^2}{n-k}\right) +o(n^{3/2})= t^2\left(\frac12 +
\frac{(\alpha n-k)^2}{k(n-k)}\right) +o(n^{3/2}).$$
Thus, up to an error term of $o(n^{3/2})$, we must have equality throughout. We conclude that $\alpha=k/n+o(1)$ and $a_i=(\alpha/m+o(1))\,n^{1/2}$, which gives the required.
Lower Bounds on $s(k,n)$ {#SLower}
========================
We know that the range of interest is $k=\Theta(n^{1/2})$. We will be proving lower bounds on the following ‘scaled’ one-parameter version of $s(k,n)$: $$\label{eq:DefS}
s(c):=\liminf_{n\to\infty} \frac{s({\lfloor cn^{1/2}\rfloor},n)}n.$$
Note that in we could have replaced ${\lfloor cn^{1/2}\rfloor}$ by anything of the form $(c+o(1))\, n^{1/2}$ without affecting the value of $s(c)$. However, we have to write $\liminf$ as the following question is open.
\[pr:limit\] Let $c$ be a fixed positive real. Suppose that $n$ tends to the infinity and $k=(c+o(1))\, n^{1/2}$. Does the ratio $s(k,n)/n$ tend to a limit?
Our lower bound on $s(c)$, provided by the following lemmata, will be given by different formulae for different ranges of $c$.
The bound of Erdős and Freud [@erdos+freud:91] implies that $$\label{eq:1}
s(c)=\frac{c^2}2,\quad c\le 2/\sqrt3.$$ Their construction can be generalised to give lower bounds on $s(c)$ for larger $c$.
\[lm:SLower1\] $$\label{eq:SL | choose 2 } } \ge \frac{t^2}2\left (\frac{|A\cap I|^2}{k}+
\frac{|A\setminus I|^2}{n - k}\right) + o(n^{3/2})= t^2\left(\frac12 +
\frac{(\alpha n - k)^2}{k(n - k)}\right) + o(n^{3/2}).$$
Thus, up to an error term of $ o(n^{3/2})$, we must have equality passim. We reason that $ \alpha = kilobyte / n+o(1)$ and $ a_i=(\alpha / m+o(1))\,n^{1/2}$, which gives the required.
low Bounds on $ s(k, n)$ { # SLower }
= = = = = = = = = = = = = = = = = = = = = = = =
We know that the stove of pastime is $ k=\Theta(n^{1/2})$. We will be proving broken bounds on the following ‘ scale ’ one - parameter version of $ s(k, n)$: $ $ \label{eq: DefS }
s(c):=\liminf_{n\to\infty } \frac{s({\lfloor cn^{1/2}\rfloor},n)}n.$$
notice that in we could have replaced $ { \lfloor cn^{1/2}\rfloor}$ by anything of the form $ (c+o(1))\, n^{1/2}$ without affecting the value of $ s(c)$. However, we have to write $ \liminf$ as the following question is exposed.
\[pr: limit\ ] Let $ c$ be a fixed positive real number. Suppose that $ n$ tends to the eternity and $ k=(c+o(1))\, n^{1/2}$. Does the proportion $ s(k, n)/n$ tend to a limit?
Our lower bound on $ s(c)$, provided by the following lemmata, will be given by different rule for different ranges of $ c$.
The bound of Erdős and Freud [ @erdos+freud:91 ] implies that $ $ \label{eq:1 }
s(c)=\frac{c^2}2,\quad c\le 2/\sqrt3.$$ Their construction can be generalised to give lower bounds on $ s(c)$ for larger $ c$.
\[lm: SLower1\ ] $ $ \label{eq: SL | cholse 2}} \ge \frac{t^2}2\left( \frac{|A\cxp I|^2}{k}+
\frac{|A\setmnbus I|^2}{i-k}\right) +o(n^{3/2})= t^2\lefg(\frac12 +
\frac{(\alpha n-k)^2}{k(n-k)}\right) +o(i^{3/2}).$$
Thuw, up uj an error term of $o(n^{3/2})$, we mudt have wquaouty througikut. We gjnclhfe tkav $\alpha=k/n+o(1)$ and $s_i=(\alpha/m+o(1))\,n^{1/2}$, which gives tve rzquired.
Lower Bounds on $s(k,n)$ {#SLower}
========================
We hnow thst the range of yntegeft ia $k=\Theta(n^{1/2})$. We will be proving lower bounds on the folloeing ‘scaled’ one-parameter vfrsiln of $s(k,n)$: $$\label{eq:DffS}
s(c):=\limind_{n\to\ybfty} \frac{s({\lfuoor cn^{1/2}\rfloor},n)}n.$$
Note thzt in we could have replaced ${\lfuoor en^{1/2}\rfloor}$ by abytjhng of the horm $(c+j(1))\, n^{1/2}$ without affectinc the vslue of $s(c)$. Howcver, xe hqve to write $\liminf$ av the following qugstion is mpzn.
\[pr:limit\] Let $c$ be a dized pmsithve fwal. Sukpoxe that $j$ txnds to the infinity abd $k=(c+o(1))\, n^{1/2}$. Does the rstyi $s(k,n)/n$ tend tk a liiie?
Our lower bound on $s(c)$, provided by the xolmowing lemmata, will be tiven by different fotmulae for different ranges of $c$.
The bound of Erdős and Freud [@ardos+hrdud:91] iiooifs that $$\label{eq:1}
s(c)=\frac{c^2}2,\quad c\le 2/\sqrt3.$$ Their cogattubtion can be genevalised to give loeeg njunds on $s(c)$ fur larysr $c$.
\[lm:SLower1\] $$\label{eq:DL | choose 2}} \ge \frac{t^2}2\left( \frac{|A\cap I|^2}{k}+ \frac{|A\setminus t^2\left(\frac12 \frac{(\alpha n-k)^2}{k(n-k)}\right) Thus, up to we have equality throughout. conclude that $\alpha=k/n+o(1)$ $a_i=(\alpha/m+o(1))\,n^{1/2}$, which gives the required. Lower on $s(k,n)$ {#SLower} ======================== We know that the range of interest is $k=\Theta(n^{1/2})$. will be proving lower bounds on the following ‘scaled’ one-parameter version of $s(k,n)$: s(c):=\liminf_{n\to\infty} cn^{1/2}\rfloor},n)}n.$$ that we could have replaced ${\lfloor cn^{1/2}\rfloor}$ by anything of the form $(c+o(1))\, n^{1/2}$ without affecting the of $s(c)$. However, we have to write $\liminf$ the following question is \[pr:limit\] Let $c$ be a positive Suppose that tends the and $k=(c+o(1))\, n^{1/2}$. the ratio $s(k,n)/n$ tend to a limit? Our lower bound on $s(c)$, provided by the following lemmata, be given formulae for ranges $c$. bound of Erdős [@erdos+freud:91] implies that $$\label{eq:1} s(c)=\frac{c^2}2,\quad c\le can be generalised to give lower bounds on for larger \[lm:SLower1\] $$\label{eq:SL | choose 2}} \ge \frac{t^2}2\left( \frac{|A\cap i|^2}{k}+
\frac{|A\setMinus i|^2}{n-k}\RigHt) +O(n^{3/2})= t^2\lEft(\fRac12 +
\frac{(\alpha n-k)^2}{K(N-k)}\riGht) +o(n^{3/2}).$$
Thus, up to an error teRm of $o(N^{3/2})$, wE Must HAvE equaLity thrOUgHOUt. WE cOnCluDe THaT $\alphA=k/n+O(1)$ and $a_i=(\aLpha/m+o(1))\,n^{1/2}$, whiCh gIvEs the requireD.
loWer Bounds oN $s(k,N)$ {#SLower}
========================
We knoW thAt the rAnGe oF InterEst Is $k=\ThEta(n^{1/2})$. We WIll be pRoving lowEr BOunds oN The follOWInG ‘scaLed’ one-parameter veRSiON of $s(k,n)$: $$\label{eq:DEfS}
s(c):=\lImINf_{N\TO\inFty} \Frac{s({\lflooR cN^{1/2}\rfloOR},n)}n.$$
Note THaT IN We cOUld have replacEd ${\lfloor cn^{1/2}\rFLooR}$ by anyThIng OF the foRm $(c+o(1))\, n^{1/2}$ WiTHouT affecting tHe vaLue of $s(c)$. HoWever, wE Have to wRIte $\limiNf$ as thE foLloWing QUeStIon Is OPen.
\[PR:lImiT\] let $C$ be a fixeD pOsItive Real. sUPPOse tHat $N$ tenDs to tHe infinity and $K=(c+o(1))\, N^{1/2}$. DoeS The Ratio $S(k,n)/n$ tEnd tO a Limit?
our lowEr bouNd On $s(c)$, provided by tHe foLlowing leMmaTa, WilL bE giveN By diffEreNt fOrmulae For diffERenT rANGEs Of $c$.
The bound of Erdős AnD fReUd [@erdos+fReud:91] imPLiEs THat $$\label{Eq:1}
S(c)=\fRac{c^2}2,\QUAd c\le 2/\Sqrt3.$$ tHeIr constrUction CAn Be GeneralIsEd to giVe LowEr bOunds ON $s(c)$ fOr largEr $c$.
\[lm:SLoWer1\] $$\laBEl{eq:SL | choose 2}} \ge \frac{t^2}2 \left( \fr ac{|A \ca p I| ^2}{ k}+\frac{|A\setmi n us I|^2}{n-k}\right) +o(n ^{3/2 }) = t^2 \ le ft(\f rac12 + \f r a c{( \a lp han- k )^ 2}{k( n-k )}\righ t) +o(n^{3 /2} ). $$
Thus, up to an errorter m of $o(n^{3 /2} )$, we m ust haveequ ality throu g hout.We conclu de that $ \ alpha=k / n +o (1)$ and $a_i=(\alpha / m+ o (1))\,n^{1/2}$ , whic hg iv e s th e r equired.
Lo wer B o unds on $s ( k , n)$ {#SLower}
=== =========== = === ======
Wek now th at th er ang e of intere st i s $k=\The ta(n^{ 1 /2})$.W e willbe pro vin g l ower bo un dson the fo llo w ing ‘scaled ’on e-par amet e r v ersi onof $ s(k,n )$: $$\label{ eq: DefS }
s (c):= \limi nf_{ n\ to\in fty} \ frac{ s( {\lfloor cn^{1/ 2}\r floor},n) }n. $$
N ot e tha t in w e c oul d havereplace d ${ \l f l o or cn^{1/2}\rfloor}$ b y an ything o f thef or m$ (c+o(1)) \, n^ {1/2 } $ with outa ff ecting t he val u eof $s(c)$ .Howeve r, we ha ve to writ e $\li minf$ as thef ollowing quest i on is open.
\ [p r : li m it\] Le t $c$ be afixe d pos itiv e r eal . Supp ose t ha t $ n $ tends to the infi ni ty and $k=( c+o(1))\, n^{ 1/2}$. Doe s t he ratio $s( k ,n ) /n$ tend to alimit ?
Our low e r boundon $s (c)$, pr ovided by t he follo win g l emm ata , wi ll be given b y diff er ent for mul ae fordif fer ent ra ng es of $c$ .
The b ou nd of Er dős a n d Freud[@ erd os +fr eud:9 1 ] impl ies t hat$$ \l a bel {eq:1}s( c ) =\fr ac {c ^2}2 ,\q ua d c\l e 2/ \ sqr t3.$$ T heir cons tru c tion c an be gen eralised to g iv e lower bo un dson $s( c ) $ for la rger $c$.
\[lm:SLower1 \ ] $$\la bel {eq:S L | choose 2}}_\ge \frac{t^2}2\left(_\frac{|A\cap I|^2}{k}+
\frac{|A\setminus _I|^2}{n-k}\right) +o(n^{3/2})=_t^2\left(\frac12_+
\frac{(\alpha n-k)^2}{k(n-k)}\right)_+o(n^{3/2}).$$
Thus,_up to an_error term of_$o(n^{3/2})$, we must have_equality throughout. We_conclude_that $\alpha=k/n+o(1)$ and $a_i=(\alpha/m+o(1))\,n^{1/2}$, which gives the required.
Lower Bounds on $s(k,n)$ {#SLower}
========================
We know that_the_range of_interest_is_$k=\Theta(n^{1/2})$. We will be proving_lower bounds on the following_‘scaled’ one-parameter_version of $s(k,n)$: $$\label{eq:DefS}
s(c):=\liminf_{n\to\infty} \frac{s({\lfloor cn^{1/2}\rfloor},n)}n.$$
Note that_in _we could have_replaced ${\lfloor cn^{1/2}\rfloor}$ by anything of the form $(c+o(1))\,_n^{1/2}$ without affecting the value of_$s(c)$. However, we_have_to_write $\liminf$ as the_following question is open.
\[pr:limit\] Let $c$_be a fixed positive real. Suppose_that $n$ tends to the infinity and_$k=(c+o(1))\, n^{1/2}$. Does the ratio $s(k,n)/n$_tend to a limit?
Our lower_bound on_$s(c)$, provided by the following_lemmata, will be_given by_different formulae for_different ranges of $c$.
The bound of_Erdős and Freud [@erdos+freud:91]_implies that $$\label{eq:1}
s(c)=\frac{c^2}2,\quad c\le 2/\sqrt3.$$_Their_construction can be_generalised_to_give lower_bounds on $s(c)$_for_larger $c$.
\[lm:SLower1\]_$$\label{eq:SL |
to-noise ratio reported with these refined positions to determine whether the burst could be a questionable detection. For a few dozen bursts, the signal-to-noise ratio associated with the refined positions found by this auto pipeline is lower than 7 (the typical image-trigger threshold). However, most of these cases are due to a long quiescent period of the burst emission before the spacecraft slews. We thus rerun the search for detections for these bursts using images created with the time interval and energy range determined by the flight software. If the signal-to-noise ratio is still below 7 and there is not an accompanying an XRT afterglows, we will mark it as a “tentative detection” in Table \[tab:summary\_general\]. The readers are advised to treat these bursts with special caution.
For spectral analysis, we use the commonly-adopted X-ray fitting package, XSPEC[^14]. When the spectrum covers time periods including spacecraft slew time, we generate multiple response files in this period in order to create an “average” response file for the whole period. [*Swift*]{} slews at a rate of $\sim$ 1 deg per second [@Markwardt07] and hence the telescope motion can be safely ignored within 5 seconds [@Sakamoto11_calibrate]. Therefore, we create a response file for each five seconds during slew time, and a response file for each time segment when the spacecraft is settled. We create an average response file for the whole time period using the HEASARC tool [*addrmf*]{}, with weighting factors equal to the fraction of photon counts in the specific time periods of each response file.
Following the BAT2 catalog, we fit the GRB spectra with two different models: simple power law (PL) and cutoff power law (CPL). The simple power-law model is described by the following equation, \[eq:PL\] f(E) = K\^[PL]{}\_[50]{} ( )\^[\^[PL]{}]{}, where f(E) is the photon flux at energy E, $\alpha^{\rm PL}$ is the PL index, and ${K^{PL}_{50} }$ is the normalization factor at 50 keV, with units of photons $\rm cm^{-2} \ s^{-1} \ keV^{-1}$. The cutoff power-law model is expressed as, \[eq:CPL\] f(E) = K\^[CPL]{}\_[50]{} ( )\^[\^[CPL]{}]{} (), where $\alpha^{\ | to - noise ratio reported with these refined position to settle whether the burst could be a questionable detection. For a few dozen fusillade, the signal - to - noise ratio associate with the refined positions recover by this auto pipeline is low than 7 (the typical image - trigger threshold). However, most of these cases are due to a retentive quiescent period of the burst emission before the spacecraft slews. We therefore rerun the search for detections for these bursts practice images make with the time interval and energy range determined by the escape software. If the signal - to - noise ratio is still below 7 and there is not an accompanying an XRT afterglows, we will mark it as a “ tentative detection ” in Table \[tab: summary\_general\ ]. The readers are advised to treat these bursts with special caution.
For spectral analysis, we practice the normally - adopted X - ray adjustment package, XSPEC[^14 ]. When the spectrum covers time period including spacecraft slew time, we generate multiple reception files in this period in order to create an “ average ” response file for the hale period. [ * Swift * ] { } slews at a rate of $ \sim$ 1 deg per second [ @Markwardt07 ] and therefore the telescope motion can be safely neglect within 5 second [ @Sakamoto11_calibrate ]. Therefore, we create a response file for each five second during slew time, and a response file for each time section when the spacecraft is settled. We produce an average response file for the hale time period use the HEASARC tool [ * addrmf * ] { }, with weight factors equal to the fraction of photon counts in the specific time period of each response file.
Following the BAT2 catalog, we equip the GRB spectra with two unlike models: simple power police (PL) and cutoff power law (CPL). The simple power - jurisprudence model is described by the follow equation, \[eq: PL\ ] f(E) = K\^[PL]{}\_[50 ] { } () \^[\^[PL ] { } ] { }, where f(E) is the photon flux at energy E, $ \alpha^{\rm PL}$ is the PL index, and $ { K^{PL}_{50 } } $ is the normalization factor at 50 keV, with unit of photons $ \rm cm^{-2 } \ s^{-1 } \ keV^{-1}$. The cutoff power - law model is expressed as, \[eq: CPL\ ] f(E) = K\^[CPL]{}\_[50 ] { } () \^[\^[CPL ] { } ] { } (), where $ \alpha^{\ | to-nlise ratio reported with these refined kowitionv to dstermine whether the burst could be e quwstiobable detection. For a wew dozen bursts, rhe wugnal-to-noise ratio assocjwted xith the refinec positionv found by thiv xuco pipeline is lower than 7 (the typicwl imagr-tgigger threshojd). Hpreved, most of these cases are due to a long qliescent period og the burst emission beforf thf spacecraft slews. We thus rethn eye search fof detections for these bursts using images created wigh thz time intetrql wtd energy renge dvtermined by bne flicht sofyware. If the slgnal-vo-nouse ratio is still benow 7 and there is got an acwojpanying an XRT adtwrglofs, wa wiuo mxrk iv aa a “tejtavive detectjon” in Tablw \[tab:summary\_general\]. Trv readers are advisqd to treat these bursts with special cauuion.
Fkr spectral analysis, we use the commonly-adopjed X-ray fytting package, XSPEC[^14]. When the spectrum covers tima permoas nkgludkbg spacecraft slew time, we generate multiple reflomsv files in this pcriod in order to vrfayg an “average” rgsponse fime for the whole pfriod. [*Srift*]{} wlews at w raye of $\sim$ 1 deg per second [@Mqrkwardt07] and yence the telescopz motion can be xafelu ignored within 5 seconbs [@Sakzmoto11_calibrwte]. Thererure, we create a fesionsa file for each five seconqs during slec time, avd a respogse file flr eagv time segment whej the s[acecraft ls settled. We create an average response file fmr nhe whole time period using the HEASARC tpol [*addxmf*]{}, wigh weightihg factmrs equal tj the fractiot of photon cmunts in the spexific tkoe periods of rach response file.
Fillowing the BAT2 cetaluf, we fit the GRY wpectra with twp dkffqrvnt mjgels: simple [owef lxe (PL) xnd cutoff iowdr lsw (CPL). The simple pofer-lzw model is descrineq by the followigg equation, \[ea:PL\] f(E) = K\^[PL]{}\_[50]{} ( )\^[\^[PL]{}]{}, whtre f(E) is the lhojon flux at energy E, $\alpha^{\rm PL}$ is the PP ikdex, and ${K^{PL}_{50} }$ is bhe gormalizatnon factor at 50 keV, with units of photons $\rm cm^{-2} \ s^{-1} \ keV^{-1}$. Thg cutoff power-law modgl ls expressed as, \[ez:CPL\] f(E) = N\^[CPL]{}\_[50]{} ( )\^[\^[CPL]{}]{} (), where $\alpya^{\ | to-noise ratio reported with these refined positions whether burst could a questionable detection. the ratio associated with refined positions found this auto pipeline is lower than (the typical image-trigger threshold). However, most of these cases are due to a quiescent period of the burst emission before the spacecraft slews. We thus rerun search detections these using images created with the time interval and energy range determined by the flight software. If signal-to-noise ratio is still below 7 and there not an accompanying an afterglows, we will mark it a detection” in \[tab:summary\_general\]. readers advised to treat bursts with special caution. For spectral analysis, we use the commonly-adopted X-ray fitting package, XSPEC[^14]. When the covers time spacecraft slew we multiple files in this order to create an “average” response whole period. [*Swift*]{} slews at a rate of 1 deg second [@Markwardt07] and hence the telescope can be safely ignored within 5 seconds [@Sakamoto11_calibrate]. we create a response file for each five seconds during slew time, and a response each time segment when spacecraft is settled. create average file the whole period using the HEASARC tool [*addrmf*]{}, with weighting factors equal to fraction of photon counts in the specific time periods of file. the BAT2 catalog, fit the GRB spectra two models: simple power law cutoff law power-law is by the following equation, f(E) = K\^[PL]{}\_[50]{} ( )\^[\^[PL]{}]{}, f(E) is the photon PL}$ is the PL index, and ${K^{PL}_{50} }$ the normalization factor at 50 keV, with of photons $\rm cm^{-2} \ s^{-1} \ keV^{-1}$. The cutoff power-law model expressed as, = K\^[CPL]{}\_[50]{} ( )\^[\^[CPL]{}]{} (), where $\alpha^{\ | to-noise ratio reported with tHese refineD posiTioNs tO dEterMine Whether the bursT CoulD be a questionable detectIon. FoR a FEw doZEn BurstS, the sigNAl-TO-NoiSe RaTio AsSOcIated WitH the refIned positiOns FoUnd by this autO PiPeline is loWer Than 7 (the typicAl iMage-trIgGer THreshOld). howevEr, most OF these Cases are dUe TO a long QUiescenT PErIod oF the burst emission BEfORe the spacecrafT slews. we THuS REruN thE search for DeTectiONs for thESe BURSts USing images creAted with the TIme IntervAl And ENergy rAnge dEtERmiNed by the fliGht sOftware. If The sigNAl-to-noiSE ratio iS still BelOw 7 aNd thERe Is Not An ACcoMPaNyiNG an xRT afterGlOwS, we wiLl maRK IT As a “tEntAtivE deteCtion” in Table \[tAb:sUmmaRY\_geNeral\]. the reAderS aRe advIsed to Treat ThEse bursts with spEciaL caution.
FOr sPeCtrAl AnalySIs, we usE thE coMmonly-aDopted X-RAy fItTING pAckage, XSPEC[^14]. When the SpECTrUm covers Time peRIoDs INcluding SpAceCrafT SLew tiMe, we GEnErate mulTiple rESpOnSe files In This peRiOd iN orDer to CReatE an “aveRage” respOnse fILe for the whole pERiod. [*Swift*]{} slewS At A RAtE Of $\siM$ 1 deG per second [@MArkwARdt07] aNd heNCe The TElescOpe moTiON cAN be safely ignored witHiN 5 seconDs [@SakAmoto11_calibratE]. Therefore, WE CReate a reSponSE fILe for each five sEcondS during sleW Time, and a RespoNse file fOr each timE SEgment whEn tHe sPacEcrAFT iS settled. We creATE an aVeRage resPonSe file fOr tHe wHolE tiMe Period usiNg the HEAsArC ToOl [*AddRmf*]{}, wiTH weightiNg FacToRs eQual tO The fraCtion Of phOtOn COunTs in the SPeCIFic tImE pErioDs oF eAch reSponSE fiLe.
FolloWing the BAt2 caTAlog, We FiT the GRB Spectra with twO dIfferent moDeLs: sImple pOWEr law (PL) aNd cutoff power law (CPL). The sIMple powEr-lAw modEl is Described By tHe follOwiNG equatIon, \[eq:Pl\] f(E) = K\^[Pl]{}\_[50]{} ( )\^[\^[Pl]{}]{}, whERE f(E) is THE pHotOn Flux at enerGY e, $\alPha^{\rm pL}$ Is thE PL indeX, and ${K^{PL}_{50} }$ is the normalIZatIon factor at 50 kev, wiTh unITS oF phOToNS $\rm Cm^{-2} \ S^{-1} \ KeV^{-1}$. tHE cutoff power-law Model is expReSSeD as, \[eq:CPL\] f(E) = k\^[cPL]{}\_[50]{} ( )\^[\^[cPl]{}]{} (), where $\aLpha^{\ | to-noise ratio reported wi th these r efine d p osi ti onsto d etermine wheth e r th e burst could be a que stion ab l e de t ec tion. For af ew d oze nbu rst s, th e sig nal -to-noi se ratio a sso ci ated with th e r efined pos iti ons found by th is aut opip e lineislower than7 (thetypical i ma g e-trig g er thre s h ol d).However, most oft he s e cases are du e to a l o ng q uie sce nt periodof theb urst em i ss i o n be f ore the space craft slews . We thusre run the se archfo r de tections fo r th ese burst s usin g images created withthe ti me i n te rv alan d en e rg y r a nge determi ne dby th e fl i g h t sof twa re.If th e signal-to-n ois e ra t iois st ill b elow 7 andthereis no tan accompanying anXRT after glo ws , w ewillm ark it as a“tentat ive det e cti on ” i nTable \[tab:summar y\ _ g en eral\].The re a de rs are advi se d t o tr e a t the se b u rs ts withspecia l c au tion.
Fo r spec tr alana lysis , weuse th e common ly-ad o pted X-ray fit t ing package,X SP E C [^ 1 4].Whe n the spect rumc over s ti m eper i ods i nclud in g s p acecraft slew time, w e gene ratemultiple resp onse files i n this pe riod in order to creat e an“average”r esponsefilefor thewhole per i o d. [*Swi ft* ]{} sl ews a ta rate of $\s i m $ 1de g per s eco nd [@Ma rkw ard t07 ] a nd hence th e telesc op emo ti oncan b e safelyig nor ed wi thin5 secon ds [@ Saka mo to 1 1_c alibrat e ]. T here fo re , we cr ea te aresp o nse file f or each f ive seco nd sduringslew time, an da response f ile for e a c h time s egment when the spacecr a ft is s ett led.We c reate anave rage r esp o nse fi le for thewh ole t ime p e r io d u si ng the HEA S A RCtool[* addr mf*]{}, with weighting fa c tor s equal to th e f ract i o nofp ho t onco u nts i n the specifictime perio ds of each resp o nse f ile.
F ollowin g the BAT2 ca talog, we fit theGR B sp e c tra with twodifferen t models: simpl e p owerlaw (PL)an d c utoff power law (CPL ). The s implepower -l aw model is described by the fo llowin g equ ati on, \[eq: PL\ ] f( E) = K\^[ PL]{ }\_[50]{}( ) \^[ \^[PL ]{} ] {}, w here f( E)i s the pho t on flux a t e ner g y E , $\alpha^{ \ r m PL }$ is th e PL in dex, and ${K^{PL}_{50 } }$ is the nor mali z a tio n f a ctor a t 50 keV, with un it s of photo ns $\rm cm^{- 2} \ s^{ -1 } \ ke V^{-1} $. The cutoff p ow e r-lawmode l i s express edas , \[eq:C PL \] f(E) = K\^ [C PL]{}\ _[50]{ } ( ) \ ^ [\^[CPL]{}]{} () , whe r e $\al p ha^ {\ | to-noise ratio_reported with_these refined positions to_determine whether_the_burst could_be_a questionable detection._For a few_dozen bursts, the signal-to-noise_ratio associated with_the_refined positions found by this auto pipeline is lower than 7 (the typical image-trigger_threshold)._However, most_of_these_cases are due to a_long quiescent period of the_burst emission_before the spacecraft slews. We thus rerun the_search_for detections for_these bursts using images created with the time interval_and energy range determined by the_flight software. If_the_signal-to-noise_ratio is still below_7 and there is not an_accompanying an XRT afterglows, we will_mark it as a “tentative detection” in_Table \[tab:summary\_general\]. The readers are advised_to treat these bursts with_special caution.
For_spectral analysis, we use the_commonly-adopted X-ray fitting_package, XSPEC[^14]._When the spectrum_covers time periods including spacecraft slew_time, we generate_multiple response files in this period_in_order to create_an_“average”_response file_for the whole_period._[*Swift*]{} slews_at_a rate of $\sim$ 1 deg_per_second [@Markwardt07] and hence the telescope motion_can be safely ignored_within_5 seconds [@Sakamoto11_calibrate]. Therefore,_we create a response file_for each five seconds during slew_time, and_a response_file for each time segment when the spacecraft is settled. We_create an average response file for_the whole time period_using the_HEASARC_tool [*addrmf*]{}, with_weighting_factors equal_to the fraction of photon counts in_the specific_time periods of each response file.
Following_the BAT2 catalog, we_fit_the GRB spectra with two different_models: simple power law (PL) and_cutoff power law (CPL). The_simple_power-law_model is described by the_following equation, \[eq:PL\] f(E) = K\^[PL]{}\_[50]{}_( )\^[\^[PL]{}]{}, where_f(E) is the photon flux at energy_E,_$\alpha^{\rm PL}$ is the PL index,_and_${K^{PL}_{50} }$ is the normalization factor_at_50_keV, with units of photons_$\rm cm^{-2} \ s^{-1} \ keV^{-1}$._The cutoff power-law model is expressed as, \[eq:CPL\] f(E)_= K\^[CPL]{}\_[50]{} (_)\^[\^[CPL]{}]{} (), where $\alpha^{\ |
dose not require construction of wave operators $W_{\pm}$ to reconstruct the potentials. Thus the method of the $v$-expansion is simpler and less complicated than the method of the ${\varepsilon}$-expansion.
In order to prove that the operator defined as is a scattering operator to the equation, we require a time-decay $L^{\infty}$-estimate on the solution to the Cauchy problem of the equation. Such estimate on the solution to the Hartree type equation has been extensively investigated for the case where the potential has the form $\lambda |x|^{-\gamma}$ for some constants $\lambda$ and $\gamma$ (see, e.g., Wada [@Wada]; Hayashi and Naumkin [@Hayashi-Naumkin98]; Hayashi and Ozawa [@Hayashi-Ozawa1988]). Few researchers have addressed the problem of the time-decay $L^{\infty}$-estimate on the solution to the Hartree-Fock equation with general interaction and external potentials. Making use of the pseudo-conformal conservation law and the Gagliardo-Nirenberg inequality, with the help of assumptions on the potentials, we give a $L^{\infty}$-estimate on the solution to the equation.
This paper is organized as follows: Section 2 proves that the operator $S$ defined as is the scattering operator for the Hartree-Fock equation, after proving a time-decay $L^{\infty}$-estimate on the solution to the Cauchy problem for the equation. We give the asymptotic expansion of the scattering operator with the high-velocity initial states in Section 3. Section 4 is devoted to reconstructions of the interaction potential and the external potential.
Representation of the scattering operator {#sec:2}
=========================================
This section shows that the scattering operator for the Hartree-Fock equation has a representation. We first recall that the unique existence of the scattering solution of Hartree equations with the external potential are studied in [@Watanabe0]. This result can be easily applicable to the Hartree-Fock equation. Assume that potentials $V_{ext}$ and $V_{int}$ satisfy Assumption \[ass:interaction-2\] and Assumption \[ass:external\], respectively. Then there exists $\varepsilon_0 >0$ such that the equation satisfying the condition $u_j(t)\to e^{-itH_0}\varphi_j$ as $t\to -\infty$ in $L^2({\ | dose not require construction of wave operators $ W_{\pm}$ to reconstruct the potential. therefore the method of the $ v$-expansion is simpler and less complicated than the method acting of the $ { \varepsilon}$-expansion.
In order to raise that the operator defined as is a break up operator to the equation, we necessitate a time - decay $ L^{\infty}$-estimate on the solution to the Cauchy problem of the equality. Such estimate on the solution to the Hartree type equation has been extensively investigated for the shell where the potential has the form $ \lambda |x|^{-\gamma}$ for some constants $ \lambda$ and $ \gamma$ (see, for example, Wada [ @Wada ]; Hayashi and Naumkin [ @Hayashi - Naumkin98 ]; Hayashi and Ozawa [ @Hayashi - Ozawa1988 ]). Few researchers have addressed the problem of the fourth dimension - decay $ L^{\infty}$-estimate on the solution to the Hartree - Fock equation with general interaction and external potential. Making function of the pseudo - conformal conservation law and the Gagliardo - Nirenberg inequality, with the help of assumptions on the potentials, we give a $ L^{\infty}$-estimate on the solution to the equality.
This paper is organized as follows: Section 2 proves that the operator $ S$ define as is the break up operator for the Hartree - Fock equation, after proving a time - decay $ L^{\infty}$-estimate on the solution to the Cauchy problem for the equation. We afford the asymptotic expansion of the disperse operator with the gamey - speed initial department of state in Section 3. Section 4 is devote to reconstructions of the interaction potential and the external potential.
Representation of the scattering hustler { # sec:2 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
This section shows that the scattering operator for the Hartree - Fock equation has a representation. We first recall that the unique existence of the scattering solution of Hartree equation with the external potential are study in [ @Watanabe0 ]. This result can be easily applicable to the Hartree - Fock equation. Assume that potentials $ V_{ext}$ and $ V_{int}$ satisfy Assumption \[ass: interaction-2\ ] and Assumption \[ass: external\ ], respectively. Then there exists $ \varepsilon_0 > 0 $ such that the equality satisfying the condition $ u_j(t)\to e^{-itH_0}\varphi_j$ as $ t\to -\infty$ in $ L^2({\ | dode not require constructlon of wave opercrors $W_{\'m}$ to rsconstruzt the potentials. Thus the mxthoe of uke $v$-expansion is simoler and pess compliceted than the mevgod of bke ${\vadcpsilmi}$-expansion.
In orcer to proee that the oparxtlr defined as is a scattering operaeor to yhf equation, we tequige a tjme-decay $L^{\infty}$-estimate on the solhtion tm the Cauchy lroblem of the equation. Sufh edtimate on the solktion to thg Hawrree type eqjation has been extensjvely investigated for the case wherz the potenjnql jds the form $\lambqa |x|^{-\gamma}$ fov some wonstanys $\lambda$ and $\namma$ (sew, e.g., Wada [@Wada]; Hayashm and Naumkin [@Hayashy-Naumkin98]; Vavashi and Ozawa [@Hayasyi-Izawa1988]). Few resdqrcfera iavs addrfssxd the probmem of the rime-decay $L^{\infty}$-estomwnr on the solhtion eo the Hartree-Fock equation with general pntedaction and external porentials. Making use ov the psetdo-conformal conservation law and the Gagliardo-Nisenbecg intqmwuutj, with the help of assumptions on the potentiwms, wv give a $L^{\infty}$-esbimate on the soluyiln jo the equatiov.
This 'zpsr is organized as followf: Secrion 2 prodes yhat the operator $S$ defined as is the sbattwring operator for the Hartrez-Fock gquatipn, after proving a time-becay $M^{\infty}$-estimwte on ths solution to the Calchy problem for the equation. Re give tie asvmptotic expsnsion of the scwtterltg operator with tje hiyh-velmcity initlal states in Section 3. Section 4 is devoted to racotstructijns on the interactijn potential aud the erternau potentiam.
Represxntation of ehe scatterinc operator {#ser:2}
=========================================
This seceion shoqs that ghe scattering operator for the Hqrtree-Fock equatiok has z representatiou. Qe first recall thxt ehv uiique axistence of the scxyterivg soluunon uf Hsrtree equations witv ths external potentisl are stueied in [@Ratanabe0]. This result can be easlly a'plicaule to thg Hartree-Fock equation. Assume tgat potenhiajs $V_{ext}$ and $D_{int}$ satisfy Assbmption \[ass:interaction-2\] and Assumption \[asw:external\], respecticely. Then there exixus $\varepsilon_0 >0$ sucr that tha equation satisfyint the condition $u_m(t)\to e^{-itH_0}\varphi_j$ as $t\tk -\inftf$ in $P^2({\ | dose not require construction of wave operators reconstruct potentials. Thus method of the complicated the method of ${\varepsilon}$-expansion. In order prove that the operator defined as a scattering operator to the equation, we require a time-decay $L^{\infty}$-estimate on the to the Cauchy problem of the equation. Such estimate on the solution to Hartree equation been investigated for the case where the potential has the form $\lambda |x|^{-\gamma}$ for some constants $\lambda$ $\gamma$ (see, e.g., Wada [@Wada]; Hayashi and Naumkin Hayashi and Ozawa [@Hayashi-Ozawa1988]). researchers have addressed the problem the $L^{\infty}$-estimate on solution the equation with general and external potentials. Making use of the pseudo-conformal conservation law and the Gagliardo-Nirenberg inequality, with the help assumptions on we give $L^{\infty}$-estimate the to the equation. is organized as follows: Section 2 operator $S$ defined as is the scattering operator the Hartree-Fock after proving a time-decay $L^{\infty}$-estimate on solution to the Cauchy problem for the equation. give the asymptotic expansion of the scattering operator with the high-velocity initial states in Section 4 is devoted to of the interaction and external Representation the scattering {#sec:2} ========================================= This section shows that the scattering operator for the equation has a representation. We first recall that the unique the solution of Hartree with the external potential studied [@Watanabe0]. This result can applicable the that $V_{ext}$ $V_{int}$ satisfy Assumption \[ass:interaction-2\] Assumption \[ass:external\], respectively. Then there $\varepsilon_0 >0$ such that $u_j(t)\to e^{-itH_0}\varphi_j$ as $t\to -\infty$ in $L^2({\ | dose not require constructioN of wave opeRatorS $W_{\pM}$ to ReConsTrucT the potentials. tHus tHe method of the $v$-expansioN is siMpLEr anD LeSs comPlicateD ThAN The MeThOd oF tHE ${\vArepsIloN}$-expansIon.
In order To pRoVe that the opeRAtOr defined aS is A scattering oPerAtor to ThE eqUAtion, We rEquirE a time-DEcay $L^{\iNfty}$-estimAtE On the sOLution tO THe caucHy problem of the equATiON. Such estimate oN the soLuTIoN TO thE HaRtree type eQuAtion HAs been eXTeNSIVelY Investigated fOr the case whERe tHe poteNtIal HAs the fOrm $\laMbDA |x|^{-\gAmma}$ for some ConsTants $\lambDa$ and $\gAMma$ (see, e.G., wada [@WadA]; HayasHi aNd NAumkIN [@HAyAshI-NAUmkIN98]; HAyaSHi aNd Ozawa [@HAyAsHi-OzaWa1988]). FeW RESEarcHerS havE addrEssed the problEm oF the TIme-Decay $l^{\inftY}$-estImAte on The solUtion To The Hartree-Fock eQuatIon with geNerAl IntErActioN And extErnAl pOtentiaLs. MakinG Use Of THE PsEudo-conformal conseRvATIoN law and tHe GaglIArDo-nIrenberg InEquAlitY, WIth thE helP Of AssumptiOns on tHE pOtEntials, We Give a $L^{\InFty}$-EstImate ON the SolutiOn to the eQuatiON.
This paper is orGAnized as folloWS: SECTiON 2 proVes That the operAtor $s$ DefiNed aS Is The SCatteRing oPeRAtOR for the Hartree-Fock eQuAtion, aFter pRoving a time-deCay $L^{\infty}$-eSTIMate on thE solUTiON to the Cauchy prOblem For the equaTIon. We givE the aSymptotiC expansioN OF the scatTerIng OpeRatOR WiTh the high-veloCITy inItIal statEs iN SectioN 3. SeCtiOn 4 iS deVoTed to recoNstructiOnS oF tHe IntEractIOn potentIaL anD tHe eXternAL potenTial.
REpreSeNtATioN of the sCAtTERing OpErAtor {#Sec:2}
=========================================
thIs secTion SHowS that thE scatteriNg oPEratOr FoR the HarTree-Fock equatIoN has a repreSeNtaTion. We FIRst recalL that the unique existence OF the scaTteRing sOlutIon of HartRee EquatiOns WIth the ExternAl potEnTiaL ARe stuDIEd In [@WAtAnabe0]. This rESUlt Can be EaSily ApplicaBle to the Hartree-FocK EquAtion. Assume thAt pOtenTIAlS $V_{eXT}$ aND $V_{iNt}$ SAtiSFY Assumption \[ass:iNteraction-2\] AnD asSumption \[asS:ExtErNal\], respEctivelY. Then THere exiSts $\varepsIlon_0 >0$ such tHaT the EQUatIon satisfyIng the coNdition $u_j(T)\To e^{-ith_0}\VaRphi_j$ As $t\To -\inftY$ iN $L^2({\ | dose not require construc tion of wa ve op era tor s$W_{ \pm} $ to reconstru c t th e potentials. Thus the meth od of t h e$v$-e xpansio n i s sim pl er an dl es s com pli cated t han the me tho dof the ${\va r ep silon}$-ex pan sion.
In or der to pr ov e t h at th e o perat or def i ned as is a sca tt e ring o p eratort o t he e quation, we requi r ea time-decay $L ^{\inf ty } $- e s tim ate on the so lu tiont o the C a uc h y pro b lem of the eq uation. Suc h es timate o n t h e solu tionto the Hartree ty pe e quation h as bee n extens i vely in vestig ate d f or t h eca sewh e ret he po t ent ial hasth eform$\la m b d a |x| ^{- \gam ma}$for some cons tan ts $ \ lam bda$and $ \gam ma $ (se e, e.g ., Wa da [@Wada]; Hayas hi a nd Naumki n [ @H aya sh i-Nau m kin98] ; H aya shi and Ozawa[ @Ha ya s h i -O zawa1988]). Few re se a r ch ers have addre s se dt he probl em of the t ime-d ecay $L ^{\infty }$-est i ma te on the s olutio ntothe Hart r ee-F ock eq uation w ith g e neral interact i on and extern a lp o te n tial s.Making useof t h e ps eudo - co nfo r mal c onser va t io n law and the Gaglia rd o-Nire nberg inequality,with the h e l p of assu mpti o ns on the potenti als,we give a$ L^{\inft y}$-e stimateon the so l u tion tothe eq uat ion .
T his paper iso r gani ze d as fo llo ws: Sec tio n 2 pr ove sthat theoperator $ S$ d ef ine d asi s the sc at ter in g o perat o r forthe H artr ee -F o ckequatio n ,a f terpr ov inga t im e-dec ay $ L ^{\ infty}$ -estimate on theso lu tion to the Cauchy p ro blem for t he eq uation . We givethe asymptotic expansio n of the sc atter ingoperatorwit h thehig h -veloc ity in itial s tat e s in S e c ti on3. Section 4 i s d evote dto r econstr uctions of the int e rac tion potentia l a nd t h e e xte r na l po te n tia l .
Representatio n of the s ca t te ring opera t or{# sec:2}======= ===== = ======= ========= ========= == =
T h i s s ection sho ws thatthe scatt e ringo pe rator fo r theHa rtr ee-Fo ck equ a tio n has a rep re sentat ion.We first r ecall that the unique e xisten ce of th e scatter ing sol ution ofHart ree equati ons wi th th e e x terna l po t en tia l arestud i ed in [@W a ta nab e 0 ]. This resul t c anbe ea sil y appli cabl e to the Hartree- F ock equation.Assu m e th atp oten ti als $V_{ext}$and $ V _ {int}$ s at isfy Assump tion \[a ss : inter action -2\] a nd Assu m p ti o n \[as s:ex ter nal\], re spe ct i vely. T he nt here e xist s$\vare psilon _ 0 >0 $ such that the eq uatio n satis f yin g the c onditio n $u_ j(t)\to e^ {-itH_0}\va rphi_j $ as $t\t o -\inf ty $ in $ L^2 ({ \ | dose_not require_construction of wave operators_$W_{\pm}$ to_reconstruct_the potentials._Thus_the method of_the $v$-expansion is_simpler and less complicated_than the method_of_the ${\varepsilon}$-expansion.
In order to prove that the operator defined as is a scattering operator_to_the equation,_we_require_a time-decay $L^{\infty}$-estimate on the_solution to the Cauchy problem_of the_equation. Such estimate on the solution to the_Hartree_type equation has_been extensively investigated for the case where the potential_has the form $\lambda |x|^{-\gamma}$ for_some constants $\lambda$_and_$\gamma$_(see, e.g., Wada [@Wada];_Hayashi and Naumkin [@Hayashi-Naumkin98]; Hayashi and_Ozawa [@Hayashi-Ozawa1988]). Few researchers have addressed_the problem of the time-decay $L^{\infty}$-estimate on_the solution to the Hartree-Fock equation_with general interaction and external_potentials. Making_use of the pseudo-conformal conservation_law and the_Gagliardo-Nirenberg inequality,_with the help_of assumptions on the potentials, we_give a $L^{\infty}$-estimate_on the solution to the equation.
This_paper_is organized as_follows:_Section_2 proves_that the operator_$S$_defined as_is_the scattering operator for the Hartree-Fock_equation,_after proving a time-decay $L^{\infty}$-estimate on the_solution to the Cauchy_problem_for the equation. We_give the asymptotic expansion of_the scattering operator with the high-velocity_initial states_in Section_3. Section 4 is devoted to reconstructions of the interaction potential_and the external potential.
Representation of the_scattering operator {#sec:2}
=========================================
This section_shows that_the_scattering operator for_the_Hartree-Fock equation_has a representation. We first recall that_the unique_existence of the scattering solution of_Hartree equations with the_external_potential are studied in [@Watanabe0]. This_result can be easily applicable to_the Hartree-Fock equation. Assume that_potentials_$V_{ext}$_and $V_{int}$ satisfy Assumption \[ass:interaction-2\]_and Assumption \[ass:external\], respectively. Then there_exists $\varepsilon_0 >0$_such that the equation satisfying the condition_$u_j(t)\to_e^{-itH_0}\varphi_j$ as $t\to -\infty$ in $L^2({\ |
at each site of the spin chain[^2]. In that case we denote the spin chain Hilbert space as $$\label{Hilbert.space.inhom}
\mathcal{V}_m=\mathbb{C}^2_{m_1} \otimes \mathbb{C}^2_{m_2} \otimes \ldots \otimes \mathbb{C}^2_{m_N}\,,$$ and the Bethe equations turn into $$\label{eq:Bethe.eqn2}
\prod_{i=1}^N\frac{\sigma_a-m_i+\frac{\hbar}{2}}{\sigma_a-m_i-\frac{\hbar}{2}} = q\,\prod_{b\neq a}\frac{\sigma_a-\sigma_b+\hbar}{\sigma_a-\sigma_b-\hbar}\,, \qquad a=1,\ldots,k.$$ For a given number of excitations $k$, there are $\binom{N}{k}$ solutions distinct solutions of the Bethe equations. The solutions $\sigma_a^I$ can be labelled by a subset $I=\{I_1,\ldots,I_k\}\subset\{1,\ldots,N\}$ such that, expanding around $q \to 0$, the solutions are of the form $\sigma_a^I=m_{I_a}-\tfrac{\hbar}{2}+\cO(q)$.
R-matrices
----------
There are many independent ways to arrive at the Bethe equations. Usually, the most powerful method is the algebraic Bethe ansatz, which is based on the construction of an R-matrix. For the inhomogeneous spin chain, this is an operator acting on two sites, $$R_{ij}(m_j-m_i):\mathbb{C}^2_{m_i} \otimes \mathbb{C}^2_{m_j} \to \mathbb{C}^2_{m_j} \otimes \mathbb{C}^2_{m_i}\,.$$ It has rational dependence on $m_j-m_i$ and satisfies the regularity property $R_{ij}(0)\sim \mathbb{P}_{ij}$ where $\mathbb{P}_{ij}$ is the permutation operator, together with the Yang-Baxter equation (shown graphically in figure \[Fig:YangBaxter.equation\]) $$\label{Yang.Baxter}
R_{12}(m_2-m_1)R | at each site of the spin chain[^2 ]. In that case we denote the tailspin range Hilbert outer space as $ $ \label{Hilbert.space.inhom }
\mathcal{V}_m=\mathbb{C}^2_{m_1 } \otimes \mathbb{C}^2_{m_2 } \otimes \ldots \otimes \mathbb{C}^2_{m_N}\,,$$ and the Bethe equations turn into $ $ \label{eq: Bethe.eqn2 }
\prod_{i=1}^N\frac{\sigma_a - m_i+\frac{\hbar}{2}}{\sigma_a - m_i-\frac{\hbar}{2 } } = q\,\prod_{b\neq a}\frac{\sigma_a-\sigma_b+\hbar}{\sigma_a-\sigma_b-\hbar}\, , \qquad a=1,\ldots, k.$$ For a yield number of excitations $ k$, there be $ \binom{N}{k}$ solution distinct solution of the Bethe equations. The solutions $ \sigma_a^I$ can be label by a subset $ I=\{I_1,\ldots, I_k\}\subset\{1,\ldots, N\}$ such that, expanding around $ q \to 0 $, the solutions are of the class $ \sigma_a^I = m_{I_a}-\tfrac{\hbar}{2}+\cO(q)$.
radius - matrices
----------
There are many independent ways to arrive at the Bethe equations. normally, the most powerful method is the algebraic Bethe ansatz, which is base on the construction of an R - matrix. For the inhomogeneous tailspin chain, this is an hustler acting on two sites, $ $ R_{ij}(m_j - m_i):\mathbb{C}^2_{m_i } \otimes \mathbb{C}^2_{m_j } \to \mathbb{C}^2_{m_j } \otimes \mathbb{C}^2_{m_i}\,.$$ It has rational dependence on $ m_j - m_i$ and satisfies the regularity place $ R_{ij}(0)\sim \mathbb{P}_{ij}$ where $ \mathbb{P}_{ij}$ is the permutation operator, together with the Yang - Baxter equation (shown graphically in number \[Fig: YangBaxter.equation\ ]) $ $ \label{Yang. Baxter }
R_{12}(m_2 - m_1)R | at each site of the spin cmain[^2]. In that casg qe denmte ths spin cfain Hilbert space as $$\label{Hmlbeet.spaxe.inhom}
\mathcal{V}_m=\mathbb{Z}^2_{m_1} \otimes \mathbb{C}^2_{n_2} \otmmes \ldots \otimxa \mathbn{E}^2_{m_N}\,,$$ ahf thz Uethe equations turn into $$\label{eq:Bethe.exn2}
\orld_{i=1}^N\frac{\sigma_a-m_i+\frac{\hbar}{2}}{\sigma_a-m_i-\fras{\hbar}{2}} = a\,\pgod_{b\neq a}\frac{\sygma_s-\figmz_b+\hbar}{\sigma_a-\sigma_b-\hbar}\,, \qquad a=1,\ldota,k.$$ For e given number pf excitations $k$, there are $\binlm{N}{k}$ solutions dishinct solutuons if the Bethe equations. The solutiohs $\sigma_a^I$ can be labelled by a subszt $I=\{I_1,\ldots,I_j\}\sybsfj\{1,\ldots,N\}$ such that, expanding avpund $q \to 0$, thr solutions arc of vhe dorm $\sigma_a^I=m_{I_a}-\tfrac{\huar}{2}+\cO(q)$.
R-matrices
----------
There are many iudependent ways to areice at the Betfw eduauiois. Hsuallj, tie most powsrful methoe is the algebraic Neeye ansatz, whidh is fafed on the construction of an R-matrix. Fmr fhe inhomogeneous spin xhain, this is an operwtor actigg on two sites, $$R_{ij}(m_j-m_i):\mathbb{C}^2_{m_i} \otimes \mathbb{C}^2_{m_j} \to \metfbb{E}^2_{n_j} \otkned \mathbb{C}^2_{m_i}\,.$$ It has rational dependence on $m_j-m_y$ amd satisfies the regularity pgolgrty $R_{ij}(0)\sim \majhbb{P}_{ij}$ whsre $\mathbb{P}_{ij}$ is tje permotatiob operatow, tobether with the Yang-Baxter wquation (shocn traphically in figbre \[Fig:YangBarter.eqoation\]) $$\label{Yang.Baxter}
R_{12}(m_2-m_1)R | at each site of the spin chain[^2]. case denote the chain Hilbert space \otimes \otimes \mathbb{C}^2_{m_N}\,,$$ and Bethe equations turn $$\label{eq:Bethe.eqn2} \prod_{i=1}^N\frac{\sigma_a-m_i+\frac{\hbar}{2}}{\sigma_a-m_i-\frac{\hbar}{2}} = q\,\prod_{b\neq a}\frac{\sigma_a-\sigma_b+\hbar}{\sigma_a-\sigma_b-\hbar}\,, \qquad For a given number of excitations $k$, there are $\binom{N}{k}$ solutions distinct solutions the Bethe equations. The solutions $\sigma_a^I$ can be labelled by a subset $I=\{I_1,\ldots,I_k\}\subset\{1,\ldots,N\}$ that, around \to the solutions are of the form $\sigma_a^I=m_{I_a}-\tfrac{\hbar}{2}+\cO(q)$. R-matrices ---------- There are many independent ways to arrive the Bethe equations. Usually, the most powerful method the algebraic Bethe ansatz, is based on the construction an For the spin this an operator acting two sites, $$R_{ij}(m_j-m_i):\mathbb{C}^2_{m_i} \otimes \mathbb{C}^2_{m_j} \to \mathbb{C}^2_{m_j} \otimes \mathbb{C}^2_{m_i}\,.$$ It has rational dependence on $m_j-m_i$ and satisfies regularity property where $\mathbb{P}_{ij}$ the operator, with the Yang-Baxter graphically in figure \[Fig:YangBaxter.equation\]) $$\label{Yang.Baxter} R_{12}(m_2-m_1)R | at each site of the spin chain[^2]. IN that case wE denoTe tHe sPiN chaIn HiLbert space as $$\laBEl{HiLbert.space.inhom}
\mathcal{v}_m=\matHbB{c}^2_{m_1} \otIMeS \mathBb{C}^2_{m_2} \otiMEs \LDOts \OtImEs \mAtHBb{c}^2_{m_N}\,,$$ anD thE Bethe eQuations tuRn iNtO $$\label{eq:BethE.EqN2}
\prod_{i=1}^N\fraC{\siGma_a-m_i+\frac{\hbAr}{2}}{\sIgma_a-m_I-\fRac{\HBar}{2}} = q\,\pRod_{B\neq a}\Frac{\siGMa_a-\sigMa_b+\hbar}{\siGmA_A-\sigma_B-\Hbar}\,, \qquAD A=1,\lDots,K.$$ For a given number oF ExCItations $k$, there Are $\binOm{n}{K}$ sOLUtiOns Distinct soLuTions OF the BetHE eQUATioNS. The solutions $\Sigma_a^I$ can bE LabElled bY a SubSEt $I=\{I_1,\ldOts,I_k\}\SuBSet\{1,\Ldots,N\}$ such tHat, eXpanding aRound $q \TO 0$, the solUTions arE of the ForM $\siGma_a^i=M_{I_A}-\tFraC{\hBAr}{2}+\co(Q)$.
R-MatRIceS
----------
There arE mAnY indePendENT WAys tO arRive At the bethe equationS. UsUallY, The Most pOwerfUl meThOd is tHe algeBraic beThe ansatz, which iS basEd on the coNstRuCtiOn Of an R-MAtrix. FOr tHe iNhomogeNeous spIN chAiN, THIs Is an operator acting On TWO sItes, $$R_{ij}(m_J-m_i):\matHBb{c}^2_{m_I} \Otimes \maThBb{C}^2_{M_j} \to \MAThbb{C}^2_{M_j} \otIMeS \mathbb{C}^2_{M_i}\,.$$ It haS RaTiOnal depEnDence oN $m_J-m_i$ And SatisFIes tHe reguLarity prOpertY $r_{ij}(0)\sim \mathbb{P}_{iJ}$ Where $\mathbb{P}_{iJ}$ Is THE pERmutAtiOn operator, tOgetHEr wiTh thE yaNg-BAXter eQuatiOn (SHoWN graphically in figurE \[FIg:YangbaxteR.equation\]) $$\labeL{Yang.BaxteR}
r_{12}(M_2-M_1)R | at each site of the spinchain[^2]. In t hat ca se wedeno te the spin ch a in H ilbert space as $$\lab el{Hi lb e rt.s p ac e.inh om}
\ma t hc a l {V} _m =\ mat hb b {C }^2_{ m_1 } \otim es \mathbb {C} ^2 _{m_2} \otim e s \ldots \o tim es \mathbb{C }^2 _{m_N} \, ,$$ and t heBethe equat i ons tu rn into $ $\ l abel{e q :Bethe. e q n2 }
\p rod_{i=1}^N\frac{ \ si g ma_a-m_i+\frac {\hbar }{ 2 }} { \ sig ma_ a-m_i-\fra c{ \hbar } {2}} =q \, \ p r od_ { b\neq a}\frac {\sigma_a-\ s igm a_b+\h ba r}{ \ sigma_ a-\si gm a _b- \hbar}\,, \ qqua d a=1,\ld ots,k. $ $ For a given n umberofexc itat i on s$k$ ,t her e a re$ \bi nom{N}{k }$ s oluti onsd i s t inct so luti ons o f the Bethe e qua tion s . T he so lutio ns $ \s igma_ a^I$ c an be l abelled by a su bset $I=\{I_1 ,\l do ts, I_ k\}\s u bset\{ 1,\ ldo ts,N\}$ such t h at, e x p a nd ing around $q \to0$ , th e soluti ons ar e o ft he form$\ sig ma_a ^ I =m_{I _a}- \ tf rac{\hba r}{2}+ \ cO (q )$.
R- ma trices
- --- --- ---
T here are m any inde pende n t ways to arri v e at the Beth e e q u at i ons. Us ually, themost powe rful me tho d is t he al ge b ra i c Bethe ansatz, whi ch is ba sed o n the constru ction of a n R -matrix. For th e inhomogeneous spin chain, th i s is anopera tor acti ng on two s ites, $$ R_{ ij} (m_ j-m _ i ): \mathbb{C}^2_ { m _i}\o times \ mat hbb{C}^ 2_{ m_j } \ to\m athbb{C}^ 2_{m_j}\o ti me s\ma thbb{ C }^2_{m_i }\ ,.$ $Ithas r a tional depe nden ce o n $m _j-m_i$ an d sati sf ie s th e r eg ulari ty p r ope rty $R_ {ij}(0)\s im\ math bb {P }_{ij}$ where $\math bb {P}_{ij}$is th e perm u t ation op erator, together with t h e Yang- Bax ter e quat ion (show n g raphic all y in fi gure \ [Fig: Ya ngB a x ter.e q u at ion \] ) $$\label { Y ang .Baxt er }
R_ {12}(m_ 2-m_1)R | at_each site_of the spin chain[^2]._In that_case_we denote_the_spin chain Hilbert_space as $$\label{Hilbert.space.inhom}
\mathcal{V}_m=\mathbb{C}^2_{m_1}_\otimes \mathbb{C}^2_{m_2} \otimes _\ldots \otimes \mathbb{C}^2_{m_N}\,,$$_and_the Bethe equations turn into $$\label{eq:Bethe.eqn2}
\prod_{i=1}^N\frac{\sigma_a-m_i+\frac{\hbar}{2}}{\sigma_a-m_i-\frac{\hbar}{2}} = q\,\prod_{b\neq a}\frac{\sigma_a-\sigma_b+\hbar}{\sigma_a-\sigma_b-\hbar}\,, \qquad a=1,\ldots,k.$$ For a given_number_of excitations_$k$,_there_are $\binom{N}{k}$ solutions distinct solutions_of the Bethe equations. The_solutions $\sigma_a^I$_can be labelled by a subset $I=\{I_1,\ldots,I_k\}\subset\{1,\ldots,N\}$ such_that,_expanding around $q_\to 0$, the solutions are of the form $\sigma_a^I=m_{I_a}-\tfrac{\hbar}{2}+\cO(q)$.
R-matrices
----------
There_are many independent ways to arrive_at the Bethe_equations._Usually,_the most powerful method_is the algebraic Bethe ansatz, which_is based on the construction of_an R-matrix. For the inhomogeneous spin chain,_this is an operator acting on_two sites, $$R_{ij}(m_j-m_i):\mathbb{C}^2_{m_i} \otimes \mathbb{C}^2_{m_j}_\to \mathbb{C}^2_{m_j}_\otimes \mathbb{C}^2_{m_i}\,.$$ It has rational_dependence on $m_j-m_i$_and satisfies_the regularity property_$R_{ij}(0)\sim \mathbb{P}_{ij}$ where $\mathbb{P}_{ij}$ is the_permutation operator, together_with the Yang-Baxter equation (shown graphically_in_figure \[Fig:YangBaxter.equation\]) $$\label{Yang.Baxter}
R_{12}(m_2-m_1)R |
beta}|}} C\end{pmatrix}},\end{aligned}$$ $$\begin{aligned}
{\alpha}{\begin{pmatrix}
a&b&U_1&0&w\\
c&d&U_2&-w&0\\
X_1&X_2&D&\I_{p,q} U_1^t&\I_{p,q} U_2^t\\
0&z&X_1^t\I_{p,q}&-a&-c\\
-z&0&X_2^t\I_{p,q}&-b&-d\end{pmatrix}}
&:=& {\begin{pmatrix}
\frac{a+d}2&-w&\frac12U&\frac12V\\
z&-\frac{a+d}2&-\frac12Y^t\I_{p,q}&\frac12X^t\I_{p,q}\\
X_1&-\I_{p,q} U_2^t&D+\frac{d-a}2\I_n&-c\I_n\\
X_2&\I_{p,q} U_1^t&-b\I_n&D+\frac{a-d}2\I_n\end{pmatrix}}.\end{aligned}$$
\[lem1\] The map $i:Q\to\tilde P$ is an injective Lie group homomorphism, ${\alpha}:\g\to\tg$ is linear, and the pair $(i,{\alpha})$ satisfies the conditions (1)–(3) from \[3.2\].
Hence the pair $(i,{\alpha})$ gives rise to an extension functor from Cartan geometries of type $(G,Q)$ to Cartan geometries of type $(\tilde G,\tilde P)$.
The map $i$ is obviously well defined, i.e. the image of an element of $Q\subset PO(p+2,q+2)$ does not depend on the representative matrix. Further, $i$ is smooth and injective and a direct computation shows this is a homomorphism of Lie groups. The map ${\alpha}$ is linear and the compatibility of the pair $(i,{\alpha})$ follows as follows: (1) involves a little tedious but very straightforward checking, (2) is an easy exercise, (3) follows from $\g/\q\cong\ | beta}| } } C\end{pmatrix}},\end{aligned}$$ $ $ \begin{aligned }
{ \alpha}{\begin{pmatrix }
a&b&U_1&0&w\\
c&d&U_2&-w&0\\
X_1&X_2&D&\I_{p, q } U_1^t&\I_{p, q } U_2^t\\
0&z&X_1^t\I_{p, q}&-a&-c\\
-z&0&X_2^t\I_{p, q}&-b&-d\end{pmatrix } }
&: = & { \begin{pmatrix }
\frac{a+d}2&-w&\frac12U&\frac12V\\
z&-\frac{a+d}2&-\frac12Y^t\I_{p, q}&\frac12X^t\I_{p, q}\\
X_1&-\I_{p, q } U_2^t&D+\frac{d - a}2\I_n&-c\I_n\\
X_2&\I_{p, q } U_1^t&-b\I_n&D+\frac{a - d}2\I_n\end{pmatrix}}.\end{aligned}$$
\[lem1\ ] The map $ i: Q\to\tilde P$ is an injective Lie group homomorphism, $ { \alpha}:\g\to\tg$ is linear, and the pair $ (i,{\alpha})$ satisfies the condition (1)–(3) from \[3.2\ ].
therefore the pair $ (i,{\alpha})$ gives upgrade to an reference functor from Cartan geometries of type $ (G, Q)$ to Cartan geometry of type $ (\tilde G,\tilde P)$.
The map $ i$ is obviously well defined, i.e. the persona of an element of $ Q\subset PO(p+2,q+2)$ does not depend on the representative matrix. far, $ i$ is smooth and injective and a direct calculation indicate this is a homomorphism of Lie groups. The map $ { \alpha}$ is linear and the compatibility of the pair $ (i,{\alpha})$ follows as succeed: (1) involves a little long-winded but very straightforward checking, (2) is an easy exercise, (3) follows from $ \g/\q\cong\ | betw}|}} C\end{pmatrix}},\end{aligned}$$ $$\negin{aligned}
{\alphc}{\vegin{pkatrix}
z&b&U_1&0&w\\
c&d&U_2&-d&0\\
X_1&X_2&D&\I_{p,q} U_1^t&\I_{p,q} U_2^t\\
0&z&X_1^t\I_{p,q}&-a&-c\\
-z&0&X_2^v\I_{p,q}&-v&-d\end{kiatrix}}
&:=& {\begin{pmatrkx}
\frac{a+d}2&-w&\vrac12U&\frax12V\\
z&-\fcac{a+d}2&-\frac12Y^t\I_{p,q}&\fred12X^t\I_{p,q}\\
W_1&-\N_{p,q} U_2^f&F+\frae{d-e}2\I_n&-c\I_n\\
X_2&\I_{p,q} U_1^t&-b\I_k&D+\frac{a-d}2\I_n\etd{pmatrix}}.\end{alhgveb}$$
\[lem1\] The map $i:Q\to\tilde P$ is an injeceive Lir hroup homomorprism, ${\wlphz}:\g\to\tg$ is linear, and the pair $(i,{\alpga})$ sativfies the concitions (1)–(3) from \[3.2\].
Hence the palr $(i,{\wlpha})$ gives rise tl an extensuon synctor from Zartan geometries of tipe $(G,Q)$ to Cartan geometries of thpe $(\tnlde G,\tilde P)$.
Rhe kap $i$ is obtiouslj well defined, i.e. the hmage og an element on $Q\suuset PO(p+2,q+2)$ does not depend on the representajive matrif. Yurther, $i$ is smooth abd injewtiva ana a airtct ckmputahioi shows thia is a homonorphism of Lie groipf. The map ${\alphz}$ is lynqar and the compatibility of the pair $(i,{\dlpga})$ follows as follows: (1) unvolves a little tedlous but dery straightforward checking, (2) is an easy exercisa, (3) fomuoww nrom $\t/\q\fong\ | beta}|}} C\end{pmatrix}},\end{aligned}$$ $$\begin{aligned} {\alpha}{\begin{pmatrix} a&b&U_1&0&w\\ c&d&U_2&-w&0\\ X_1&X_2&D&\I_{p,q} 0&z&X_1^t\I_{p,q}&-a&-c\\ &:=& {\begin{pmatrix} z&-\frac{a+d}2&-\frac12Y^t\I_{p,q}&\frac12X^t\I_{p,q}\\ X_1&-\I_{p,q} U_2^t&D+\frac{d-a}2\I_n&-c\I_n\\ $i:Q\to\tilde is an injective group homomorphism, ${\alpha}:\g\to\tg$ linear, and the pair $(i,{\alpha})$ satisfies conditions (1)–(3) from \[3.2\]. Hence the pair $(i,{\alpha})$ gives rise to an extension from Cartan geometries of type $(G,Q)$ to Cartan geometries of type $(\tilde G,\tilde The $i$ obviously defined, i.e. the image of an element of $Q\subset PO(p+2,q+2)$ does not depend on the representative Further, $i$ is smooth and injective and a computation shows this is homomorphism of Lie groups. The ${\alpha}$ linear and compatibility the $(i,{\alpha})$ follows as (1) involves a little tedious but very straightforward checking, (2) is an easy exercise, (3) follows from | beta}|}} C\end{pmatrix}},\end{aligned}$$ $$\Begin{alignEd}
{\alpHa}{\bEgiN{pMatrIx}
a&b&u_1&0&w\\
c&d&U_2&-w&0\\
X_1&X_2&D&\I_{p,q} U_1^t&\i_{P,q} U_2^t\\
0&Z&X_1^t\I_{p,q}&-a&-c\\
-z&0&X_2^t\I_{p,q}&-b&-d\end{pmaTrix}}
&:=& {\bEgIN{pmaTRiX}
\frac{A+d}2&-w&\frac12u&\FrAC12v\\
z&-\fRaC{a+D}2&-\frAc12y^T\I_{P,q}&\fraC12X^t\i_{p,q}\\
X_1&-\I_{p,q} u_2^t&D+\frac{d-a}2\I_N&-c\I_N\\
X_2&\i_{p,q} U_1^t&-b\I_n&D+\fraC{A-d}2\i_n\end{pmatrIx}}.\eNd{aligned}$$
\[lem1\] the Map $i:Q\tO\tIldE p$ is an InjEctivE Lie grOUp homoMorphism, ${\aLpHA}:\g\to\tg$ IS linear, AND tHe paIr $(i,{\alpha})$ satisfies THe COnditions (1)–(3) from \[3.2\].
HEnce thE pAIr $(I,{\ALphA})$ giVes rise to aN eXtensIOn functOR fROM carTAn geometries oF type $(G,Q)$ to CaRTan GeometRiEs oF Type $(\tiLde G,\tIlDE P)$.
THe map $i$ is obvIousLy well defIned, i.e. THe image OF an elemEnt of $Q\SubSet pO(p+2,q+2)$ DOeS nOt dEpENd oN ThE rePResEntative MaTrIx. FurTher, $I$ IS SMootH anD injEctivE and a direct coMpuTatiON shOws thIs is a HomoMoRphisM of Lie GroupS. THe map ${\alpha}$ is linEar aNd the compAtiBiLitY oF the pAIr $(i,{\alpHa})$ fOllOws as foLlows: (1) inVOlvEs A LITtLe tedious but very stRaIGHtForward cHeckinG, (2) Is An EAsy exercIsE, (3) foLlowS FRom $\g/\q\Cong\ | beta}|}} C\end{pmatrix}},\ end{aligne d}$$$$\ beg in {ali gned }
{\alpha}{\be g in{p matrix}
a&b&U_1&0&w\\
c&d& U_ 2 &-w& 0 \\
X_1 &X_2&D& \ I_ { p ,q} U _1 ^t& \I _ {p ,q} U _2^ t\\
0& z&X_1^t\I_ {p, q} &-a&-c\\
-z& 0 &X _2^t\I_{p, q}& -b&-d\end{pm atr ix}} &: =&{ \begi n{p matri x}
\fr a c{a+d} 2&-w&\fra c1 2 U&\fra c 12V\\
z & - \f rac{ a+d}2&-\frac12Y^t \ I_ { p,q}&\frac12X^ t\I_{p ,q } \\ X_1 &-\ I_{p,q} U_ 2^ t&D+\ f rac{d-a } 2\ I _ n &-c \ I_n\\
X_2&\I_ {p,q} U_1^t & -b\ I_n&D+ \f rac { a-d}2\ I_n\e nd { pma trix}}.\end {ali gned}$$
\[lem1 \ ] The m a p $i:Q\ to\til deP$is a n i nj ect iv e Li e g rou p ho momorphi sm ,${\al pha} : \ g \ to\t g$is l inear , and the pai r $ (i,{ \ alp ha})$ sati sfie sthe c onditi ons ( 1) –(3) from \[3.2 \].
Hence th e p ai r $ (i ,{\al p ha})$giv esrise to an ext e nsi on f u nc tor from Cartan ge om e t ri es of ty pe $(G , Q) $t o Cartan g eom etri e s of t ype$ (\ tilde G, \tilde P) $.
The m ap $i$ i sobv iou sly w e ll d efined , i.e. t he im a ge of an eleme n t of $Q\subse t P O ( p+ 2 ,q+2 )$does not de pend on t he r e pr ese n tativ e mat ri x .F urther, $i$ is smoo th and i nject ive and a dir ect comput a t i on shows thi s i s a homomorphis m ofLie groups . The map ${\a lpha}$ i s lineara n d the co mpa tib ili tyo f t he pair $(i,{ \ a lpha }) $ follo wsas foll ows : ( 1)inv ol ves a lit tle tedi ou sbu tver y str a ightforw ar d c he cki ng, ( 2 ) is a n eas y ex er ci s e,(3) fol l ow s from $ \g /\q\ con g\ | beta}|}} C\end{pmatrix}},\end{aligned}$$_$$\begin{aligned}
{\alpha}{\begin{pmatrix}
a&b&U_1&0&w\\
c&d&U_2&-w&0\\_
X_1&X_2&D&\I_{p,q} U_1^t&\I_{p,q} U_2^t\\
0&z&X_1^t\I_{p,q}&-a&-c\\
-z&0&X_2^t\I_{p,q}&-b&-d\end{pmatrix}}_
&:=& {\begin{pmatrix}
\frac{a+d}2&-w&\frac12U&\frac12V\\
z&-\frac{a+d}2&-\frac12Y^t\I_{p,q}&\frac12X^t\I_{p,q}\\_
X_1&-\I_{p,q}_U_2^t&D+\frac{d-a}2\I_n&-c\I_n\\
X_2&\I_{p,q} U_1^t&-b\I_n&D+\frac{a-d}2\I_n\end{pmatrix}}.\end{aligned}$$
\[lem1\]_The_map $i:Q\to\tilde P$_is an injective_Lie group homomorphism, ${\alpha}:\g\to\tg$_is linear, and_the_pair $(i,{\alpha})$ satisfies the conditions (1)–(3) from \[3.2\].
Hence the pair $(i,{\alpha})$ gives rise to_an_extension functor_from_Cartan_geometries of type $(G,Q)$ to_Cartan geometries of type $(\tilde_G,\tilde P)$.
The_map $i$ is obviously well defined, i.e. the image_of_an element of_$Q\subset PO(p+2,q+2)$ does not depend on the representative matrix._Further, $i$ is smooth and injective_and a direct_computation_shows_this is a homomorphism_of Lie groups. The map ${\alpha}$_is linear and the compatibility of_the pair $(i,{\alpha})$ follows as follows: (1)_involves a little tedious but very_straightforward checking, (2) is an_easy exercise,_(3) follows from $\g/\q\cong\ |
has 4 singular fibers. We will use $\mathscr{W}^{\circ}\rightarrow B^{\circ}$ to denote the locus over which this morphism is smooth.
\[h1CC\] Let $C$ be a smooth member of the family $\mathscr{W}$. Then $H^0(C,\omega_{C})$ is a 10-dimensional complex representation of $\mathcal{A}_5$ that is isomorphic to $V\oplus I\oplus I^{'}$ where $V$ is a 4 dimensional irreducible representation. Moreover $H^1(C;{\mathds{C}})$ is isomorphic to $V^{\oplus2}\oplus (I\oplus I^{'})^{\oplus 2}$.
Choose a generator $\mu$ of $\wedge^3 U^\vee$ and regard this as a translation-invariant $3$-form on $U$. Note that this form is ${\mathscr I}$-invariant. Let $F\in {\mathds{C}}[U]_6^{\mathscr I}$ be a defining equation for $C$. Then a meromorphic 3-forms on $U$ defined by $$\Omega_G=G\frac{\mu}{F}, \quad \text{ with } G\in {\mathds{C}}[U]_3$$ has a simple pole along the cone over $C$ (defined by $F=0$) and at the hyperplane at infinity ${\mathds{P}}(U)$. This enabled us to define a linear map $$\begin{aligned}
{\mathds{C}}[U]_3&\rightarrow H^0(C,\omega_{C})\\
G&\mapsto {\operatorname{Res}}_C{\operatorname{Res}}_{\infty}\Omega_G.
\end{aligned}$$ This map is known to be injective see Chapter 5 of [@griffiths2014principles] and since both ${\mathds{C}}[U]_3$ and $H^0(C,\omega_{C})$ are of dimension $10$, it must be an isomorphism. As it is also ${\mathscr I}$-equivariant, it follows that $H^0(C,\omega_{C})$ is as a ${\mathscr I}$-representation isomorphic with ${\mathds{C}}[U]_3$. We know that $U^{\vee}$ is either $I$ or $I^{'}$ as a$\mathcal{A}_5$-representation. Recall the following character formula for symmetric powers $$\chi_{{\mathds{C}}[U]_3}(g)=\frac{\chi_{U^{\vee}}^3(g)+3 | has 4 singular fibers. We will use $ \mathscr{W}^{\circ}\rightarrow B^{\circ}$ to announce the venue over which this morphism is placid.
\[h1CC\ ] Let $ C$ be a legato penis of the family $ \mathscr{W}$. Then $ H^0(C,\omega_{C})$ is a 10 - dimensional complex representation of $ \mathcal{A}_5 $ that is isomorphous to $ V\oplus I\oplus I^{'}$ where $ V$ is a 4 dimensional irreducible representation. Moreover $ H^1(C;{\mathds{C}})$ is isomorphic to $ V^{\oplus2}\oplus (I\oplus I^{'})^{\oplus 2}$.
Choose a generator $ \mu$ of $ \wedge^3 U^\vee$ and see this as a translation - invariant $ 3$-form on $ U$. Note that this shape is $ { \mathscr I}$-invariant. Let $ F\in { \mathds{C}}[U]_6^{\mathscr I}$ be a defining equality for $ C$. Then a meromorphic 3 - forms on $ U$ defined by $ $ \Omega_G = G\frac{\mu}{F }, \quad \text { with } G\in { \mathds{C}}[U]_3$$ has a simple pole along the cone over $ C$ (defined by $ F=0 $) and at the hyperplane at infinity $ { \mathds{P}}(U)$. This enable us to define a linear map $ $ \begin{aligned }
{ \mathds{C}}[U]_3&\rightarrow H^0(C,\omega_{C})\\
G&\mapsto { \operatorname{Res}}_C{\operatorname{Res}}_{\infty}\Omega_G.
\end{aligned}$$ This function is known to be injective see Chapter 5 of [ @griffiths2014principles ] and since both $ { \mathds{C}}[U]_3 $ and $ H^0(C,\omega_{C})$ are of dimension $ 10 $, it must be an isomorphism. As it is also $ { \mathscr I}$-equivariant, it follows that $ H^0(C,\omega_{C})$ is as a $ { \mathscr I}$-representation isomorphous with $ { \mathds{C}}[U]_3$. We know that $ U^{\vee}$ is either $ I$ or $ I^{'}$ as a$\mathcal{A}_5$-representation. hark back the following character formula for symmetric power $ $ \chi_{{\mathds{C}}[U]_3}(g)=\frac{\chi_{U^{\vee}}^3(g)+3 | had 4 singular fibers. We wiul use $\mathscr{W}^{\eurc}\rigitarrow B^{\circ}$ tu denote the locus over whici thus moephism is smooth.
\[h1CC\] Leg $C$ be a dmooth mwmbec of the family $\mathscr{W}$. Then $G^0(G,\omegc_{C})$ is a 10-dimensiokal complex representatiot uf $\mathcal{A}_5$ that is isomorphic to $V\opjus I\opkud I^{'}$ where $V$ is a 4 cymenapokal irreducible representation. Mkreover $H^1(C;{\mathds{C}})$ is isomorphic to $V^{\oplus2}\oplus (I\oppus I^{'})^{\oplus 2}$.
Choose w generator $\mu$ jd $\wedge^3 U^\vee$ and regard this as a jranslation-invariant $3$-form on $U$. Nute tkat this fotm id ${\mathscr I}$-iivariagt. Let $F\in {\mathds{C}}[U]_6^{\mdthscr O}$ be a defininn equetiob for $C$. Then a meromocphic 3-forms on $U$ defyned by $$\Okeya_G=G\frac{\mu}{F}, \quad \text{ wuth } C\in {\matfes{C}}[J]_3$$ hzs a simplf pkle along fhe cone ovwr $C$ (defined by $F=0$) amd qt the hyperpmane ae ynfinity ${\mathds{P}}(U)$. This enabled us to dexins a linear map $$\begin{alitned}
{\mathds{C}}[U]_3&\rightarroa H^0(C,\omega_{S})\\
G&\mapsto {\operatorname{Res}}_C{\operatorname{Res}}_{\infty}\Omegd_G.
\end{elkgntd}$$ This nao is known to be injective see Chapter 5 of [@gryrfotms2014principles] and since both ${\mwtnqs{C}}[U]_3$ and $H^0(C,\omgga_{C})$ arz or dimension $10$, it mudt be ag isonorphism. Ws iy is also ${\mathscr I}$-equivariqnt, it follocs rhat $H^0(C,\omega_{C})$ is ad a ${\mathscr I}$-rekresenyation isomorphic with ${\oathss{C}}[U]_3$. We knoa that $U^{\vsd}$ is either $I$ or $I^{'}$ ss a$\mathcal{A}_5$-representation. Rqcall the folkowing zharscter sormula fog symmetric powers $$\chi_{{\mahhds{C}}[O]_3}(g)=\frac{\whi_{U^{\vee}}^3(g)+3 | has 4 singular fibers. We will use to the locus which this morphism be smooth member of family $\mathscr{W}$. Then is a 10-dimensional complex representation of that is isomorphic to $V\oplus I\oplus I^{'}$ where $V$ is a 4 dimensional representation. Moreover $H^1(C;{\mathds{C}})$ is isomorphic to $V^{\oplus2}\oplus (I\oplus I^{'})^{\oplus 2}$. Choose a generator of U^\vee$ regard as a translation-invariant $3$-form on $U$. Note that this form is ${\mathscr I}$-invariant. Let $F\in {\mathds{C}}[U]_6^{\mathscr be a defining equation for $C$. Then a 3-forms on $U$ defined $$\Omega_G=G\frac{\mu}{F}, \quad \text{ with } {\mathds{C}}[U]_3$$ a simple along cone $C$ (defined by and at the hyperplane at infinity ${\mathds{P}}(U)$. This enabled us to define a linear map $$\begin{aligned} {\mathds{C}}[U]_3&\rightarrow G&\mapsto {\operatorname{Res}}_C{\operatorname{Res}}_{\infty}\Omega_G. map is to injective Chapter 5 of since both ${\mathds{C}}[U]_3$ and $H^0(C,\omega_{C})$ are it must be an isomorphism. As it is ${\mathscr I}$-equivariant, follows that $H^0(C,\omega_{C})$ is as a I}$-representation isomorphic with ${\mathds{C}}[U]_3$. We know that $U^{\vee}$ either $I$ or $I^{'}$ as a$\mathcal{A}_5$-representation. Recall the following character formula for symmetric powers $$\chi_{{\mathds{C}}[U]_3}(g)=\frac{\chi_{U^{\vee}}^3(g)+3 | has 4 singular fibers. We will usE $\mathscr{W}^{\cIrc}\riGhtArrOw b^{\cirC}$ to dEnote the locus oVEr whIch this morphism is smootH.
\[h1CC\] LEt $c$ Be a sMOoTh memBer of thE FaMILy $\mAtHsCr{W}$. thEN $H^0(c,\omegA_{C})$ iS a 10-dimenSional compLex RePresentation OF $\mAthcal{A}_5$ thaT is Isomorphic to $v\opLus I\opLuS I^{'}$ wHEre $V$ iS a 4 dImensIonal iRReduciBle represEnTAtion. MOReover $H^1(c;{\MAtHds{C}})$ Is isomorphic to $V^{\opLUs2}\OPlus (I\oplus I^{'})^{\oplUs 2}$.
ChooSe A GeNERatOr $\mU$ of $\wedge^3 U^\vEe$ And reGArd this AS a TRANslATion-invariant $3$-Form on $U$. Note THat This foRm Is ${\mAThscr I}$-InvarIaNT. LeT $F\in {\mathds{C}}[u]_6^{\matHscr I}$ be a dEfininG EquatioN For $C$. TheN a meroMorPhiC 3-forMS oN $U$ DefInED by $$\oMeGa_G=g\FraC{\mu}{F}, \quad \TeXt{ With } G\In {\maTHDS{c}}[U]_3$$ haS a sImplE pole Along the cone oVer $c$ (defINed By $F=0$) anD at thE hypErPlane At infiNity ${\mAtHds{P}}(U)$. This enableD us tO define a lIneAr Map $$\BeGin{alIGned}
{\maThdS{C}}[U]_3&\RightarRow H^0(C,\omEGa_{C})\\
g&\mAPSTo {\Operatorname{Res}}_C{\opErATOrName{Res}}_{\iNfty}\OmEGa_g.
\eND{aligned}$$ thIs mAp is KNOwn to Be inJEcTive see CHapter 5 OF [@gRiFfiths2014pRiNcipleS] aNd sIncE both ${\MAthdS{C}}[U]_3$ and $h^0(C,\omega_{C})$ Are of DImension $10$, it must BE an isomorphisM. as IT Is ALso ${\mAthScr I}$-equivarIant, IT folLows THaT $H^0(C,\OMega_{C})$ Is as a ${\MaTHsCR I}$-representation isoMoRphic wIth ${\maThds{C}}[U]_3$. We know tHat $U^{\vee}$ is eITHEr $I$ or $I^{'}$ as A$\matHCaL{a}_5$-representatioN. RecaLl the folloWIng charaCter fOrmula foR symmetriC POwers $$\chi_{{\MatHds{c}}[U]_3}(g)=\FraC{\CHi_{u^{\vee}}^3(g)+3 | has 4 singular fibers. We will use$\mat hsc r{W }^ {\ci rc}\ rightarrow B^{ \ circ }$ to denote the locus over w h icht hi s mor phism i s s m o oth .
\ [h1 CC \ ]Let $ C$be a sm ooth membe r o fthe family $ \ ma thscr{W}$. Th en $H^0(C,\o meg a_{C}) $isa 10-d ime nsion al com p lex re presentat io n of $\ m athcal{ A } _5 $ th at is isomorphict o$ V\oplus I\oplu s I^{' }$ wh e r e $ V$is a 4 dim en siona l irredu c ib l e rep r esentation. M oreover $H^ 1 (C; {\math ds {C} } )$ isisomo rp h icto $V^{\opl us2} \oplus (I \oplus I^{'})^ { \oplus2}$.
Cho ose a g e ne ra tor $ \ mu$ of $\ w edg e^3 U^\v ee $and r egar d t h is a s a tra nslat ion-invariant $3 $-fo r m o n $U$ . Not e th at this formis ${ \m athscr I}$-inva rian t. Let $F \in { \ma th ds{C} } [U]_6^ {\m ath scr I}$ be a d e fin in g e qu ation for $C$. The na me romorphi c 3-fo r ms o n $U$ def in edby $ $ \ Omega _G=G \ fr ac{\mu}{ F}, \q u ad \ text{ w it h } G\ in { \ma thds{ C }}[U ]_3$$has a si mplep ole along thec one over $C$( de f i ne d by$F= 0$) and attheh yper plan e a t i n finit y ${\ ma t hd s {P}}(U)$. This enab le d us t o def ine a linearmap $$\beg i n { aligned}
{\m a th d s{C}}[U]_3&\ri ghtar row H^0(C, \ omega_{C })\\G&\mapst o {\opera t o rname{Re s}} _C{ \op era t o rn ame{Res}}_{\i n f ty}\ Om ega_G.\en d{align ed} $$Thi s m ap is known to be i nj ec ti ve se e Cha p ter 5 of [ @gr if fit hs201 4 princi ples] and s in c e b oth ${\ m at h d s{C} }[ U] _3$and $ H^0(C ,\om e ga_ {C})$ a re of dim ens i on $ 10 $, it mus t be an isomo rp hism. As i tisalso $ { \ mathscrI}$-equivariant, it fol l ows tha t $ H^0(C ,\om ega_{C})$ is as a${\ m athscr I}$-r epres en tat i o n iso m o rp hic w ith ${\mat h d s{C }}[U] _3 $. W e knowthat $U^{\vee}$ is eit her $I$ or $I ^{' }$ a s a$ \ma t hc a l{A }_ 5 $-r e p resentation. Re call the f ol l ow ing charac t erfo rmula f or symm etric powers$$\chi_{{ \mathds{C }} [U]_ 3 } (g) =\frac{\ch i_{U^{\v ee}}^3(g) + 3 | has_4 singular_fibers. We will use_$\mathscr{W}^{\circ}\rightarrow B^{\circ}$_to_denote the_locus_over which this_morphism is smooth.
\[h1CC\]_Let $C$ be a_smooth member of_the_family $\mathscr{W}$. Then $H^0(C,\omega_{C})$ is a 10-dimensional complex representation of $\mathcal{A}_5$ that is isomorphic_to_$V\oplus I\oplus_I^{'}$_where_$V$ is a 4 dimensional_irreducible representation. Moreover $H^1(C;{\mathds{C}})$ is_isomorphic to_$V^{\oplus2}\oplus (I\oplus I^{'})^{\oplus 2}$.
Choose a generator $\mu$ of_$\wedge^3_U^\vee$ and regard_this as a translation-invariant $3$-form on $U$. Note that_this form is ${\mathscr I}$-invariant. Let_$F\in {\mathds{C}}[U]_6^{\mathscr I}$_be_a_defining equation for $C$._Then a meromorphic 3-forms on $U$_defined by $$\Omega_G=G\frac{\mu}{F}, \quad \text{ with_} G\in {\mathds{C}}[U]_3$$ has a simple_pole along the cone over $C$_(defined by $F=0$) and at_the hyperplane_at infinity ${\mathds{P}}(U)$. This enabled_us to define_a linear_map $$\begin{aligned}
{\mathds{C}}[U]_3&\rightarrow H^0(C,\omega_{C})\\
G&\mapsto_{\operatorname{Res}}_C{\operatorname{Res}}_{\infty}\Omega_G.
\end{aligned}$$ This map is known to_be injective see_Chapter 5 of [@griffiths2014principles] and since_both_${\mathds{C}}[U]_3$ and $H^0(C,\omega_{C})$_are_of_dimension $10$,_it must be_an_isomorphism. As_it_is also ${\mathscr I}$-equivariant, it follows_that_$H^0(C,\omega_{C})$ is as a ${\mathscr I}$-representation isomorphic_with ${\mathds{C}}[U]_3$. We know_that_$U^{\vee}$ is either $I$_or $I^{'}$ as a$\mathcal{A}_5$-representation. Recall_the following character formula for symmetric_powers $$\chi_{{\mathds{C}}[U]_3}(g)=\frac{\chi_{U^{\vee}}^3(g)+3 |
which is finite only if $w$ is in $A_{p}$ and seek estimates for this in order to improve the above estimate (\[a\]). One may add more variables to the above Bellman function as the integrals of $w$ and of $w^{-1/(p-1)}$ over $Q$ but we will not treat those here. The estimates here will be proved in the general setting of tree like families on probability spaces and its related maximal operator, as will be described in the next section.
‘We will derive two types of estimates related to the above problems. In the first we will use a related condition on some symmetrization of the weight to find the exact form of a related to weights Bellman function and this is done in section 2. Then in section 3 we obtain certain new estimates for the above Bellman function related to $A_{p}$ with respect to a tree, and to the corresponding maximal operator, by using an estimate of the maximal operator via two applications of maximal operators on the same tree but with different measures, and this is described in section 3.
There are several other problems in Harmonic Analysis where Bellman functions naturally arise. Such problems (including the dyadic Carleson imbedding and weighted inequalities) are described in [@Naz2] (see also [@Naz], [@Naz1]) and also connections to Stochastic Optimal Control are provided, from which it follows that the corresponding Bellman functions satisfy certain nonlinear second order PDE.
The exact computation of a Bellman function is a difficult task which is connected with the deeper structure of the corresponding Harmonic Analysis problem. Thus far several Bellman functions have been computed (see [@Burk1], [@Burk2], [@Mel1], [@Sla], [@Sla1], [@Vas], [@Vas1], [@Vas2]). L.Slavin and A.Stokolos [@SlSt] linked the Bellman function computation to solving certain PDE’s of the Monge Ampere type, and in this way they obtained an alternative proof of the Bellman functions relate to the dyadic maximal operator in [@Mel1]. Also in [@Vas2] using the Monge-Ampere equation approach a more general Bellman function than the one related to the dyadic Carleson imbedding Theorem has be precisely evaluated thus generalizing the corresponding result in [@Mel1].
Trees, maximal operators and symmetrization
===========================================
As | which is finite only if $ w$ is in $ A_{p}$ and seek estimates for this in order to better the above appraisal (\[a\ ]). One may add more variables to the above Bellman affair as the integral of $ w$ and of $ w^{-1/(p-1)}$ over $ Q$ but we will not treat those here. The estimates here will be proved in the cosmopolitan mount of tree like families on probability quad and its related maximal operator, as will be described in the future section.
‘ We will derive two type of estimates relate to the above problems. In the first we will practice a related condition on some symmetrization of the weight to find the accurate form of a related to weights Bellman function and this is done in part 2. Then in section 3 we obtain certain new estimates for the above Bellman affair related to $ A_{p}$ with respect to a tree, and to the corresponding maximal operator, by using an estimate of the maximal operator via two applications of maximal operators on the same tree but with different measures, and this is described in section 3.
There are several other problem in Harmonic Analysis where Bellman functions naturally arise. Such trouble (including the dyadic Carleson imbedding and weighted inequalities) are trace in [ @Naz2 ] (see also [ @Naz ], [ @Naz1 ]) and also connections to Stochastic Optimal Control are provide, from which it follows that the corresponding Bellman functions satisfy certain nonlinear second order PDE.
The accurate computation of a Bellman function is a difficult task which is connected with the deeper structure of the comparable Harmonic Analysis trouble. therefore far several Bellman functions have been computed (see [ @Burk1 ], [ @Burk2 ], [ @Mel1 ], [ @Sla ], [ @Sla1 ], [ @Vas ], [ @Vas1 ], [ @Vas2 ]). L.Slavin and A.Stokolos [ @SlSt ] linked the Bellman function computation to solving sealed PDE ’s of the Monge Ampere type, and in this means they obtained an alternative validation of the Bellman functions relate to the dyadic maximal operator in [ @Mel1 ]. Also in [ @Vas2 ] using the Monge - Ampere equality approach a more general Bellman function than the one related to the dyadic Carleson implant Theorem has be precisely evaluated thus popularize the corresponding result in [ @Mel1 ].
tree, maximal operator and symmetrization
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
As | whlch is finite only if $w$ ls in $A_{p}$ and seek estimetes fod this iv order to improve the above ewtimauv (\[a\]). One may add more xariables to the qbovt Bellman functioi as the integdwls mh $w$ and of $w^{-1/(p-1)}$ oyer $Q$ but wa will not tredt tkose here. The estimates here will be proved ij the general fettpnd of nrte like families on probability slaces aid its related kaximal operator, as will bf dedcribed in the nexh section.
‘We wilj derive two gypes of estimates relzted to the above problems. In tfe fixst we will uwe w related coiditiog on some symmetrizadion of the weight to fiid tye exact form of a renated to weights Bgllman funwtnon and this is done un secthon 2. Thev in sedtmon 3 we ohtamn certain hew estimatws for the above Bekliqn function rslated tj $A_{p}$ with respect to a tree, and to the borrssponding maximal operaror, by using an estimwte of thq maximal operator via two applications of maximan opecagorw jv tje same tree but with different measures, and egix ps described in scction 3.
There are srvfrsj other problgms in Kzrjonic Analysis whege Bellian fynctions gatutally arise. Such problems (ibcluding the eyadic Carleson imyedding and ceightgd ineaualities) are described in [@Haz2] (see alsl [@Naz], [@Naz1]) xnd also connectkonx do Stochastic Optimal Contwol are pcovidzd, from dhicn it fjllows thah the corresponding Belllan fonctiots satisfy certain nonlinear second order PDE.
The exact vokpunation of a Beklman functiog is a difficukt task whicf is connebted with the deeper structure of jhe corresponving Harmjnic Anaoysis pfublem. Thus far several Ytllman funxtions have been cpmpotsd (see [@Burk1], [@Burj2], [@Mtl1], [@Sla], [@Sla1], [@Vas], [@Vss1], [@Xas2]). L.Dlevin wtd A.Stokolos [@SlSg] lkmked ghe Bellman wuncyion computation to volvjng certain PDE’s og bhe Monge Ampere eype, and in tnis way they obtaijed ai altecnativr ptoof of the Bellman functions rslate to hhe dyadic maxiial iperator in [@Kel1]. Also in [@Vas2] using the Monge-Ampere ewuation approach a more general Bellmcn function tian thq one reldted to the dyadic Cqrleson imbedding Theorem has be preciaely eealuahed thus generalizing the corresponding result in [@Mel1].
Trees, maximal operatirs anv fymmetrizatjon
===========================================
Ax | which is finite only if $w$ is and estimates for in order to One add more variables the above Bellman as the integrals of $w$ and $w^{-1/(p-1)}$ over $Q$ but we will not treat those here. The estimates here be proved in the general setting of tree like families on probability spaces its maximal as be described in the next section. ‘We will derive two types of estimates related to the problems. In the first we will use a condition on some symmetrization the weight to find the form a related weights function this is done section 2. Then in section 3 we obtain certain new estimates for the above Bellman function related $A_{p}$ with a tree, to corresponding operator, by using of the maximal operator via two operators on the same tree but with different and this described in section 3. There are other problems in Harmonic Analysis where Bellman functions arise. Such problems (including the dyadic Carleson imbedding and weighted inequalities) are described in [@Naz2] [@Naz], [@Naz1]) and also to Stochastic Optimal are from it that the Bellman functions satisfy certain nonlinear second order PDE. The exact computation a Bellman function is a difficult task which is connected deeper of the corresponding Analysis problem. Thus far Bellman have been computed (see [@Mel1], [@Sla1], L.Slavin A.Stokolos linked the Bellman function to solving certain PDE’s of Monge Ampere type, and an alternative proof of the Bellman functions relate the dyadic maximal operator in [@Mel1]. Also [@Vas2] using the Monge-Ampere equation approach a more general Bellman function than one related dyadic Carleson imbedding Theorem has be precisely evaluated generalizing the corresponding result [@Mel1]. Trees, maximal operators and symmetrization =========================================== As | which is finite only if $w$ is in $A_{P}$ and seek esTimatEs fOr tHiS in oRder To improve the abOVe esTimate (\[a\]). One may add more vaRiablEs TO the ABoVe BelLman funCTiON As tHe InTegRaLS oF $w$ and Of $w^{-1/(P-1)}$ over $Q$ bUt we will noT trEaT those here. ThE EsTimates herE wiLl be proved in The GeneraL sEttINg of tRee Like fAmilieS On probAbility spAcES and itS Related MAXiMal oPerator, as will be deSCrIBed in the next seCtion.
‘WE wILl DERivE twO types of esTiMates RElated tO ThE ABOve PRoblems. In the fIrst we will uSE a rElated CoNdiTIon on sOme syMmETriZation of the WeigHt to find tHe exacT Form of a RElated tO weighTs BEllMan fUNcTiOn aNd THis IS dOne IN seCtion 2. TheN iN sEctioN 3 we oBTAIN cerTaiN new EstimAtes for the aboVe BEllmAN fuNctioN relaTed tO $A_{P}$ with RespecT to a tReE, and to the corresPondIng maximaL opErAtoR, bY usinG An estiMatE of The maxiMal operATor ViA TWO aPplications of maximAl OPErAtors on tHe same TReE bUT with difFeRenT meaSURes, anD thiS Is DescribeD in secTIoN 3.
THere are SeVeral oThEr pRobLems iN harmOnic AnAlysis whEre BeLLman functions nATurally arise. SUCh PRObLEms (iNclUding the dyaDic CARlesOn imBEdDinG And weIghteD iNEqUAlities) are described In [@naz2] (see Also [@NAz], [@Naz1]) and also cOnnections TO sTochastiC OptIMaL control are provIded, fRom which it FOllows thAt the CorrespoNding BellMAN functioNs sAtiSfy CerTAIn Nonlinear secoND OrdeR PdE.
The exAct ComputaTioN of A BeLlmAn Function iS a difficUlT tAsK wHicH is coNNected wiTh The DeEpeR struCTure of The coRresPoNdINg HArmonic aNaLYSis pRoBlEm. ThUs fAr SeverAl BeLLmaN functiOns have beEn cOMputEd (SeE [@Burk1], [@BuRk2], [@Mel1], [@Sla], [@Sla1], [@VaS], [@VAs1], [@Vas2]). L.SlavIn And a.StokoLOS [@SlSt] linKed the Bellman function coMPutatioN to SolviNg ceRtain PDE’s Of tHe MongE AmPEre typE, and in This wAy TheY OBtainED An AltErNative prooF OF thE BellMaN funCtions rElate to the dyadic maXImaL operator in [@MeL1]. AlSo in [@vAS2] uSinG ThE monGe-aMpeRE Equation approacH a more geneRaL beLlman functIOn tHaN the one Related To the DYadic CaRleson imbEdding TheOrEm haS BE prEcisely evaLuated thUs generalIZing tHE cOrresPonDing reSuLt iN [@Mel1].
TRees, maXImaL operAtors aNd SymmetRizatIoN
===========================================
As | which is finite only if $ w$ is in $ A_{p} $ a ndse ek e stim ates for thisi n or der to improve the abo ve es ti m ate( \[ a\]). One ma y a d d mo re v ari ab l es to t heabove B ellman fun cti on as the inte g ra ls of $w$and of $w^{-1/( p-1 )}$ ov er $Q $ butwewillnot tr e at tho se here.Th e estim a tes her e wi ll b e proved in the g e ne r al setting oftree l ik e f a m ili eson probabi li ty sp a ces and it s r ela t ed maximal op erator, asw ill be de sc rib e d in t he ne xt sec tion.
‘Wewill derive t wo typ e s of es t imatesrelate d t o t he a b ov epro bl e ms. In th e fi rst we w il luse a rel a t e d con dit ionon so me symmetriza tio n of the weig ht to fin dthe e xact f orm o fa related to we ight s Bellman fu nc tio nand t h is isdon e i n secti on 2. T h enin s e ct ion 3 we obtain ce rt a i nnew esti matesf or t h e aboveBe llm an f u n ction rel a te d to $A_ {p}$ w i th r espectto a tre e, an d t o the corr espond ing maxi mal o p erator, by usi n g an estimate of t he maxi mal operator v ia t w o ap plic a ti ons of ma ximal o p er a tors on the same tr ee but w ith d ifferent meas ures, andt h i s is des crib e di n section 3.
There are sever a l otherprobl ems in H armonic A n a lysis wh ere Be llm anf u nc tions natural l y ari se . Suchpro blems ( inc lud ing th edyadic Ca rleson i mb ed di ng an d wei g hted ine qu ali ti es) ared escrib ed in [@N az 2] (se e also[ @N a z ], [ @N az 1])and a lso c onne c tio ns to S tochastic Op t imal C on trol ar e provided, f ro m which it f oll ows th a t the cor responding Bellman func t ions sa tis fy ce rtai n nonline arsecond or d er PDE .
The exac tcom p u tatio n of aBe llman func t i onis adi ffic ult tas k which is connect e d w ith the deepe r s truc t u re of th e co rr e spo n d ing Harmonic An alysis pro bl e m. Thus fars eve ra l Bellm an func tions have be en comput ed (see [ @B urk1 ] , [@ Burk2], [@ Mel1], [ @Sla], [@ S la1], [@ Vas], [@ Vas1], [ @Va s2]). L.Sla v inand A .Stoko lo s [@Sl St] l in ked theBellman function comput ationto so lvi ng certai n P D E’s of the M onge Ampere ty pe, an d inthi s waythey ob tai n ed an alt e rnative p r oo f o f th e Bellman f u n c tio ns re lat e to th e dy adic maximal oper a tor in [@Mel1] . Al s o in [@ V as2] u sing the Monge -Am pe r e equatio napproach amore gen er a l Bel lman f unctio n thant h eo ne rel ated to the dyad icCa r leson i mb ed d ing Th eore mhas be preci s elye v aluated thus gen erali z i ng th e co rresp on ding re s ultin [@Mel1] .
Trees, m aximal ope rator s and s ym metriz ati on
========= = ========= ===== ======= == ==== === ===
A s | which_is finite_only if $w$ is_in $A_{p}$_and_seek estimates_for_this in order_to improve the_above estimate (\[a\]). One_may add more_variables_to the above Bellman function as the integrals of $w$ and of $w^{-1/(p-1)}$ over_$Q$_but we_will_not_treat those here. The estimates_here will be proved in_the general_setting of tree like families on probability spaces_and_its related maximal_operator, as will be described in the next section.
‘We_will derive two types of estimates_related to the_above_problems._In the first we_will use a related condition on_some symmetrization of the weight to_find the exact form of a related_to weights Bellman function and this_is done in section 2._Then in_section 3 we obtain certain_new estimates for_the above_Bellman function related_to $A_{p}$ with respect to a_tree, and to_the corresponding maximal operator, by using_an_estimate of the_maximal_operator_via two_applications of maximal_operators_on the_same_tree but with different measures, and_this_is described in section 3.
There are several_other problems in Harmonic_Analysis_where Bellman functions naturally_arise. Such problems (including the_dyadic Carleson imbedding and weighted inequalities)_are described_in [@Naz2]_(see also [@Naz], [@Naz1]) and also connections to Stochastic Optimal Control_are provided, from which it follows_that the corresponding Bellman_functions satisfy_certain_nonlinear second order_PDE.
The_exact computation_of a Bellman function is a difficult_task which_is connected with the deeper structure_of the corresponding Harmonic_Analysis_problem. Thus far several Bellman functions_have been computed (see [@Burk1], [@Burk2],_[@Mel1], [@Sla], [@Sla1], [@Vas], [@Vas1],_[@Vas2])._L.Slavin_and A.Stokolos [@SlSt] linked the_Bellman function computation to solving certain_PDE’s of the_Monge Ampere type, and in this way_they_obtained an alternative proof of the_Bellman_functions relate to the dyadic maximal_operator_in_[@Mel1]. Also in [@Vas2] using_the Monge-Ampere equation approach a more_general Bellman function than the one related to the_dyadic Carleson imbedding_Theorem has be precisely evaluated_thus_generalizing_the corresponding result in [@Mel1].
Trees, maximal operators and symmetrization
===========================================
As |
\overline{\Hv^{p}_{\alpha, \beta n}(n)}$.
The next lemma is the first step towards the proof of \[prop:structural\] and motivates the subsequent discussion. In its statement, the support of a vector $a=(a_1,\dots,a_n) \in \F_{p}^{n}$, denoted by $\operatorname{supp}(a)$, refers to the set of indices $i\in [n]$ such that $a_i \neq 0 \mod p$.
\[lemma:eliminate-small-support\] Let $M_{n}$ be an $n\times n$, $\{\pm 1\}$-valued symmetric matrix, chosen uniformly at random from among all such matrices. Let $1 \leq d \leq n$ be an integer, and let $p$ be a prime. Let $\Sv^{p}_{\geq d, \beta n}(n)$ denote the event that every vector in $\F_{p}^{n}\setminus\{\operatorname{\textbf{0}}\}$ which is orthogonal (over $\F_{p}$) to at least $(1-\beta)n$ many rows of $M_{n}$ has support of size at least $d$. Suppose further that $\beta \leq 1/2$, $d\leq n/2$, $p^{\beta n} \leq 2^{n/2}$, $p^{d} \leq 2^{n/8}$, $H(\beta) \leq 1/4$, and $H(d/n) \leq 1/16$ (where $H(x):= -x\log_{2}(x) - (1-x)\log_{2}(1-x)$ is the binary entropy function for $x\in [0,1]$) Then, $$\Pr_{M_n}\left[\overline{\Sv^{p}_{\geq d,\beta n}(n)}\right] \leq 2^{-n/16}.$$
The proof of this lemma will use the following simple, yet powerful, observation.
\[obs:inject-permutation\] Let $\Sigma$ be an $n\times n$ permutation matrix. Then, for a uniformly random $n\times n$ symmetric $\{\pm 1\}$-matrix $M_n$, the random matrix $\Sigma^{-1}M_n\Sigma$ is also a uniformly distributed $n\times n$ symmetric $\{\pm 1\}$-matrix.
It is clear than | \overline{\Hv^{p}_{\alpha, \beta n}(n)}$.
The next lemma is the first step towards the proof of \[prop: structural\ ] and motivate the subsequent discussion. In its instruction, the support of a vector $ a=(a_1,\dots, a_n) \in \F_{p}^{n}$, denoted by $ \operatorname{supp}(a)$, refer to the bent of indices $ i\in [ n]$ such that $ a_i \neq 0 \mod p$.
\[lemma: eliminate - little - support\ ] get $ M_{n}$ be an $ n\times n$, $ \{\pm 1\}$-valued symmetric matrix, chosen uniformly at random from among all such matrices. Let $ 1 \leq five hundred \leq n$ be an integer, and let $ p$ be a prime. get $ \Sv^{p}_{\geq d, \beta n}(n)$ announce the event that every vector in $ \F_{p}^{n}\setminus\{\operatorname{\textbf{0}}\}$ which is orthogonal (over $ \F_{p}$) to at least $ (1-\beta)n$ many row of $ M_{n}$ has support of size at least $ d$. think further that $ \beta \leq 1/2 $, $ d\leq n/2 $, $ p^{\beta n } \leq 2^{n/2}$, $ p^{d } \leq 2^{n/8}$, $ H(\beta) \leq 1/4 $, and $ H(d / n) \leq 1/16 $ (where $ H(x):= -x\log_{2}(x) - (1 - x)\log_{2}(1 - x)$ is the binary entropy function for $ x\in [ 0,1]$) Then, $ $ \Pr_{M_n}\left[\overline{\Sv^{p}_{\geq d,\beta n}(n)}\right ] \leq 2^{-n/16}.$$
The proof of this lemma will practice the following simple, so far powerful, notice.
\[obs: inject - permutation\ ] Let $ \Sigma$ be an $ n\times n$ substitution matrix. Then, for a uniformly random $ n\times n$ symmetric $ \{\pm 1\}$-matrix $ M_n$, the random matrix $ \Sigma^{-1}M_n\Sigma$ is also a uniformly distributed $ n\times n$ symmetric $ \{\pm 1\}$-matrix.
It is well-defined than | \ovfrline{\Hv^{p}_{\alpha, \beta n}(n)}$.
Tht next lemma is tkw firsv step fowards ghe proof of \[prop:structural\] end notivqtes the subsequent dircussion. Pn its stqtemtnt, the support oh a vector $a=(a_1,\dkbs,a_n) \nn \F_{p}^{n}$, denoted bi $\operatornake{supp}(a)$, refers tu che set of indices $i\in [n]$ such that $a_y \neq 0 \kof p$.
\[lemma:eliminwte-skwll-shipirt\] Let $M_{n}$ be an $n\times n$, $\{\pm 1\}$-balued vymmetric mattix, chosen uniformly at rajdom from among all sufh matrices. Let $1 \leq d \leq n$ be an inttgzr, and let $k$ be a prime. Let $\Sv^{p}_{\geq d, \beta n}(v)$ denpte the evgur tjdt every vertor ig $\F_{p}^{n}\setminus\{\operatosname{\tectbf{0}}\}$ which is prtiogobal (over $\F_{p}$) to at leavt $(1-\beta)n$ many rows of $M_{n}$ hav aupport of size ar oeast $d$. Sgpporw fjrtgec tgat $\beha \meq 1/2$, $d\leq h/2$, $p^{\beta n} \lwq 2^{n/2}$, $p^{d} \leq 2^{n/8}$, $H(\beta) \lqw 1/4$, and $H(d/n) \les 1/16$ (whewe $H(x):= -x\log_{2}(x) - (1-x)\log_{2}(1-x)$ is the binary entropy xundtion for $x\in [0,1]$) Then, $$\Pr_{M_b}\left[\overline{\Sv^{p}_{\geq d,\heta n}(n)}\ridht] \leq 2^{-n/16}.$$
The proof of this lemma will use the folnowinj rimklc, hwt powerful, observation.
\[obs:inject-permutation\] Let $\Aibms$ be an $n\times n$ permutatooj kwtrix. Then, fot a uniykrjly random $n\times j$ symmejric $\{\pn 1\}$-matrix $I_n$, tne random matrix $\Sigma^{-1}M_n\Signa$ is also a yniformly distribuced $n\times n$ symketriv $\{\pm 1\}$-matrix.
It is clear chan | \overline{\Hv^{p}_{\alpha, \beta n}(n)}$. The next lemma is step the proof \[prop:structural\] and motivates statement, support of a $a=(a_1,\dots,a_n) \in \F_{p}^{n}$, by $\operatorname{supp}(a)$, refers to the set indices $i\in [n]$ such that $a_i \neq 0 \mod p$. \[lemma:eliminate-small-support\] Let $M_{n}$ an $n\times n$, $\{\pm 1\}$-valued symmetric matrix, chosen uniformly at random from among such Let \leq \leq n$ be an integer, and let $p$ be a prime. Let $\Sv^{p}_{\geq d, \beta n}(n)$ the event that every vector in $\F_{p}^{n}\setminus\{\operatorname{\textbf{0}}\}$ which orthogonal (over $\F_{p}$) to least $(1-\beta)n$ many rows of has of size least Suppose that $\beta \leq $d\leq n/2$, $p^{\beta n} \leq 2^{n/2}$, $p^{d} \leq 2^{n/8}$, $H(\beta) \leq 1/4$, and $H(d/n) \leq 1/16$ (where -x\log_{2}(x) - the binary function $x\in Then, $$\Pr_{M_n}\left[\overline{\Sv^{p}_{\geq d,\beta 2^{-n/16}.$$ The proof of this lemma following simple, yet powerful, observation. \[obs:inject-permutation\] Let $\Sigma$ an $n\times permutation matrix. Then, for a uniformly $n\times n$ symmetric $\{\pm 1\}$-matrix $M_n$, the random $\Sigma^{-1}M_n\Sigma$ is also a uniformly distributed $n\times n$ symmetric $\{\pm 1\}$-matrix. It is clear than | \overline{\Hv^{p}_{\alpha, \beta n}(n)}$.
The Next lemma iS the fIrsT stEp TowaRds tHe proof of \[prop:sTRuctUral\] and motivates the subSequeNt DIscuSSiOn. In iTs stateMEnT, THe sUpPoRt oF a VEcTor $a=(a_1,\DotS,a_n) \in \F_{p}^{N}$, denoted by $\OpeRaTorname{supp}(a)$, REfErs to the seT of Indices $i\in [n]$ sUch That $a_i \NeQ 0 \moD P$.
\[lemmA:elIminaTe-smalL-SupporT\] Let $M_{n}$ be aN $n\TImes n$, $\{\pM 1\}$-Valued sYMMeTric Matrix, chosen unifoRMlY At random from amOng all SuCH mATRicEs. LEt $1 \leq d \leq n$ Be An intEGer, and lET $p$ BE A PriME. Let $\Sv^{p}_{\geq d, \beTa n}(n)$ denote tHE evEnt thaT eVerY Vector In $\F_{p}^{n}\SeTMinUs\{\operatornAme{\tExtbf{0}}\}$ whicH is ortHOgonal (oVEr $\F_{p}$) to aT least $(1-\BetA)n$ mAny rOWs Of $m_{n}$ hAs SUppORt Of sIZe aT least $d$. SUpPoSe furTher THAT $\Beta \Leq 1/2$, $D\leq N/2$, $p^{\betA n} \leq 2^{n/2}$, $p^{d} \leq 2^{n/8}$, $H(\BetA) \leq 1/4$, ANd $H(D/n) \leq 1/16$ (Where $h(x):= -x\lOg_{2}(X) - (1-x)\log_{2}(1-X)$ is the BinarY eNtropy function fOr $x\iN [0,1]$) Then, $$\Pr_{M_n}\LefT[\oVerLiNe{\Sv^{p}_{\GEq d,\betA n}(n)}\RigHt] \leq 2^{-n/16}.$$
THe proof OF thIs LEMMa Will use the followinG sIMPlE, yet poweRful, obSErVaTIon.
\[obs:inJeCt-pErmuTATion\] LEt $\SiGMa$ Be an $n\timEs n$ perMUtAtIon matrIx. then, foR a UniForMly raNDom $n\Times n$ SymmetriC $\{\pm 1\}$-maTRix $M_n$, the random MAtrix $\Sigma^{-1}M_n\SIGmA$ IS aLSo a uNifOrmly distriButeD $N\timEs n$ sYMmEtrIC $\{\pm 1\}$-maTrix.
IT iS ClEAr than | \overline{\Hv^{p}_{\alpha , \beta n} (n)}$ .
The n extlemm a is the first step towards the proof of\[pro p: s truc t ur al\]and mot i va t e s t he s ubs eq u en t dis cus sion. I n its stat eme nt , the suppor t o f a vector $a =(a_1,\dots, a_n ) \in\F _{p } ^{n}$ , d enote d by $ \ operat orname{su pp } (a)$,r efers t o th e se t of indices $i\i n [ n ]$ such that $ a_i \n eq 0\ m odp$.
\[lemma: el imina t e-small - su p p o rt\ ] Let $M_{n}$be an $n\ti m esn$, $\ {\ pm1 \}$-va luedsy m met ric matrix, cho sen unifo rmly a t random from am ong al l s uch mat r ic es . L et $1\ le q d \le q n$ bean i ntege r, a n d l et $ p$be a prim e. Let $\Sv^{ p}_ {\ge q d, \bet a n}( n)$de notethe ev ent t ha t every vectorin $ \F_{p}^{n }\s et min us \{\op e ratorn ame {\t extbf{0 }}\}$ w h ich i s o rt hogonal (over $\F_ {p } $ )to at le ast $( 1 -\ be t a)n$ man yrow s of $ M_{n} $ ha s s upport o f size at l east $d $. Suppo se fu rth er th a t $\ beta \ leq 1/2$ , $d\ l eq n/2$, $p^{\ b eta n} \leq 2 ^ {n / 2 }$ , $p^ {d} \leq 2^{n/ 8}$, $H(\ beta ) \ leq 1/4$, and$H ( d/ n ) \leq 1/16$ (where $ H(x):= -x\l og_{2}(x) - ( 1-x)\log_{ 2 } ( 1-x)$ is the bi n ary entropy fu nctio n for $x\i n [0,1]$) Then , $$\Pr_ {M_n}\lef t [ \overlin e{\ Sv^ {p} _{\ g e qd,\beta n}(n) } \ righ t] \leq 2 ^{- n/16}.$ $
The pr oof o f this le mma will u se t he fo llowi n g simple ,yet p owe rful, observ ation .
\ [o bs : inj ect-per m ut a t ion\ ]Le t $\ Sig ma $ bean $ n \ti mes n$permutati onm atri x. T hen, fo r a uniformly r andom $n\t im esn$ sym m e tric $\{ \pm 1\}$-matrix $M_n$,t he rand ommatri x $\ Sigma^{-1 }M_ n\Sigm a$i s also a uni forml ydis t r ibute d $n \ti me s n$ symme t r ic$\{\p m1\}$ -matrix .
It is clear tha n | \overline{\Hv^{p}_{\alpha,_\beta n}(n)}$.
The_next lemma is the_first step_towards_the proof_of_\[prop:structural\] and motivates_the subsequent discussion._In its statement, the_support of a_vector_$a=(a_1,\dots,a_n) \in \F_{p}^{n}$, denoted by $\operatorname{supp}(a)$, refers to the set of indices $i\in [n]$_such_that $a_i_\neq_0_\mod p$.
\[lemma:eliminate-small-support\] Let $M_{n}$ be_an $n\times n$, $\{\pm 1\}$-valued_symmetric matrix,_chosen uniformly at random from among all such_matrices._Let $1 \leq_d \leq n$ be an integer, and let $p$_be a prime. Let $\Sv^{p}_{\geq d,_\beta n}(n)$ denote_the_event_that every vector in_$\F_{p}^{n}\setminus\{\operatorname{\textbf{0}}\}$ which is orthogonal (over $\F_{p}$)_to at least $(1-\beta)n$ many rows_of $M_{n}$ has support of size at_least $d$. Suppose further that $\beta_\leq 1/2$, $d\leq n/2$, $p^{\beta_n} \leq_2^{n/2}$, $p^{d} \leq 2^{n/8}$, $H(\beta)_\leq 1/4$, and_$H(d/n) \leq_1/16$ (where $H(x):=_-x\log_{2}(x) - (1-x)\log_{2}(1-x)$ is the binary_entropy function for_$x\in [0,1]$) Then, $$\Pr_{M_n}\left[\overline{\Sv^{p}_{\geq d,\beta n}(n)}\right]_\leq_2^{-n/16}.$$
The proof of_this_lemma_will use_the following simple,_yet_powerful, observation.
\[obs:inject-permutation\]_Let_$\Sigma$ be an $n\times n$ permutation_matrix._Then, for a uniformly random $n\times n$_symmetric $\{\pm 1\}$-matrix $M_n$,_the_random matrix $\Sigma^{-1}M_n\Sigma$ is_also a uniformly distributed $n\times_n$ symmetric $\{\pm 1\}$-matrix.
It is clear_than |
all the functions $q_{t_i}$ have finite support then all the needed analytic assumptions are automatically satisfied because we are working in an $L^2$ space on a finite set. Of course, such a restriction is unnecessarily harsh, but we will not pursue this issue here any further.
Proofs of the results from Section \[sec:examples\] {#sec:app}
===================================================
We recall the following facts about trace class and Hilbert-Schmidt norms (see e.g. [@Simon]) of operators in $L^2(X)$ for some measurable space $(X,\Sigma,\mu)$, which we will use repeatedly without reference: $$\begin{gathered}
\|AB\|_1\leq\|A\|_2\|B\|_2,\qquad\|AB\|_1\leq\|A\|_{\rm op}\|B\|_1,\\
\shortintertext{and if $A$ has integral kernel $A(x,y)$,}
\|A\|_2=\left(\int \mu(dx)\mu(dy)\,|A(x,y)|^2\right)^{1/2},\end{gathered}$$ for each $A,B$ in the appropriate space, where $\|\cdot\|_{\rm op}$ denotes the operator norm in $L^2(X)$.
Throughout this section $c$ and $c'$ will denote positive constants whose value may change from line to line.
Checking Assumption \[assum:1\] is straightforward. We will take in this case $V_{t_i}=V_{t_i}'=U_{t_i}=U_{t_i}'=I$, and thus Assumption \[assum:0\] is contained in Assumption \[assum:2\], which we check next.
Condition (i) is trivial. Given functions $\psi_1$ and $\psi_2$ write $\psi_1\otimes\psi_2$ for the kernel $\psi_1(x)\psi_2(y)$ and let $\phi$ be any function with $\int\!\phi^2=1$. Then we can write ${\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)=({\overline{Q}}_{t_i}\varphi_k\otimes\phi)(\phi\otimes\varphi_k)$, so that $\|{\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)\|_1\leq\|{\overline{Q | all the functions $ q_{t_i}$ have finite support then all the needed analytic assumption are mechanically satisfied because we are working in an $ L^2 $ quad on a finite bent. Of course, such a restriction is unnecessarily coarse, but we will not pursue this issue here any further.
proof of the results from Section \[sec: examples\ ] { # sec: app }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
We recall the keep up facts about tracing course and Hilbert - Schmidt norms (see for example [ @Simon ]) of operators in $ L^2(X)$ for some measurable quad $ (X,\Sigma,\mu)$, which we will practice repeatedly without reference: $ $ \begin{gathered }
\|AB\|_1\leq\|A\|_2\|B\|_2,\qquad\|AB\|_1\leq\|A\|_{\rm op}\|B\|_1,\\
\shortintertext{and if $ A$ has integral kernel $ A(x, y)$, }
\|A\|_2=\left(\int \mu(dx)\mu(dy)\,|A(x, y)|^2\right)^{1/2},\end{gathered}$$ for each $ vitamin a, B$ in the appropriate space, where $ \|\cdot\|_{\rm op}$ denotes the hustler norm in $ L^2(X)$.
Throughout this department $ c$ and $ c'$ will denote positivist constants whose value may change from line to cable.
Checking Assumption \[assum:1\ ] is straightforward. We will take in this case $ V_{t_i}=V_{t_i}'=U_{t_i}=U_{t_i}'=I$, and thus Assumption \[assum:0\ ] is contain in Assumption \[assum:2\ ], which we determine next.
Condition (i) is trivial. Given functions $ \psi_1 $ and $ \psi_2 $ write $ \psi_1\otimes\psi_2 $ for the kernel $ \psi_1(x)\psi_2(y)$ and let $ \phi$ be any affair with $ \int\!\phi^2=1$. Then we can publish $ { \overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)=({\overline{Q}}_{t_i}\varphi_k\otimes\phi)(\phi\otimes\varphi_k)$, so that $ \|{\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)\|_1\leq\|{\overline{Q | alp the functions $q_{t_i}$ have finite support then anl the needed xnalytic assumptions are autlmqticaoly satisfied because de are wogking in qn $L^2$ wpace on a finite set. Of goursz, wuch a restricjion is unnewessarily harsv, cuc we will not pursue this issue here any futtjer.
Proofs of tre rtsujts rgon Section \[sec:examples\] {#sec:app}
===================================================
Ws recaln the followimg facts about trace class and Hilbert-Schmidt nogms (see e.g. [@Wimog]) of operatorr in $L^2(X)$ fog some measorable space $(X,\Sigma,\mu)$, which we wkll uxe repeateelt wljhout refereice: $$\bedin{gathered}
\|AB\|_1\leq\|A\|_2\|B\|_2,\xquad\|AB\|_1\keq\|A\|_{\rm op}\|B\|_1,\\
\shprtmnteetext{and if $A$ has intxgral kernel $A(x,y)$,}
\|A\|_2=\lgft(\int \mu(df)\mb(dy)\,|A(x,y)|^2\right)^{1/2},\end{gatheree}$$ dor edch $D,B$ iv thd akprppdiate dpare, where $\|\cdkt\|_{\rm op}$ denites the operator npri in $L^2(X)$.
Throughkut thys section $c$ and $c'$ will denote positive cmnsfants whose value may cyange from line to lije.
Checkind Assumption \[assum:1\] is straightforward. We will taka in vhks eqse $V_{g_u}=V_{h_i}'=U_{t_i}=U_{t_i}'=I$, and thus Assumption \[assum:0\] is contaigsd ik Assumption \[assmm:2\], which we check medt.
Vjndition (i) is trivicm. Fiven functions $\psl_1$ and $\pfi_2$ wrute $\psi_1\otymes\lsi_2$ for the kernel $\psi_1(x)\psi_2(y)$ and let $\phi$ ve any function wich $\int\!\phi^2=1$. Thzn we van wtite ${\overline{Q}}_{t_i}(\varphi_k\utimss\varphi_k)=({\ovfrline{Q}}_{t_i}\bxrphi_k\otimes\phi)(\pfi\onimev\varphi_k)$, so that $\|{\overline{Q}}_{e_i}(\varphi_k\itimzs\varphi_y)\|_1\leq\|{\pverlige{Q | all the functions $q_{t_i}$ have finite support the analytic assumptions automatically satisfied because $L^2$ on a finite Of course, such restriction is unnecessarily harsh, but we not pursue this issue here any further. Proofs of the results from Section {#sec:app} =================================================== We recall the following facts about trace class and Hilbert-Schmidt norms e.g. of in for some measurable space $(X,\Sigma,\mu)$, which we will use repeatedly without reference: $$\begin{gathered} \|AB\|_1\leq\|A\|_2\|B\|_2,\qquad\|AB\|_1\leq\|A\|_{\rm op}\|B\|_1,\\ \shortintertext{and $A$ has integral kernel $A(x,y)$,} \|A\|_2=\left(\int \mu(dx)\mu(dy)\,|A(x,y)|^2\right)^{1/2},\end{gathered}$$ for $A,B$ in the appropriate where $\|\cdot\|_{\rm op}$ denotes the norm $L^2(X)$. Throughout section and will denote positive whose value may change from line to line. Checking Assumption \[assum:1\] is straightforward. We will take in case $V_{t_i}=V_{t_i}'=U_{t_i}=U_{t_i}'=I$, Assumption \[assum:0\] contained Assumption which we check (i) is trivial. Given functions $\psi_1$ $\psi_1\otimes\psi_2$ for the kernel $\psi_1(x)\psi_2(y)$ and let $\phi$ any function $\int\!\phi^2=1$. Then we can write ${\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)=({\overline{Q}}_{t_i}\varphi_k\otimes\phi)(\phi\otimes\varphi_k)$, that $\|{\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)\|_1\leq\|{\overline{Q | all the functions $q_{t_i}$ have finIte support Then aLl tHe nEeDed aNalyTic assumptions ARe auTomatically satisfied beCause We ARe woRKiNg in aN $L^2$ space ON a FINitE sEt. of cOuRSe, Such a ResTrictioN is unnecesSarIlY harsh, but we wILl Not pursue tHis Issue here any FurTher.
PrOoFs oF The reSulTs froM SectiON \[sec:exAmples\] {#sec:ApP}
===================================================
we recaLL the folLOWiNg faCts about trace clasS AnD hilbert-Schmidt Norms (sEe E.G. [@SIMOn]) oF opErators in $L^2(x)$ fOr somE MeasuraBLe SPACe $(X,\sIgma,\mu)$, which we Will use repeATedLy withOuT reFErence: $$\Begin{GaTHerEd}
\|AB\|_1\leq\|A\|_2\|B\|_2,\qqUad\|Ab\|_1\leq\|A\|_{\rm op}\|b\|_1,\\
\shortINtertexT{And if $A$ hAs inteGraL keRnel $a(X,y)$,}
\|a\|_2=\lEft(\InT \Mu(dX)\Mu(Dy)\,|A(X,Y)|^2\riGht)^{1/2},\end{gaThErEd}$$ for Each $a,b$ IN The aPprOpriAte spAce, where $\|\cdot\|_{\rM op}$ DenoTEs tHe opeRator Norm In $l^2(X)$.
ThrOughouT this SeCtion $c$ and $c'$ will dEnotE positive ConStAntS wHose vALue may ChaNge From linE to line.
cHecKiNG aSsUmption \[assum:1\] is straIgHTFoRward. We wIll takE In ThIS case $V_{t_i}=v_{t_I}'=U_{t_I}=U_{t_i}'=i$, ANd thuS AssUMpTion \[assuM:0\] is conTAiNeD in AssuMpTion \[asSuM:2\], whIch We cheCK nexT.
CondiTion (i) is tRiviaL. given functions $\PSi_1$ and $\psi_2$ write $\PSi_1\OTImES\psi_2$ For The kernel $\psI_1(x)\psI_2(Y)$ and Let $\pHI$ bE anY FunctIon wiTh $\INt\!\PHi^2=1$. Then we can write ${\oveRlIne{Q}}_{t_i}(\VarphI_k\otimes\varphI_k)=({\overline{q}}_{T_I}\Varphi_k\oTimeS\PhI)(\Phi\otimes\varphI_k)$, so tHat $\|{\overlinE{q}}_{t_i}(\varphI_k\otiMes\varphI_k)\|_1\leq\|{\overLINe{Q | all the functions $q_{t_i }$ have fi nitesup por tthen all the needed an a lyti c assumptions are auto matic al l y sa t is fiedbecause we a rewo rk ing i n a n $L^ 2$space o n a finite se t. Of course,s uc h a restri cti on is unnece ssa rily h ar sh, but w e w ill n ot pur s ue thi s issue h er e any f u rther.P ro ofsof the results fr o mS ection \[sec:e xample s\ ] { # s ec: app }
======== == ===== = ======= = == = = = === = ============= =====
We r e cal l thefo llo w ing fa cts a bo u t t race classandHilbert-S chmidt norms ( s ee e.g. [@Sim on] ) o f op e ra to rsin $L^ 2 (X )$f orsome mea su ra ble s pace $ ( X ,\Si gma ,\mu )$, w hich we willuse rep e ate dly w ithou t re fe rence : $$\b egin{ ga thered}
\|AB\ |_1\ leq\|A\|_ 2\| B\ |_2 ,\ qquad \ |AB\|_ 1\l eq\ |A\|_{\ rm op}\ | B\| _1 , \ \
\shortintertext{a nd i f$A$ hasintegr a lke r nel $A(x ,y )$, }
\ | A\|_2 =\le f t( \int \mu (dx)\m u (d y) \,|A(x, y) |^2\ri gh t)^ {1/ 2},\e n d{ga thered }$$ foreach$ A,B$ in the ap p ropriate spac e ,w h er e $\| \cd ot\|_{\rm o p}$d enot es t h eope r atornormin $L ^ 2(X)$.
Throughoutth is sec tion$c$ and $c'$will denot e p ositivecons t an t s whose valuemay c hange from line toline.
Checki ng Assump t i on \[ass um: 1\] is st r a ig htforward. We w illta ke in t his case $ V_{ t_i }=V _{t _i }'=U_{t_i }=U_{t_i }' =I $, a ndthusA ssumptio n\[a ss um: 0\] i s conta inedin A ss um p tio n \[ass u m: 2 \ ], w hi ch weche ck next .
C o ndi tion (i ) is triv ial . Giv en f unction s $\psi_1$ an d$\psi_2$ w ri te$\psi_ 1 \ otimes\p si_2$ for the kernel $\ p si_1(x) \ps i_2(y )$ a nd let $\ phi $ be a nyf unctio n with $\in t\ !\p h i ^2=1$ . Th enwe can write $ {\o verli ne {Q}} _{t_i}( \varphi_k\otimes\v a rph i_k)=({\overl ine {Q}} _ { t_ i}\ v ar p hi_ k\ o tim e s \phi)(\phi\otim es\varphi_ k) $ ,so that $\ | {\o ve rline{Q }}_{t_i }(\va r phi_k\o times\var phi_k)\|_ 1\ leq\ | { \ov erline{Q | all_the functions_$q_{t_i}$ have finite support_then all_the_needed analytic_assumptions_are automatically satisfied_because we are_working in an $L^2$_space on a_finite_set. Of course, such a restriction is unnecessarily harsh, but we will not pursue_this_issue here_any_further.
Proofs_of the results from Section_\[sec:examples\] {#sec:app}
===================================================
We recall the following_facts about_trace class and Hilbert-Schmidt norms (see e.g. [@Simon])_of_operators in $L^2(X)$_for some measurable space $(X,\Sigma,\mu)$, which we will use_repeatedly without reference: $$\begin{gathered}
\|AB\|_1\leq\|A\|_2\|B\|_2,\qquad\|AB\|_1\leq\|A\|_{\rm_op}\|B\|_1,\\
\shortintertext{and_if_$A$_has integral kernel $A(x,y)$,}
_ \|A\|_2=\left(\int \mu(dx)\mu(dy)\,|A(x,y)|^2\right)^{1/2},\end{gathered}$$ for each $A,B$_in the appropriate space, where $\|\cdot\|_{\rm_op}$ denotes the operator norm in $L^2(X)$.
Throughout_this section $c$ and $c'$ will_denote positive constants whose value_may change_from line to line.
Checking Assumption_\[assum:1\] is straightforward._We will_take in this_case $V_{t_i}=V_{t_i}'=U_{t_i}=U_{t_i}'=I$, and thus Assumption \[assum:0\]_is contained in_Assumption \[assum:2\], which we check next.
Condition_(i)_is trivial. Given_functions_$\psi_1$_and $\psi_2$_write $\psi_1\otimes\psi_2$ for_the_kernel $\psi_1(x)\psi_2(y)$_and_let $\phi$ be any function with_$\int\!\phi^2=1$._Then we can write ${\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)=({\overline{Q}}_{t_i}\varphi_k\otimes\phi)(\phi\otimes\varphi_k)$, so that_$\|{\overline{Q}}_{t_i}(\varphi_k\otimes\varphi_k)\|_1\leq\|{\overline{Q |
mu_{i,0}^A - \mu_{i,0,-S}^A)\right)(Y_{i,1} - \overline Y_1),\\
A_{2n} &=& \frac{1}{\sqrt{n}v^2}\sum_{i \in N}(\mu_{i,0}^A - \mu_{i,0,-S}^A)\left((Y_{i,1} - \overline Y_1) - \mathbf{E}_\mathscr{F}[Y_{i,1} - \overline Y_1]\right),\text{ and }\\
A_{3n} &=& \frac{1}{\sqrt{n}}\left( \frac{1}{\hat v^2} - \frac{1}{v^2} \right) \sum_{i \in N}(\hat \mu_{i,0}^A - \hat \mu_{i,0,-S}^A)(Y_{i,1} - \overline Y_1).\end{aligned}$$
As for $A_{1n}$, we write $$\begin{aligned}
\label{decomp31}
&& \frac{1}{\sqrt{n}}\sum_{i \in N}\left(\hat \mu_{i,0}^A - \mu_{i,0}^A\right)(Y_{i,1} - \overline Y_1)\\ \notag
&=& \frac{1}{\sqrt{n}}\sum_{i \in N} \frac{1}{|N(i)|}\sum_{j \in N(i)} \frac{1}{n}\sum_{k \in N: k \ne j} \varepsilon_{k,0} w_{kj} \left(Y_{i,1} - \overline Y_1\right)\\ \notag
&=& \frac{1}{\sqrt{n}}\sum_{i \in N} \rho_{i,0}^A \left(Y_{i,1} - \overline Y_1\right).\end{aligned}$$ Applying the same argument to the term with $X_{i,-S}$’s, and Lemma \[lemm: deg U\], we conclude that $$\begin{aligned}
A_{1n} = \frac{1}{v^2 \sqrt{n}}\sum_{i \in N} (\rho_{i,0}^A - \rho_{i, | mu_{i,0}^A - \mu_{i,0,-S}^A)\right)(Y_{i,1 } - \overline Y_1),\\
A_{2n } & = & \frac{1}{\sqrt{n}v^2}\sum_{i \in N}(\mu_{i,0}^A - \mu_{i,0,-S}^A)\left((Y_{i,1 } - \overline Y_1) - \mathbf{E}_\mathscr{F}[Y_{i,1 } - \overline Y_1]\right),\text { and } \\
A_{3n } & = & \frac{1}{\sqrt{n}}\left (\frac{1}{\hat v^2 } - \frac{1}{v^2 } \right) \sum_{i \in N}(\hat \mu_{i,0}^A - \hat \mu_{i,0,-S}^A)(Y_{i,1 } - \overline Y_1).\end{aligned}$$
As for $ A_{1n}$, we write $ $ \begin{aligned }
\label{decomp31 }
& & \frac{1}{\sqrt{n}}\sum_{i \in N}\left(\hat \mu_{i,0}^A - \mu_{i,0}^A\right)(Y_{i,1 } - \overline Y_1)\\ \notag
& = & \frac{1}{\sqrt{n}}\sum_{i \in N } \frac{1}{|N(i)|}\sum_{j \in N(i) } \frac{1}{n}\sum_{k \in N: k \ne j } \varepsilon_{k,0 } w_{kj } \left(Y_{i,1 } - \overline Y_1\right)\\ \notag
& = & \frac{1}{\sqrt{n}}\sum_{i \in N } \rho_{i,0}^A \left(Y_{i,1 } - \overline Y_1\right).\end{aligned}$$ Applying the same argument to the condition with $ X_{i,-S}$ ’s, and Lemma \[lemm: deg U\ ], we reason that $ $ \begin{aligned }
A_{1n } = \frac{1}{v^2 \sqrt{n}}\sum_{i \in N } (\rho_{i,0}^A - \rho_{i, | mu_{i,0}^W - \mu_{i,0,-S}^A)\right)(Y_{i,1} - \overlint Y_1),\\
A_{2n} &=& \frac{1}{\sqtt{b}v^2}\sum_{i \in N}(\mh_{i,0}^A - \mu_{i,0,-R}^A)\left((Y_{i,1} - \overline Y_1) - \mathbf{X}_\matyscr{F}[T_{i,1} - \overline Y_1]\right),\texg{ and }\\
A_{3n} &=& \frax{1}{\sqru{n}}\left( \frac{1}{\hat v^2} - \frac{1}{v^2} \rlyht) \shl_{i \iu I}(\hat \mu_{i,0}^A - \hat \ku_{i,0,-S}^A)(Y_{i,1} - \oeerline Y_1).\end{alhgveb}$$
As for $A_{1n}$, we write $$\begin{aligned}
\labej{decomp31}
&& \frac{1}{\sqrt{n}}\sui_{i \im N}\lern(\hct \mu_{i,0}^A - \mu_{i,0}^A\right)(Y_{i,1} - \overline G_1)\\ \notag
&=& \frac{1}{\sqrt{m}}\sum_{i \in N} \frac{1}{|N(i)|}\sum_{j \in N(l)} \frwc{1}{n}\sum_{k \in N: k \ne u} \varepsilob_{k,0} w_{hh} \left(Y_{i,1} - \ovdrline Y_1\right)\\ \notag
&=& \frac{1}{\sqrt{n}}\sum_{i \in N} \rho_{i,0}^A \left(H_{i,1} - \orerline Y_1\rithr).\enf{dligned}$$ Appoying the same arnlment tm the trrm with $X_{i,-S}$’s, snd Lenma \[lemm: deg U\], we conrlude that $$\begin{aligged}
A_{1n} = \yrac{1}{v^2 \sqrt{n}}\sum_{i \in N} (\ehi_{i,0}^A - \tho_{i, | mu_{i,0}^A - \mu_{i,0,-S}^A)\right)(Y_{i,1} - \overline Y_1),\\ A_{2n} \in - \mu_{i,0,-S}^A)\left((Y_{i,1} \overline Y_1) - }\\ &=& \frac{1}{\sqrt{n}}\left( \frac{1}{\hat - \frac{1}{v^2} \right) \in N}(\hat \mu_{i,0}^A - \hat \mu_{i,0,-S}^A)(Y_{i,1} \overline Y_1).\end{aligned}$$ As for $A_{1n}$, we write $$\begin{aligned} \label{decomp31} && \frac{1}{\sqrt{n}}\sum_{i \in N}\left(\hat - \mu_{i,0}^A\right)(Y_{i,1} - \overline Y_1)\\ \notag &=& \frac{1}{\sqrt{n}}\sum_{i \in N} \frac{1}{|N(i)|}\sum_{j \in N(i)} \in k j} w_{kj} \left(Y_{i,1} - \overline Y_1\right)\\ \notag &=& \frac{1}{\sqrt{n}}\sum_{i \in N} \rho_{i,0}^A \left(Y_{i,1} - \overline Y_1\right).\end{aligned}$$ Applying same argument to the term with $X_{i,-S}$’s, and \[lemm: deg U\], we that $$\begin{aligned} A_{1n} = \frac{1}{v^2 \in (\rho_{i,0}^A - | mu_{i,0}^A - \mu_{i,0,-S}^A)\right)(Y_{i,1} - \overline Y_1),\\
a_{2n} &=& \frac{1}{\sqrt{N}v^2}\sum_{I \in n}(\mu_{I,0}^A - \Mu_{i,0,-S}^a)\lefT((Y_{i,1} - \overline Y_1) - \maTHbf{E}_\Mathscr{F}[Y_{i,1} - \overline Y_1]\rigHt),\texT{ aND }\\
A_{3n} &=& \fRAc{1}{\Sqrt{n}}\Left( \fraC{1}{\HaT V^2} - \FraC{1}{v^2} \RiGht) \SuM_{I \iN N}(\hat \Mu_{i,0}^a - \hat \mu_{i,0,-s}^A)(Y_{i,1} - \overliNe Y_1).\EnD{aligned}$$
As foR $a_{1n}$, We write $$\begIn{aLigned}
\label{dEcoMp31}
&& \frac{1}{\SqRt{n}}\SUm_{i \in n}\leFt(\hat \Mu_{i,0}^A - \mu_{I,0}^a\right)(y_{i,1} - \overlinE Y_1)\\ \NOtag
&=& \frAC{1}{\sqrt{n}}\sUM_{I \iN N} \frAc{1}{|N(i)|}\sum_{j \in N(i)} \frac{1}{n}\SUm_{K \In N: k \ne j} \varepsiLon_{k,0} w_{kJ} \lEFt(y_{I,1} - \OveRliNe Y_1\right)\\ \noTaG
&=& \frac{1}{\SQrt{n}}\sum_{I \In n} \RHO_{i,0}^A \LEft(Y_{i,1} - \overline y_1\right).\end{alIGneD}$$ ApplyInG thE Same arGumenT tO The Term with $X_{i,-S}$’S, and lemma \[lemm: Deg U\], we COnclude THat $$\begiN{alignEd}
A_{1N} = \frAc{1}{v^2 \sQRt{N}}\sUm_{i \In n} (\Rho_{I,0}^a - \rHo_{i, | mu_{i,0}^A - \mu_{i,0,-S}^ A)\right)( Y_{i, 1}- \ ov erli ne Y _1),\\
A_{ 2 n} & =& \frac{1}{\sqrt{n}v^ 2}\su m_ { i \i n N }(\mu _{i,0}^ A - \ mu_ {i ,0 ,-S }^ A )\ left( (Y_ {i,1} - \overline Y_ 1) - \mathbf{E } _\ mathscr{F} [Y_ {i,1} - \ove rli ne Y_1 ]\ rig h t),\t ext { and }\\
A_{3 n} &=& \f ra c {1}{\s q rt{n}}\ l e ft ( \f rac{1}{\hat v^2}- \ f rac{1}{v^2} \r ight)\s u m_ { i \i n N }(\hat \mu _{ i,0}^ A - \hat \m u _ { i,0 , -S}^A)(Y_{i,1 } - \overli n e Y _1).\e nd {al i gned}$ $
As f o r $ A_{1n}$, we wri te $$\beg in{ali g ned}
\l a bel{dec omp31}
& & \f r ac {1 }{\ sq r t{n } }\ sum _ {i\in N}\l ef t( \hat\mu_ { i , 0 }^A- \ mu_{ i,0}^ A\right)(Y_{i ,1} - \ o ver lineY_1)\ \ \n ot ag
&=&\frac {1 }{\sqrt{n}}\sum _{i\in N} \f rac {1 }{| N( i)|}\ s um_{j\in N( i)} \fr ac{1}{n } \su m_ { k \i n N: k \ne j} \var ep s i lo n_{k,0}w_{kj} \l ef t (Y_{i,1} - \o verl i n e Y_1 \rig h t) \\ \nota g
& =& \ frac{1} {\ sqrt{n }} \su m_{ i \in N} \ rho_{i ,0}^A \l eft(Y _ {i,1} - \overl i ne Y_1\right) . \e n d {a l igne d}$ $ Applyingthes ameargu m en t t o thetermwi t h$ X_{i,-S}$’s, and Le mm a \[le mm: d eg U\], we co nclude tha t $ $\begin{ alig n ed }
A_{1n} =\frac {1}{v^2 \s q rt{n}}\s um_{i \in N}(\rho_{i, 0 } ^A - \rh o_{ i, | mu_{i,0}^A -_\mu_{i,0,-S}^A)\right)(Y_{i,1} -_\overline Y_1),\\
_ A_{2n}_&=&_\frac{1}{\sqrt{n}v^2}\sum_{i \in_N}(\mu_{i,0}^A_- \mu_{i,0,-S}^A)\left((Y_{i,1} -_\overline Y_1) -_\mathbf{E}_\mathscr{F}[Y_{i,1} - \overline Y_1]\right),\text{_and }\\
__ A_{3n} &=& \frac{1}{\sqrt{n}}\left( \frac{1}{\hat v^2} - \frac{1}{v^2} \right) \sum_{i \in N}(\hat \mu_{i,0}^A -_\hat_\mu_{i,0,-S}^A)(Y_{i,1} -_\overline_Y_1).\end{aligned}$$
As_for $A_{1n}$, we write $$\begin{aligned}
\label{decomp31}
_ && \frac{1}{\sqrt{n}}\sum_{i_\in N}\left(\hat_\mu_{i,0}^A - \mu_{i,0}^A\right)(Y_{i,1} - \overline Y_1)\\ \notag
__ &=& \frac{1}{\sqrt{n}}\sum_{i_\in N} \frac{1}{|N(i)|}\sum_{j \in N(i)} \frac{1}{n}\sum_{k \in N: k_\ne j} \varepsilon_{k,0} w_{kj} \left(Y_{i,1} -_\overline Y_1\right)\\ \notag
___ &=& \frac{1}{\sqrt{n}}\sum_{i \in_N} \rho_{i,0}^A \left(Y_{i,1} - \overline Y_1\right).\end{aligned}$$_Applying the same argument to the_term with $X_{i,-S}$’s, and Lemma \[lemm: deg_U\], we conclude that $$\begin{aligned}
_ A_{1n} = \frac{1}{v^2_\sqrt{n}}\sum_{i \in_N} (\rho_{i,0}^A - \rho_{i, |
Observe that if a partially hyperbolic diffemorphism $f$ is $m s-$minimal (resp. $m u-$minimal), then $f$ belongs to $\mathcal{B}^s({\varepsilon},\delta)$ (resp. $\mathcal{B}^u({\varepsilon},\delta)$) for every ${\varepsilon}$ and $\delta$ positive. In particular, the rest of the proof is a directly consequence of the next lemma, Lemma \[l.aberto\], which implies $m$-minimality is a $G_{\delta}$ property.
\[l.aberto\] The subsets $\mathcal{B}^s({\varepsilon},\delta)$ and $\mathcal{B}^u({\varepsilon},\delta)$ are open subsets of $\mathcal{PH}^1_{m}(M)$.
Since $f$ belongs to $\mathcal{B}^s({\varepsilon},\delta)$ if, and only if, $f^{-1}$ belongs to $\mathcal{B}^u({\varepsilon},\delta)$, it is enough to prove that $\mathcal{B}^s({\varepsilon},\delta)$ is an open subset of $\mathcal{PH}^1_{m}(M)$.
Let $f\in \mathcal{B}^s({\varepsilon},\delta)$ be a partially hyperbolic diffeomorphism with decomposition $TM=E^s\oplus E^c\oplus E^u$.
By continuity of the partially hyperbolic splitting any diffeomorphism $g$ close enough to $f$ is also partially hyperbolic. We suppose first that every diffeomorphism $g$ close to $f$ has a partially hyperbolic splitting with stable and unstable bundle dimensions equal to the dimensions of the respectively sub-bundles in the partially hyperbolic splitting of $f$.
Now, given $x\in \mathcal{X}^s_{\delta}(f)$, there exists $K_x>0$ such that $\mathcal{F}^s_{K_x}(f,x)$ is $\delta$-dense in $M$, since $M$ is a compact manifold. Thus since the strong stable manifolds varies continuously with respect to the diffeomorphism in compact parts, there exists a neighborhood $\mathcal{V}_x$ of $f$ and a neighborhood $U_x$ of $x$, such that: $$\mathcal{F}^s_{K_x}(g,y) \text{ is } \delta-\text{dense in } M \text{ for every } y\in U_x \text{ and } g\in | Observe that if a partially hyperbolic diffemorphism $ f$ is $ m s-$minimal (resp. $ m u-$minimal), then $ f$ belongs to $ \mathcal{B}^s({\varepsilon},\delta)$ (resp. $ \mathcal{B}^u({\varepsilon},\delta)$) for every $ { \varepsilon}$ and $ \delta$ positive. In particular, the remainder of the validation is a directly consequence of the next lemma, Lemma \[l.aberto\ ], which entail $ m$-minimality is a $ G_{\delta}$ place.
\[l.aberto\ ] The subsets $ \mathcal{B}^s({\varepsilon},\delta)$ and $ \mathcal{B}^u({\varepsilon},\delta)$ are open subsets of $ \mathcal{PH}^1_{m}(M)$.
Since $ f$ belong to $ \mathcal{B}^s({\varepsilon},\delta)$ if, and only if, $ f^{-1}$ belongs to $ \mathcal{B}^u({\varepsilon},\delta)$, it is adequate to rise that $ \mathcal{B}^s({\varepsilon},\delta)$ is an open subset of $ \mathcal{PH}^1_{m}(M)$.
lease $ f\in \mathcal{B}^s({\varepsilon},\delta)$ be a partially hyperbolic diffeomorphism with decomposition $ TM = E^s\oplus E^c\oplus E^u$.
By continuity of the partially hyperbolic splitting any diffeomorphism $ g$ close enough to $ f$ is besides partially hyperbolic. We suppose first that every diffeomorphism $ g$ near to $ f$ has a partially hyperbolic splitting with static and precarious bundle dimensions equal to the proportion of the respectively sub - bundles in the partially hyperbolic splitting of $ f$.
Now, given $ x\in \mathcal{X}^s_{\delta}(f)$, there exist $ K_x>0 $ such that $ \mathcal{F}^s_{K_x}(f, x)$ is $ \delta$-dense in $ M$, since $ M$ is a compact manifold. Thus since the strong stable manifold varies continuously with respect to the diffeomorphism in compact parts, there exists a neighborhood $ \mathcal{V}_x$ of $ f$ and a neighborhood $ U_x$ of $ x$, such that: $ $ \mathcal{F}^s_{K_x}(g, y) \text { is } \delta-\text{dense in } M \text { for every } y\in U_x \text { and } g\in | Obsfrve that if a partially hyperbolic difywmorphmsm $f$ ia $m s-$minkmal (resp. $m u-$minimal), then $f$ uelobgs ti $\mathcal{B}^s({\varepsilon},\ddlta)$ (resp. $\mathcal{V}^u({\vacepsilon},\delta)$) foc every ${\ycrepsjpon}$ cnv $\delta$ positivg. In particunar, the rest ox ghz proof is a directly consequence of the nect lemma, Lemma \[l.wberuo\], rhicg implies $m$-minimality is a $G_{\delta}$ lropertj.
\[l.aberto\] The subsrts $\mathcal{B}^s({\varepsilon},\delha)$ ajd $\mathcal{B}^u({\varepsllon},\delta)$ ats o[wn subsets ow $\mathcal{PH}^1_{m}(M)$.
Since $f$ bglongs to $\mathcal{B}^s({\varepsilon},\delga)$ if, and only uf, $f^{-1}$ hglongs to $\mavhcal{B}^l({\varepsilon},\delta)$, it iv enougn to prove thab $\matical{V}^s({\varepsilon},\delta)$ is en open subset of $\majhcal{PH}^1_{m}(M)$.
Lat $f\in \mathcal{B}^s({\vaeepsilot},\delda)$ bd a oaruiaklg hypegbomic diffeojorphism wirh decomposition $TM=T^s\o[ous E^c\oplus E^h$.
By cogtynuity of the partially hyperbolic spliuting any diffeomorphism $g$ coose enough to $f$ is apso partiwlly hyperbolic. We suppose first that every diffemmorpiirm $y$ closd tl $f$ has a partially hyperbolic splitting with atsbke and unstablc bundle dimensionx fqiwl to the dimgnsions of the respectively dub-bundjes ib the paruiallu hyperbolic splitting of $f$.
Bow, given $x\iu \mqthcal{X}^s_{\delta}(f)$, thexe exists $K_x>0$ sucn thay $\mathcal{F}^s_{K_x}(f,x)$ is $\deltc$-dense in $M$, since $M$ is a ckopact manifold. Tfus shnce the strong stable manyfolds vacies eontinuojsly with wespect to the diffeomorphism in clmpacj partv, there exlsts a neighborhood $\mathcal{V}_x$ oh $f$ and a neibhtorvood $U_x$ jf $x$, xuch that: $$\matrcal{F}^s_{K_x}(g,y) \texj{ is } \delca-\text{aense in } J \text{ hor every } y\yn U_x \text{ ang } g\in | Observe that if a partially hyperbolic diffemorphism $m (resp. $m then $f$ belongs every and $\delta$ positive. particular, the rest the proof is a directly consequence the next lemma, Lemma \[l.aberto\], which implies $m$-minimality is a $G_{\delta}$ property. \[l.aberto\] subsets $\mathcal{B}^s({\varepsilon},\delta)$ and $\mathcal{B}^u({\varepsilon},\delta)$ are open subsets of $\mathcal{PH}^1_{m}(M)$. Since $f$ belongs to if, only $f^{-1}$ to $\mathcal{B}^u({\varepsilon},\delta)$, it is enough to prove that $\mathcal{B}^s({\varepsilon},\delta)$ is an open subset of $\mathcal{PH}^1_{m}(M)$. Let \mathcal{B}^s({\varepsilon},\delta)$ be a partially hyperbolic diffeomorphism with decomposition E^c\oplus E^u$. By continuity the partially hyperbolic splitting any $g$ enough to is partially We suppose first every diffeomorphism $g$ close to $f$ has a partially hyperbolic splitting with stable and unstable bundle dimensions to the the respectively in partially splitting of $f$. $x\in \mathcal{X}^s_{\delta}(f)$, there exists $K_x>0$ such $\delta$-dense in $M$, since $M$ is a compact Thus since strong stable manifolds varies continuously with to the diffeomorphism in compact parts, there exists neighborhood $\mathcal{V}_x$ of $f$ and a neighborhood $U_x$ of $x$, such that: $$\mathcal{F}^s_{K_x}(g,y) \text{ is in } M \text{ every } y\in \text{ } | Observe that if a partially hyPerbolic diFfemoRphIsm $F$ iS $m s-$mInimAl (resp. $m u-$minimaL), Then $F$ belongs to $\mathcal{B}^s({\varEpsilOn},\DElta)$ (REsP. $\mathCal{B}^u({\vaREpSILon},\DeLtA)$) foR eVErY ${\varePsiLon}$ and $\dElta$ positiVe. IN pArticular, the REsT of the prooF is A directly conSeqUence oF tHe nEXt lemMa, LEmma \[l.Aberto\], WHich imPlies $m$-minImALity is A $g_{\delta}$ pROPeRty.
\[l.Aberto\] The subsets $\mAThCAl{B}^s({\varepsilon},\Delta)$ aNd $\MAtHCAl{B}^U({\vaRepsilon},\deLtA)$ are oPEn subseTS oF $\MAThcAL{PH}^1_{m}(M)$.
Since $f$ beLongs to $\mathCAl{B}^S({\varepSiLon},\DElta)$ if, And onLy IF, $f^{-1}$ bElongs to $\matHcal{b}^u({\varepsiLon},\delTA)$, it is enOUgh to prOve thaT $\maThcAl{B}^s({\VArEpSilOn},\DEltA)$ Is An oPEn sUbset of $\mAtHcAl{PH}^1_{m}(m)$.
Let $F\IN \MAthcAl{B}^S({\varEpsilOn},\delta)$ be a parTiaLly hYPerBolic DiffeOmorPhIsm wiTh decoMposiTiOn $TM=E^s\oplus E^c\opLus E^U$.
By continUitY oF thE pArtiaLLy hypeRboLic SplittiNg any diFFeoMoRPHIsM $g$ close enough to $f$ is AlSO PaRtially hYperboLIc. we SUppose fiRsT thAt evERY diffEomoRPhIsm $g$ closE to $f$ haS A pArTially hYpErboliC sPliTtiNg witH StabLe and uNstable bUndle DImensions equal TO the dimensionS Of THE rESpecTivEly sub-bundlEs in THe paRtiaLLy HypERboliC spliTtINg OF $f$.
Now, given $x\in \mathcaL{X}^S_{\delta}(F)$, therE exists $K_x>0$ such That $\mathcaL{f}^S_{k_x}(f,x)$ is $\deLta$-dENsE In $M$, since $M$ is a coMpact Manifold. ThUS since thE stroNg stable Manifolds VARies contInuOusLy wIth RESpEct to the diffeOMOrphIsM in compAct Parts, thEre ExiSts A neIgHborhood $\mAthcal{V}_x$ Of $F$ aNd A nEigHborhOOd $U_x$ of $x$, sUcH thAt: $$\MatHcal{F}^S_{k_x}(g,y) \teXt{ is } \dElta-\TeXt{DEnsE in } M \texT{ FoR EVery } Y\iN U_X \texT{ anD } g\In | Observe that if a partiall y hyperbol ic di ffe mor ph ism$f$is $m s-$minim a l (r esp. $m u-$minimal), t hen $ f$ belo n gs to $ \mathca l {B } ^ s({ \v ar eps il o n} ,\del ta) $ (resp . $\mathca l{B }^ u({\varepsil o n} ,\delta)$) fo r every ${\v are psilon }$ an d $\de lta $ pos itive. In par ticular,th e resto f the p r o of isa directly conseq u en c e of the nextlemma, L e mm a \[l .ab erto\], wh ic h imp l ies $m$ - mi n i m ali t y is a $G_{\d elta}$ prop e rty .
\[l .a ber t o\] Th e sub se t s $ \mathcal{B} ^s({ \varepsil on},\d e lta)$ a n d $\mat hcal{B }^u ({\ vare p si lo n}, \d e lta ) $are ope n subset sof $\ma thca l { P H }^1_ {m} (M)$ .
Si nce $f$ belon gsto $ \ mat hcal{ B}^s( {\va re psilo n},\de lta)$ i f, and only if, $f^ {-1}$ bel ong sto$\ mathc a l{B}^u ({\ var epsilon },\delt a )$, i t i senough to prove th at $ \m athcal{B }^s({\ v ar ep s ilon},\d el ta) $ is a n ope n su b se t of $\m athcal { PH }^ 1_{m}(M )$ .
Let $ f\i n \math c al{B }^s({\ varepsil on},\ d elta)$ be a pa r tially hyperb o li c di f feom orp hism with d ecom p osit ion$ TM =E^ s \oplu s E^c \o p lu s E^u$.
By continui ty of th e par tially hyperb olic split t i n g any di ffeo m or p hism $g$ close enou gh to $f$i s also p artia lly hype rbolic. W e supposefir sttha t e v e ry diffeomorphi s m $g$ c lose to $f $ has a pa rti all y h yp erbolic s plitting w it hst abl e and unstable b und le di mensi o ns equ al to the d im e nsi ons oft he r espe ct iv elysub -b undle s in the partia lly hyper bol i c sp li tt ing of$f$.
Now, gi ve n $x\in \m at hca l{X}^s _ { \delta}( f)$, there exists $K_x> 0 $ suchtha t $\m athc al{F}^s_{ K_x }(f,x) $ i s $\del ta$-de nse i n$M$ , since $ M$ is a compact m a n ifo ld. T hu s si nce the strong stable man i fol ds varies con tin uous l y w ith re s pec tt o t h e diffeomorphism in compac tp ar ts, theree xis ts a neig hborhoo d $\m a thcal{V }_x$ of $ f$ and ane ighb o r hoo d $U_x$ of $x$, su ch that:$ $\mat h ca l{F}^ s_{ K_x}(g ,y ) \ text{ is }\ del ta-\t ext{de ns e in } M \t ex t{ for e very } y\in U_x \text{and }g\in | Observe that_if a_partially hyperbolic diffemorphism $f$_is $m_s-$minimal_(resp. $m_u-$minimal),_then $f$ belongs_to $\mathcal{B}^s({\varepsilon},\delta)$ (resp._$\mathcal{B}^u({\varepsilon},\delta)$) for every ${\varepsilon}$_and $\delta$ positive._In_particular, the rest of the proof is a directly consequence of the next lemma,_Lemma_\[l.aberto\], which_implies_$m$-minimality_is a $G_{\delta}$ property.
\[l.aberto\] The_subsets $\mathcal{B}^s({\varepsilon},\delta)$ and $\mathcal{B}^u({\varepsilon},\delta)$ are_open subsets_of $\mathcal{PH}^1_{m}(M)$.
Since $f$ belongs to $\mathcal{B}^s({\varepsilon},\delta)$ if, and_only_if, $f^{-1}$ belongs_to $\mathcal{B}^u({\varepsilon},\delta)$, it is enough to prove that $\mathcal{B}^s({\varepsilon},\delta)$_is an open subset of $\mathcal{PH}^1_{m}(M)$.
Let_$f\in \mathcal{B}^s({\varepsilon},\delta)$_be_a_partially hyperbolic diffeomorphism with_decomposition $TM=E^s\oplus E^c\oplus E^u$.
By continuity of_the partially hyperbolic splitting any diffeomorphism_$g$ close enough to $f$ is also_partially hyperbolic. We suppose first that_every diffeomorphism $g$ close to_$f$ has_a partially hyperbolic splitting with_stable and unstable_bundle dimensions_equal to the_dimensions of the respectively sub-bundles in_the partially hyperbolic_splitting of $f$.
Now, given $x\in \mathcal{X}^s_{\delta}(f)$,_there_exists $K_x>0$ such_that_$\mathcal{F}^s_{K_x}(f,x)$_is $\delta$-dense_in $M$, since_$M$_is a_compact_manifold. Thus since the strong stable_manifolds_varies continuously with respect to the diffeomorphism_in compact parts, there_exists_a neighborhood $\mathcal{V}_x$ of_$f$ and a neighborhood $U_x$_of $x$, such that: $$\mathcal{F}^s_{K_x}(g,y) \text{_is }_\delta-\text{dense in_} M \text{ for every } y\in U_x \text{ and }_g\in |
bar{\Omega} \left(3 (\gamma \bar{\Omega}_{m}+\Omega -1)+2 \bar{\Sigma}^2\right), \\
& \partial_{\tau}{\bar{\Omega}}_m= \bar{\Omega}_{m} \left(3 \gamma (\bar{\Omega}_{m}-1)+2 \bar{\Sigma}^2+3 \bar{\Omega} \right), \\
& \partial_{\tau}{\bar{\Sigma}}= \frac{1}{2} \bar{\Sigma} \left(3 (\gamma \bar{\Omega}_{m}+\bar{\Omega} -2)+2 \bar{\Sigma} ^2\right).\end{aligned}$$
Analyzing the equilibrium points of the system, which is the same as, we obtain points of the form $P_1 = (0,1,0)$, $P_2 = (1,0,0)$, $ P_3 =(0,0, - \sqrt{3})$, $ P_4 = (0,0, \sqrt{3}) $ and $P_5 = (0,0,0)$. By evaluating the points in the linearization of and calculating their respective eigenvalues, we obtain the results shown in the table \[PCsis3.31\]. This means that on average the dynamics of the models and is the same which indicates that the asymptotic behavior of the model on average is independent of the coupling function, of the geometry of model (FLRW, Bianchi I). Although, obviously, non-averaged systems have different dynamics.
Averaging methods for the scalar field with generalized harmonic potential $
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\frac{\phi }{f}\right)\right]
$, $f> 0$ nonminimally coupled to matter with coupling $\chi(\phi)=\chi_0 e^{\frac{\lambda \phi}{4-3\gamma}}$. {#SECT:4.5}
==============================================================================================================
In this section the averaging methods are applied for FLRW metrics and Bianchi I metrics for the generalized harmonic potential $
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\frac{\phi }{f}\right)\right]
$, $f> 0$, coupled to matter with coupling function $\chi=\chi_0 e^{\frac | bar{\Omega } \left(3 (\gamma \bar{\Omega}_{m}+\Omega -1)+2 \bar{\Sigma}^2\right), \\
& \partial_{\tau}{\bar{\Omega}}_m= \bar{\Omega}_{m } \left(3 \gamma (\bar{\Omega}_{m}-1)+2 \bar{\Sigma}^2 + 3 \bar{\Omega } \right), \\
& \partial_{\tau}{\bar{\Sigma}}= \frac{1}{2 } \bar{\Sigma } \left(3 (\gamma \bar{\Omega}_{m}+\bar{\Omega } -2)+2 \bar{\Sigma } ^2\right).\end{aligned}$$
Analyzing the equilibrium points of the system, which is the like as, we receive points of the form $ P_1 = (0,1,0)$, $ P_2 = (1,0,0)$, $ P_3 =( 0,0, - \sqrt{3})$, $ P_4 = (0,0, \sqrt{3 }) $ and $ P_5 = (0,0,0)$. By measure the point in the linearization of and calculating their respective eigenvalues, we prevail the resultant role shown in the mesa \[PCsis3.31\ ]. This means that on average the moral force of the models and is the same which indicates that the asymptotic behavior of the exemplar on average is independent of the coupling function, of the geometry of model (FLRW, Bianchi I). Although, obviously, non - average systems have different moral force.
Averaging methods for the scalar plain with generalized harmonic electric potential $
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\frac{\phi } { f}\right)\right ]
$, $ f > 0 $ nonminimally coupled to count with coupling $ \chi(\phi)=\chi_0 e^{\frac{\lambda \phi}{4 - 3\gamma}}$. { # SECT:4.5 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
In this section the averaging methods are applied for FLRW metric function and Bianchi I metrics for the generalized harmonic potential $
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\frac{\phi } { f}\right)\right ]
$, $ f > 0 $, coupled to matter with yoke affair $ \chi=\chi_0 e^{\frac | bar{\Lmega} \left(3 (\gamma \bar{\Ooega}_{m}+\Omega -1)+2 \bar{\Sigma}^2\rijht), \\
& \padtial_{\tau}{\car{\Omega}}_m= \bar{\Omega}_{m} \left(3 \galmq (\bae{\Omega}_{m}-1)+2 \bar{\Sigma}^2+3 \bar{\Omdga} \right), \\
& \partiao_{\tau}{\uar{\Sigma}}= \frac{1}{2} \bac{\Aigma} \left(3 (\gzlma \uar{\Omega}_{m}+\bar{\Omena} -2)+2 \bar{\Sigmd} ^2\right).\end{aligtea}$$
Aualyzing the equilibrium points of tre systrm, which is the fame ws, ws obtain points of the form $P_1 = (0,1,0)$, $P_2 = (1,0,0)$, $ P_3 =(0,0, - \vqrt{3})$, $ P_4 = (0,0, \sqry{3}) $ and $P_5 = (0,0,0)$. By evaluating tje plints in the lineagization of and xalculating gheir resptccive eigenvzlues, we obtain the results shodn in the table \[PXsid3.31\]. This means that on average bne dyndmics og the models akd is thw same which indicatev that the asymptojic behavimr of the model on qvwrage is hndeownddnt oh tge couoliig function, of the geonetry of model (FLRW, Byqnchi I). Althohgh, obdijusly, non-averaged systems have differenu dynzmics.
Averaging methods dor the scalar field aith genewalized harmonic potential $
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\hrxc{\pkl }{w}\eihht)\right]
$, $f> 0$ nonminimally coupled to matter wieg volpling $\chi(\phi)=\chi_0 c^{\frac{\lambda \phi}{4-3\gamka}}$. {#SRST:4.5}
==============================================================================================================
In this secjion thz aberaging methods age applyed fir FLRW mttricx and Bianchi I metrics for the generalpzed harmonic potentiap $
V(\phi)= \frac{\'hi ^2}{2}+f\lgft[1- \cox \left(\frac{\phi }{f}\right)\rigkt]
$, $f> 0$, doupled to latter wiff coupling functkon $\cvi=\chi_0 e^{\frac | bar{\Omega} \left(3 (\gamma \bar{\Omega}_{m}+\Omega -1)+2 \bar{\Sigma}^2\right), \\ \bar{\Omega}_{m} \gamma (\bar{\Omega}_{m}-1)+2 \bar{\Omega} \right), \\ (\gamma -2)+2 \bar{\Sigma} ^2\right).\end{aligned}$$ the equilibrium points the system, which is the same we obtain points of the form $P_1 = (0,1,0)$, $P_2 = (1,0,0)$, $ =(0,0, - \sqrt{3})$, $ P_4 = (0,0, \sqrt{3}) $ and $P_5 = (0,0,0)$. evaluating points the of and calculating their respective eigenvalues, we obtain the results shown in the table \[PCsis3.31\]. This that on average the dynamics of the models is the same which that the asymptotic behavior of model average is of coupling of the geometry model (FLRW, Bianchi I). Although, obviously, non-averaged systems have different dynamics. Averaging methods for the scalar field generalized harmonic V(\phi)= \frac{\phi \cos }{f}\right)\right] $f> 0$ nonminimally matter with coupling $\chi(\phi)=\chi_0 e^{\frac{\lambda \phi}{4-3\gamma}}$. this section the averaging methods are applied for metrics and I metrics for the generalized harmonic $ V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos \left(\frac{\phi }{f}\right)\right] $, 0$, coupled to matter with coupling function $\chi=\chi_0 e^{\frac | bar{\Omega} \left(3 (\gamma \bar{\Omega}_{M}+\Omega -1)+2 \bar{\SIgma}^2\rIghT), \\
& \paRtIal_{\tAu}{\baR{\Omega}}_m= \bar{\OmegA}_{M} \lefT(3 \gamma (\bar{\Omega}_{m}-1)+2 \bar{\SigmA}^2+3 \bar{\OMeGA} \rigHT), \\
& \pArtiaL_{\tau}{\bar{\sIgMA}}= \FraC{1}{2} \bAr{\sigMa} \LEfT(3 (\gammA \baR{\Omega}_{m}+\Bar{\Omega} -2)+2 \baR{\SiGmA} ^2\right).\end{aliGNeD}$$
Analyzing The Equilibrium pOinTs of thE sYstEM, whicH is The saMe as, we OBtain pOints of thE fORm $P_1 = (0,1,0)$, $P_2 = (1,0,0)$, $ P_3 =(0,0, - \sQRt{3})$, $ P_4 = (0,0, \sqrt{3}) $ AND $P_5 = (0,0,0)$. by evAluating the points IN tHE linearization Of and cAlCUlATIng TheIr respectiVe EigenVAlues, we OBtAIN The REsults shown in The table \[PCsIS3.31\]. ThIs meanS tHat ON averaGe the DyNAmiCs of the modeLs anD is the samE which INdicateS That the AsymptOtiC beHaviOR oF tHe mOdEL on AVeRagE Is iNdependeNt Of The coUpliNG FUNctiOn, oF the GeomeTry of model (FLRw, BiAnchI i). AlThougH, obviOuslY, nOn-aveRaged sYstemS hAve different dynAmicS.
AveraginG meThOds FoR the sCAlar fiEld WitH generaLized haRMonIc POTEnTial $
V(\phi)= \frac{\phi ^2}{2}+f\leFt[1- \COS \lEft(\frac{\pHi }{f}\rigHT)\rIgHT]
$, $f> 0$ nonminImAllY couPLEd to mAtteR WiTh coupliNg $\chi(\pHI)=\cHi_0 E^{\frac{\laMbDa \phi}{4-3\gAmMa}}$. {#SeCT:4.5}
==============================================================================================================
in thiS SectIon the AveraginG methODs are applied foR fLRW metrics anD biANChI i metRicS for the geneRaliZEd haRmonIC pOteNTial $
V(\Phi)= \frAc{\PHi ^2}{2}+F\Left[1- \cos \left(\frac{\phi }{f}\RiGht)\rigHt]
$, $f> 0$, coUpled to matter With coupliNG FUnction $\cHi=\chI_0 E^{\fRAc | bar{\Omega} \left(3 (\ga mma \bar{ \Omeg a}_ {m} +\ Omeg a -1 )+2 \bar{\Sigm a }^2\ right), \\
& \partial_ {\tau }{ \ bar{ \ Om ega}} _m= \ba r {\ O m ega }_ {m } \l e ft (3 \g amm a (\ba r{\Omega}_ {m} -1 )+2 \bar{\Si g ma }^2+3 \bar {\O mega} \right ),\\
& \ pa rti a l_{\t au} {\bar {\Sigm a }}= \f rac{1}{2} \ b ar{\Si g ma} \l e f t( 3 (\ gamma \bar{\Omeg a }_ { m}+\bar{\Omega } -2)+ 2\ ba r { \Si gma } ^2\right ). \end{ a ligned} $ $A n aly z ing the equil ibrium poin t s o f thesy ste m , whic h isth e sa me as, we o btai n pointsof the form $P _ 1 = (0, 1,0)$, $P _2= (1 , 0, 0) $,$P _3= (0 ,0, - \ sqrt{3}) $, $ P_4= (0 , 0 , \sqr t{3 }) $ and$P_5 = (0,0,0 )$. Bye val uatin g the poi nt s inthe li neari za tion of and cal cula ting thei r r es pec ti ve ei g envalu es, we obtain the re s ult ss h o wn in the table \[PC si s 3 .3 1\]. Thi s mean s t ha t on aver ag e t he d y n amics oft he modelsand is th esame wh ic h indi ca tes th at th e asy mptoti c behavi or of the model on a v erage is inde p en d e nt of t hecoupling fu ncti o n, o f th e g eom e try o f mod el (F L RW, Bianchi I). Alt ho ugh, o bviou sly, non-aver aged syste m s have dif fere n td ynamics.
Aver aging methods f o r the sc alarfield wi th genera l i zed harm oni c p ote nti a l $
V(\phi)= \fr a c {\ph i^2}{2}+ f\l eft[1-\co s \ lef t(\ fr ac{\phi } {f}\righ t) \r ig ht ]
$ , $f> 0$ nonmi ni mal ly co upled to mat ter w ithco up l ing $\chi( \ ph i ) =\ch i_ 0e^{\ fra c{ \lamb da \ p hi} {4-3\ga mma}}$. { #SE C T:4. 5}
= ======= ============= == ========== == === ====== = = ======== ======================= = ======= === ===== ==== ========= === =
Inthi s secti on the aver ag ing m ethod s ar e a pp lied for F L R W m etric sandBianchi I metrics for the gen eralized harm oni c po t e nt ial $V (\p hi ) = \ f r ac{\phi ^2}{2}+ f\left[1-\c o s\left(\fra c {\p hi }{f}\r ight)\r ight] $, $f>0$, coupl ed to mat te r wi t h co upling fun ction $\ chi=\chi_ 0 e^{\ f ra c | bar{\Omega} _ \left(3_(\gamma \bar{\Omega}_{m}+\Omega -1)+2_\bar{\Sigma}^2\right), \\
&_\partial_{\tau}{\bar{\Omega}}_m=_\bar{\Omega}_{m} _\left(3_\gamma (\bar{\Omega}_{m}-1)+2_\bar{\Sigma}^2+3 \bar{\Omega} \right),_\\
& \partial_{\tau}{\bar{\Sigma}}= \frac{1}{2} \bar{\Sigma}_ \left(3 (\gamma__\bar{\Omega}_{m}+\bar{\Omega} -2)+2 \bar{\Sigma} ^2\right).\end{aligned}$$
Analyzing the equilibrium points of the system, which is the same_as,_we obtain_points_of_the form $P_1 = (0,1,0)$,_$P_2 = (1,0,0)$, $ P_3_=(0,0, -_\sqrt{3})$, $ P_4 = (0,0, \sqrt{3}) $ and_$P_5_= (0,0,0)$. By_evaluating the points in the linearization of and calculating_their respective eigenvalues, we obtain the_results shown in_the_table_\[PCsis3.31\]. This means that_on average the dynamics of the_models and is the same which_indicates that the asymptotic behavior of the_model on average is independent of_the coupling function, of the_geometry of_model (FLRW, Bianchi I). Although,_obviously, non-averaged systems_have different_dynamics.
Averaging methods for_the scalar field with generalized harmonic_potential $
V(\phi)= \frac{\phi_^2}{2}+f\left[1- \cos \left(\frac{\phi }{f}\right)\right]
$, $f> 0$_nonminimally_coupled to matter_with_coupling_$\chi(\phi)=\chi_0 e^{\frac{\lambda_\phi}{4-3\gamma}}$. {#SECT:4.5}
==============================================================================================================
In this_section_the averaging_methods_are applied for FLRW metrics and_Bianchi_I metrics for the generalized harmonic potential_$
V(\phi)= \frac{\phi ^2}{2}+f\left[1- \cos_\left(\frac{\phi_}{f}\right)\right]
$, $f> 0$, coupled_to matter with coupling function_$\chi=\chi_0 e^{\frac |
G$-cells, there are only finitely many conjugacy classes of finite subgroups of $G$ that occur as isotropy groups of points of $X$. In particular, the isotropy groups of $X$ have bounded order. Since $G$ is discrete, Corollary 2.7 of [@luck-oliver:completion] thus provides a hermitian $G$-vector bundle $\zeta$ over $X$ such that for every point $x\in X$, the fiber $\zeta_x$ is a multiple of the regular representation of the isotropy group $G_x$. In particular, the $G_x$-action on $\zeta_x$ is faithful, and hence $\operatorname{Sym}(\zeta_x)$ is a complete complex $G_x$-universe, compare [@schwede:global Rk.6.3.22]. So the bundle $\zeta$ is ample.
\[prop:ample bundle\] Let $G$ be a discrete group and $X$ a finite proper $G$-CW-complex. Then the map ${\langle -\rangle}: \operatorname{Vect}_G(X){\longrightarrow}{\mathbf{ku}}_G{\llbracket X\rrbracket}$ defined in is additive.
The sum in the group ${\mathbf{ku}}_G{\llbracket X\rrbracket}$ is defined by addition via a fiberwise pinch map. We will relate the pinch sum to a different binary operation, the ‘bundle sum’, by an Eckmann-Hilton argument.
Proposition \[prop:universal bundle\] (ii) provides an ample hermitian $G$-vector bundle $\zeta$ over $X$. We let $\mu=u\zeta$ denote the underlying euclidean vector bundle. Because $\zeta$ embeds into $(u\zeta)_{\mathbb C}=\mu_{\mathbb C}$, the hermitian $G$-vector bundle $\mu_{\mathbb C}$ is also ample. By adding a trivial complex line bundle to $\zeta$, if necessary, we can moreover assume that there exists a $G$-equivariant linear isometric embedding $j:X\times \underline{{\mathbb R}}{\longrightarrow}\mu=u\zeta$ of the trivial ${\mathbb R}$-line bundle over $X$. The embedding $j$ parameterizes a trivial 1-dimensional summand in $\mu$, and hence a pinch map $p:S^\mu{\longrightarrow}S^\mu\vee_X S^\mu$. The ‘pinch sum’ | G$-cells, there are only finitely many conjugacy classes of finite subgroups of $ G$ that happen as isotropy group of points of $ X$. In particular, the isotropy group of $ X$ have bound order. Since $ G$ is discrete, Corollary 2.7 of [ @luck - oliver: completion ] thus put up a hermitian $ G$-vector pile $ \zeta$ over $ X$ such that for every point $ x\in X$, the roughage $ \zeta_x$ is a multiple of the regular representation of the isotropy group $ G_x$. In finical, the $ G_x$-action on $ \zeta_x$ is faithful, and hence $ \operatorname{Sym}(\zeta_x)$ is a arrant complex $ G_x$-universe, compare [ @schwede: global Rk.6.3.22 ]. therefore the bundle $ \zeta$ is ample.
\[prop: ample bundle\ ] Let $ G$ be a discrete group and $ X$ a finite proper $ G$-CW - complex. Then the map $ { \langle -\rangle }: \operatorname{Vect}_G(X){\longrightarrow}{\mathbf{ku}}_G{\llbracket X\rrbracket}$ defined in is additive.
The kernel in the group $ { \mathbf{ku}}_G{\llbracket X\rrbracket}$ is defined by addition via a fiberwise pinch function. We will relate the pinch kernel to a different binary process, the ‘ bundle sum ’, by an Eckmann - Hilton argument.
Proposition \[prop: universal bundle\ ] (ii) provides an ample hermitian $ G$-vector bundle $ \zeta$ over $ X$. We lease $ \mu = u\zeta$ denote the underlying euclidean vector bundle. Because $ \zeta$ embeds into $ (u\zeta)_{\mathbb C}=\mu_{\mathbb C}$, the hermitian $ G$-vector bundle $ \mu_{\mathbb C}$ is also ample. By adding a trivial complex line bundle to $ \zeta$, if necessary, we can moreover assume that there exists a $ G$-equivariant linear isometric embedding $ j: X\times \underline{{\mathbb R}}{\longrightarrow}\mu = u\zeta$ of the fiddling $ { \mathbb R}$-line bundle over $ X$. The embedding $ j$ parameterizes a superficial 1 - dimensional summand in $ \mu$, and therefore a apprehension map $ p: S^\mu{\longrightarrow}S^\mu\vee_X S^\mu$. The ‘ pinch kernel ’ | G$-cepls, there are only finittly many conjugaci xlassev of fjnite sucgroups of $G$ that occur as idorropy groups of points of $X$. In partibular, the isouropy groups of $X$ have bomuded kvder. Vmnce $G$ is discrgte, Corollarf 2.7 of [@luck-olivar:zolpletion] thus provides a hermitian $D$-vector bkndle $\zeta$ ovet $X$ slcr thzn nor every point $x\in X$, the fiber $\aeta_x$ iv a multiple pf the regular representatlon lf the isotropy grlup $G_x$. In pqrtisylar, the $G_x$-aztion on $\zttc_x$ is faithrul, and hence $\operatorname{Sym}(\zega_x)$ ix a complejz colklex $G_x$-univecse, coipare [@schwedc:blobal Rk.6.3.22]. So yhe bundle $\zets$ iv anple.
\[prop:ample bundle\] Net $G$ be a discretg group ang $R$ a finite proper $G$-CW-xonplex. Thet thd mao ${\lznjle -\ranglf}: \o'eratorname{Bect}_G(X){\longrughtarrow}{\mathbf{ku}}_G{\lkbwqcket X\rrbraciet}$ desiged in is additive.
The sum in the group ${\katgbf{ku}}_G{\llbracket X\rrbracjet}$ is defined by addltion via a fiberwise pinch map. We will relate the pinch sgm to x dnnnerevr hinary operation, the ‘bundle sum’, by an Eckmann-Rjluon argument.
Propofition \[prop:inlvrtsal bundle\] (ii) provibss an ample hermitiaj $G$-vectjr bubdle $\zeta$ ovet $X$. We let $\mu=u\zeta$ denote tye underlyiny eyclidean vector buudle. Because $\zets$ embrds into $(u\zeta)_{\mathbb C}=\mb_{\mathbg C}$, the herlitian $G$-vsztor bundle $\mu_{\maghbn C}$ is also ample. By adding a trivial romplzx line cundke to $\seta$, if nefessavf, we can moreover wssumg that there exidts a $G$-equivariant linear isomevcic embedding $j:F\tikes \undexline{{\msthbb R}}{\longridhtarrow}\mu=u\zetc$ of the trivkal ${\mathbb R}$-line uundle over $V$. The embeddith $j$ parametecizes a twiviql 1-dumensiovxl summand in $\ku$, and heuee a pincy map $p:S^\mu{\longrighbarrod}A^\mu\vee_X S^\mu$. The ‘pibch sum’ | G$-cells, there are only finitely many conjugacy finite of $G$ occur as isotropy In the isotropy groups $X$ have bounded Since $G$ is discrete, Corollary 2.7 [@luck-oliver:completion] thus provides a hermitian $G$-vector bundle $\zeta$ over $X$ such that for point $x\in X$, the fiber $\zeta_x$ is a multiple of the regular representation the group In the $G_x$-action on $\zeta_x$ is faithful, and hence $\operatorname{Sym}(\zeta_x)$ is a complete complex $G_x$-universe, compare [@schwede:global So the bundle $\zeta$ is ample. \[prop:ample bundle\] $G$ be a discrete and $X$ a finite proper Then map ${\langle \operatorname{Vect}_G(X){\longrightarrow}{\mathbf{ku}}_G{\llbracket defined is additive. The in the group ${\mathbf{ku}}_G{\llbracket X\rrbracket}$ is defined by addition via a fiberwise pinch map. We will relate pinch sum different binary the sum’, an Eckmann-Hilton argument. bundle\] (ii) provides an ample hermitian over $X$. We let $\mu=u\zeta$ denote the underlying vector bundle. $\zeta$ embeds into $(u\zeta)_{\mathbb C}=\mu_{\mathbb C}$, hermitian $G$-vector bundle $\mu_{\mathbb C}$ is also ample. adding a trivial complex line bundle to $\zeta$, if necessary, we can moreover assume that a $G$-equivariant linear isometric $j:X\times \underline{{\mathbb R}}{\longrightarrow}\mu=u\zeta$ the ${\mathbb bundle $X$. The $j$ parameterizes a trivial 1-dimensional summand in $\mu$, and hence a map $p:S^\mu{\longrightarrow}S^\mu\vee_X S^\mu$. The ‘pinch sum’ | G$-cells, there are only finitelY many conjuGacy cLasSes Of FiniTe suBgroups of $G$ that OCcur As isotropy groups of poinTs of $X$. in PArtiCUlAr, the IsotropY GrOUPs oF $X$ HaVe bOuNDeD ordeR. SiNce $G$ is dIscrete, CorOllArY 2.7 of [@luck-oliveR:CoMpletion] thUs pRovides a hermItiAn $G$-vecToR buNDle $\zeTa$ oVer $X$ sUch thaT For eveRy point $x\iN X$, THe fibeR $\Zeta_x$ is A MUlTiplE of the regular reprESeNTation of the isoTropy gRoUP $G_X$. iN paRtiCular, the $G_x$-AcTion oN $\Zeta_x$ is FAiTHFUl, aND hence $\operatoRname{Sym}(\zetA_X)$ is A complEtE coMPlex $G_x$-UniveRsE, ComPare [@schwede:GlobAl Rk.6.3.22]. So the Bundle $\ZEta$ is amPLe.
\[prop:aMple buNdlE\] LeT $G$ be A DiScRetE gROup ANd $x$ a fINitE proper $G$-cW-CoMplex. then THE MAp ${\laNglE -\ranGle}: \opEratorname{VecT}_G(X){\LongRIghTarroW}{\mathBf{ku}}_g{\lLbracKet X\rrBrackEt}$ Defined in is addiTive.
the sum in tHe gRoUp ${\mAtHbf{ku}}_g{\LlbracKet x\rrBracket}$ Is definED by AdDITIoN via a fiberwise pincH mAP. we Will relaTe the pINcH sUM to a diffErEnt BinaRY OperaTion, THe ‘Bundle suM’, by an ECKmAnN-Hilton ArGument.
prOpoSitIon \[prOP:uniVersal Bundle\] (ii) ProviDEs an ample hermiTIan $G$-vector bunDLe $\ZETa$ OVer $X$. we lEt $\mu=u\zeta$ deNote THe unDerlYInG euCLideaN vectOr BUnDLe. Because $\zeta$ embeds InTo $(u\zetA)_{\mathBb C}=\mu_{\mathbb C}$, tHe hermitiaN $g$-VEctor bunDle $\mU_{\MaTHbb C}$ is also amplE. By adDing a triviAL complex Line bUndle to $\zEta$, if neceSSAry, we can MorEovEr aSsuME ThAt there exists A $g$-EquiVaRiant liNeaR isometRic EmbEddIng $J:X\Times \undeRline{{\matHbB R}}{\LoNgRigHtarrOW}\mu=u\zeta$ Of The TrIviAl ${\matHBb R}$-linE bundLe ovEr $x$. THE emBedding $J$ PaRAMeteRiZeS a trIviAl 1-DimenSionAL suMmand in $\Mu$, and hencE a pINch mAp $P:S^\Mu{\longrIghtarrow}S^\mu\vEe_x S^\mu$. The ‘pinCh Sum’ | G$-cells, there are only f initely ma ny co nju gac yclas sesof finite subg r oups of $G$ that occur asisotr op y gro u ps of p oints o f $ X $ . I npa rti cu l ar , the is otropygroups of$X$ h ave boundedo rd er. Since$G$ is discrete , C orolla ry 2. 7 of [ @lu ck-ol iver:c o mpleti on] thuspr o videsa hermit i a n$G$- vector bundle $\z e ta $ over $X$ such thatfo r e v e rypoi nt $x\in X $, thef iber $\ z et a _ x $ i s a multiple o f the regul a r r eprese nt ati o n of t he is ot r opy group $G_x $. I n particu lar, t h e $G_x$ - actionon $\z eta _x$ isf ai th ful ,a ndh en ce$ \op eratorna me {S ym}(\ zeta _ x ) $ isa c ompl ete c omplex $G_x$- uni vers e , c ompar e [@s chwe de :glob al Rk. 6.3.2 2] . So the bundle $\z eta$ is a mpl e.
\ [p rop:a m ple bu ndl e\] Let $G $ be ad isc re t e gr oup and $X$ a fini te p ro per $G$- CW-com p le x. Then the m ap${\l a n gle - \ran g le }: \oper atorna m e{ Ve ct}_G(X ){ \longr ig hta rro w}{\m a thbf {ku}}_ G{\llbra cketX \rrbracket}$ d e fined in is a d di t i ve .
Th e s um in the g roup ${\m athb f {k u}} _ G{\ll brack et X\ r rbracket}$ is defin ed by ad ditio n via a fiber wise pinch m a p. We wi ll r e la t e the pinch su m toa differen t binaryopera tion, th e ‘bundle s um’, byanEck man n-H i l to n argument.
P r opos it ion \[p rop :univer sal bu ndl e\] ( ii) provi des an a mp le h er mit ian $ G $-vector b und le $\ zeta$ over $ X$. W e le t$\ m u=u \zeta$d en o t e th eun derl yin geucli dean vec tor bun dle. Beca use $\ze ta $embedsinto $(u\zeta )_ {\mathbb C }= \mu _{\mat h b b C}$, t he hermitian $G$-vector bundle$\m u_{\m athb b C}$ isals o ampl e.B y addi ng a t rivia lcom p l ex li n e b und le to $\zeta $ , if nece ss ary, we can moreover assume t h atthere existsa $ G$-e q u iv ari a nt lin ea r is o m etric embedding $j:X\time s\ un derline{{\ m ath bb R}}{\l ongrigh tarro w }\mu=u\ zeta$ ofthe trivi al ${\ m a thb b R}$-line bundleover $X$. The e m be dding $j $ para me ter izesa triv i al1-dim ension al summa nd in $ \mu$, an d hence a pinch map $p: S^\mu{ \long rig htarrow}S ^\m u \ve e_X S^\mu $. T he ‘pinchsum ’ | G$-cells, there_are only_finitely many conjugacy classes_of finite_subgroups_of $G$_that_occur as isotropy_groups of points_of $X$. In particular,_the isotropy groups_of_$X$ have bounded order. Since $G$ is discrete, Corollary 2.7 of [@luck-oliver:completion] thus provides_a_hermitian $G$-vector_bundle_$\zeta$_over $X$ such that for_every point $x\in X$, the_fiber $\zeta_x$_is a multiple of the regular representation of_the_isotropy group $G_x$._In particular, the $G_x$-action on $\zeta_x$ is faithful, and_hence $\operatorname{Sym}(\zeta_x)$ is a complete complex_$G_x$-universe, compare [@schwede:global_Rk.6.3.22]._So_the bundle $\zeta$ is_ample.
\[prop:ample bundle\] Let $G$ be a_discrete group and $X$ a finite_proper $G$-CW-complex. Then the map ${\langle -\rangle}:_\operatorname{Vect}_G(X){\longrightarrow}{\mathbf{ku}}_G{\llbracket X\rrbracket}$ defined in is additive.
The_sum in the group ${\mathbf{ku}}_G{\llbracket_X\rrbracket}$ is_defined by addition via a_fiberwise pinch map._We will_relate the pinch_sum to a different binary operation,_the ‘bundle sum’,_by an Eckmann-Hilton argument.
Proposition \[prop:universal bundle\]_(ii)_provides an ample_hermitian_$G$-vector_bundle $\zeta$_over $X$. We_let_$\mu=u\zeta$ denote_the_underlying euclidean vector bundle. Because $\zeta$_embeds_into $(u\zeta)_{\mathbb C}=\mu_{\mathbb C}$, the hermitian $G$-vector_bundle $\mu_{\mathbb C}$ is_also_ample. By adding a_trivial complex line bundle to_$\zeta$, if necessary, we can moreover_assume that_there exists_a $G$-equivariant linear isometric embedding $j:X\times \underline{{\mathbb R}}{\longrightarrow}\mu=u\zeta$ of the trivial_${\mathbb R}$-line bundle over $X$. The_embedding $j$ parameterizes a_trivial 1-dimensional_summand_in $\mu$, and_hence_a pinch_map $p:S^\mu{\longrightarrow}S^\mu\vee_X S^\mu$. The ‘pinch sum’ |
]\
& $V=23.04\pm$0.08 $^{\dagger}$ & &[@vrees99]\
& $R=22.58\pm$0.06\
& $I=21.95\pm$0.25\
& $J=20.87\pm$0.11\
& $H=20.27\pm$0.19\
& $K=19.62\pm$0.12\
& $U=23.6\pm$0.15 & & [@castro99]\
& $B=24.23\pm$0.1 $^{\dagger}$ &\
& $V=24.20\pm$0.15\
& $R=23.77\pm$0.1\
& $I=23.65\pm$0.15\
& $K=21.7\pm$0.3 & Keck I/NIRC& [@chary01]\
& $U$ = 23.12$\pm$0.05 &ESO-3.6m& [@lise03]\
& $B$ = 23.36$\pm$0.09& DK-1.5m/DFOSC&\
& $V$= 22.39$\pm$0.03 & DK-1.5m/DFOSC&\
& $R$= 21.84$\pm$0.02 & DK-1.5m/DFOSC&\
& $I$= 21.41$\pm$0.03& DK-1.5m/DFOSC&\
& $J$= 20.81$\pm$0.17& NTT/SOFI\
& $H$= 20.25$\pm$0.19& NTT/SOFI\
& $K\!s$=20.05$\pm$0.1& & [@lefloch03]\
& $B=25.19\pm$0.17 $^{\dagger}$ & & [@cas01]\
& $V=24.55\pm$0.16\
& $R=24.26\pm$0.15\
& $I=23.3\pm$0.2\
& $K=21.7 \pm$0.2 & NIRC/Keck I &[@chary | ] \
& $ V=23.04\pm$0.08 $ ^{\dagger}$ & & [ @vrees99]\
& $ R=22.58\pm$0.06\
& $ I=21.95\pm$0.25\
& $ J=20.87\pm$0.11\
& $ H=20.27\pm$0.19\
& $ K=19.62\pm$0.12\
& $ U=23.6\pm$0.15 & & [ @castro99]\
& $ B=24.23\pm$0.1 $ ^{\dagger}$ & \
& $ V=24.20\pm$0.15\
& $ R=23.77\pm$0.1\
& $ I=23.65\pm$0.15\
& $ K=21.7\pm$0.3 & Keck I / NIRC & [ @chary01]\
& $ U$ = 23.12$\pm$0.05 & ESO-3.6 m & [ @lise03]\
& $ B$ = 23.36$\pm$0.09 & DK-1.5m / DFOSC&\
& $ V$= 22.39$\pm$0.03 & DK-1.5m / DFOSC&\
& $ R$= 21.84$\pm$0.02 & DK-1.5m / DFOSC&\
& $ I$= 21.41$\pm$0.03 & DK-1.5m / DFOSC&\
& $ J$= 20.81$\pm$0.17 & NTT / SOFI\
& $ H$= 20.25$\pm$0.19 & NTT / SOFI\
& $ K\!s$=20.05$\pm$0.1 & & [ @lefloch03]\
& $ B=25.19\pm$0.17 $ ^{\dagger}$ & & [ @cas01]\
& $ V=24.55\pm$0.16\
& $ R=24.26\pm$0.15\
& $ I=23.3\pm$0.2\
& $ K=21.7 \pm$0.2 & NIRC / Keck I & [ @chary | ]\
& $V=23.04\pl$0.08 $^{\dagger}$ & &[@vrees99]\
& $R=22.58\pm$0.06\
& $I=21.95\pm$0.25\
& $M=20.87\pm$0.11\
& $H=20.27\pm$0.19\
& $K=19.62\pm$0.12\
& $U=23.6\pm$0.15 & & [@castrm99]\
& $B=24.23\pm$0.1 $^{\sagger}$ &\
& $X=24.20\pm$0.15\
& $R=23.77\pm$0.1\
& $I=23.65\pm$0.15\
& $K=21.7\pm$0.3 & Keck I/NIRC& [@rhart01]\
& $U$ = 23.12$\ki$0.05 &ESO-3.6m& [@lise03]\
& $B$ = 23.36$\pm$0.09& AK-1.5m/DFOSC&\
& $N$= 22.39$\pm$0.03 & DK-1.5m/EFOSR&\
& $R$= 21.84$\pm$0.02 & DK-1.5m/DFOSC&\
& $I$= 21.41$\pm$0.03& DK-1.5m/DFOSC&\
& $M$= 20.81$\pm$0.17& UTV/SOFI\
& $H$= 20.25$\pm$0.19& NTT/SPFI\
& $K\!s$=20.05$\pm$0.1& & [@nefloch03]\
& $B=25.19\pm$0.17 $^{\dacgdr}$ & & [@cas01]\
& $V=24.55\pm$0.16\
& $R=24.26\pm$0.15\
& $I=23.3\pm$0.2\
& $K=21.7 \pm$0.2 & NIRC/Keck I &[@chary | ]\ & $V=23.04\pm$0.08 $^{\dagger}$ & &[@vrees99]\ & $I=21.95\pm$0.25\ $J=20.87\pm$0.11\ & & $K=19.62\pm$0.12\ & $B=24.23\pm$0.1 &\ & $V=24.20\pm$0.15\ $R=23.77\pm$0.1\ & $I=23.65\pm$0.15\ $K=21.7\pm$0.3 & Keck I/NIRC& [@chary01]\ & = 23.12$\pm$0.05 &ESO-3.6m& [@lise03]\ & $B$ = 23.36$\pm$0.09& DK-1.5m/DFOSC&\ & $V$= 22.39$\pm$0.03 & & $R$= 21.84$\pm$0.02 & DK-1.5m/DFOSC&\ & $I$= 21.41$\pm$0.03& DK-1.5m/DFOSC&\ & $J$= 20.81$\pm$0.17& NTT/SOFI\ $H$= NTT/SOFI\ $K\!s$=20.05$\pm$0.1& [@lefloch03]\ & $B=25.19\pm$0.17 $^{\dagger}$ & & [@cas01]\ & $V=24.55\pm$0.16\ & $R=24.26\pm$0.15\ & $I=23.3\pm$0.2\ & $K=21.7 \pm$0.2 NIRC/Keck I &[@chary | ]\
& $V=23.04\pm$0.08 $^{\dagger}$ & &[@vrees99]\
& $R=22.58\pm$0.06\
& $I=21.95\pm$0.25\
& $J=20.87\pm$0.11\
& $H=20.27\pM$0.19\
& $K=19.62\pm$0.12\
& $U=23.6\pm$0.15 & & [@casTro99]\
& $B=24.23\pM$0.1 $^{\daGgeR}$ &\
& $V=24.20\Pm$0.15\
& $R=23.77\pM$0.1\
& $I=23.65\pm$0.15\
& $k=21.7\pm$0.3 & Keck I/NIRC& [@chARy01]\
& $U$ = 23.12$\pM$0.05 &ESO-3.6m& [@lise03]\
& $B$ = 23.36$\pm$0.09& DK-1.5m/DFOSC&\
& $V$= 22.39$\pM$0.03 & DK-1.5m/DfOsc&\
& $R$= 21.84$\pm$0.02 & dk-1.5m/dFOSC&\
& $i$= 21.41$\pm$0.03& DK-1.5m/DfoSc&\
& $j$= 20.81$\Pm$0.17& NtT/sOfI\
& $H$= 20.25$\Pm$0.19& ntT/sOFI\
& $K\!S$=20.05$\pm$0.1& & [@Lefloch03]\
& $b=25.19\pm$0.17 $^{\dagger}$ & & [@cAs01]\
& $V=24.55\Pm$0.16\
& $r=24.26\pm$0.15\
& $I=23.3\pm$0.2\
& $K=21.7 \pm$0.2 & NIRc/keCk I &[@chary | ]\
& $V=23.04\pm$0.08 $^{\ dagger}$ & &[@v ree s99 ]\
& $ R=22 .58\pm$0.06\
& $I=2 1.95\pm$0.25\
& $J=20. 87\pm $0 . 11\& $ H=20. 27\pm$0 . 19 \ & $ K= 19 .62 \p m $0 .12\& $ U=23.6\ pm$0.15 && [ @c astro99]\
&$ B= 24.23\pm$0 .1$^{\dagger}$ &\
& $V= 24 .20 \ pm$0. 15\
& $R =23.77 \ pm$0.1 \
& $I=23 .6 5 \pm$0. 1 5\
& $K = 2 1. 7\pm $0.3 & Keck I/NIR C &[ @chary01]\
& $ U$ = 2 3. 1 2$ \ p m$0 .05 &ESO-3.6m &[@lis e 03]\
&$ B$ = 23. 3 6$\pm$0.09& D K-1.5m/DFOS C &\& $V$= 2 2.3 9 $\pm$0 .03 & D K -1. 5m/DFOSC&\& $R $= 21.84$ \pm$0. 0 2 & DK- 1 .5m/DFO SC&\
& $I $=21.4 1 $\ pm $0. 03 & DK - 1. 5m/ D FOS C&\
& $J $= 2 0.81$ \pm$ 0 . 1 7 & NT T/S OFI\
& $H $= 20.25$\pm$ 0.1 9& N T T/S OFI\& $K\ !s$= 20 .05$\ pm$0.1 & & [ @l efloch03]\
& $B =25. 19\pm$0.1 7 $ ^{ \da gg er}$& & [@c as0 1]\
& $V=2 4.55\pm $ 0.1 6\ & $R =24.26\pm$0.15\
&$I = 2 3. 3\pm$0.2 \
& $K = 21 .7 \pm$0.2&NIR C/Ke c k I &[ @cha r y | ]\
& $V=23.04\pm$0.08_$^{\dagger}$ &_&[@vrees99]\
& $R=22.58\pm$0.06\
& $I=21.95\pm$0.25\
& $J=20.87\pm$0.11\
&_$H=20.27\pm$0.19\
& $K=19.62\pm$0.12\
&_$U=23.6\pm$0.15_& &_[@castro99]\
&_$B=24.23\pm$0.1 $^{\dagger}$ &\
&_$V=24.20\pm$0.15\
& $R=23.77\pm$0.1\
& $I=23.65\pm$0.15\
&_$K=21.7\pm$0.3 & Keck I/NIRC&_[@chary01]\
& $U$ =_23.12$\pm$0.05_&ESO-3.6m& [@lise03]\
& $B$ = 23.36$\pm$0.09& DK-1.5m/DFOSC&\
& $V$= 22.39$\pm$0.03 & DK-1.5m/DFOSC&\
& $R$= 21.84$\pm$0.02 & DK-1.5m/DFOSC&\
&_$I$=_21.41$\pm$0.03& DK-1.5m/DFOSC&\
&_$J$=_20.81$\pm$0.17&_NTT/SOFI\
& $H$= 20.25$\pm$0.19& NTT/SOFI\
& $K\!s$=20.05$\pm$0.1&_& [@lefloch03]\
& $B=25.19\pm$0.17 $^{\dagger}$ &_& [@cas01]\
&_$V=24.55\pm$0.16\
& $R=24.26\pm$0.15\
& $I=23.3\pm$0.2\
& $K=21.7 \pm$0.2 & NIRC/Keck I_&[@chary |
$|E^{\text{v}}|$), whereas Basic DP examines the whole DP space. 2) ILSP and SLSP effectively reduce loco-state space to $X$; 3) The rounding strategy scales the total possible v-path lengths from $O(2^{|E^{\text{v}}|})$ to $O(\frac{|X|}{\epsilon})$.
Enhancements and Extensions {#sec:extensions}
===========================
In this section, we propose an enhancement for DROP, then show that DEWN can serve as a building block to support other spatial queries.
Critical Orientation Simplification
-----------------------------------
While DEWN is efficient and effective in solving DROP, the practical efficiency is still a concern in some applications. In Simp-DEWN (see also Section \[sec:exp\]) as a simplified implementation, we introduce the COS strategy to efficiently simplify the loco-states by ignoring the orientations for the computationally intensive parts of DEWN. The idea of COS is as follows. On solving DROP, DEWN always examine the user orientations in the dual worlds carefully to leverage RO to steer users away from obstacles, incurring the side-effect of a large time complexity, since a single pair of locations $(\gamma^{\text{v}}, \gamma^{\text{p}})$ corresponds to $|\Theta|^2$ valid loco-states. The final time complexity is then dependent on $O(\Theta^4)$. However, through careful investigation of the RW paths, we found that the complexity can be significantly reduced by allowing an additional RW operation at each loco-state along the RW path. More specifically, if an additional RO (or Reset, but more costly) is allowed at each loco-state along the RW path, then the orientations of the user need not to be considered in DEWN, which trades solution quality for computational efficiency.
Concretely, this is achieved by setting $\Theta = \{ 0^{\circ} \}$, or equivalently speaking, merging all the loco-states with identically locations $(\gamma^{\text{v}}, \gamma^{\text{p}})$. The additional RW cost incurred can be bounded since this incurs at most the largest cost of an RO or Reset at each loco-state. Denote $C_\theta$ as the maximum RW cost incurred, depending on the cost model, to correct the user orientation. The following theorem shows that this simplification gives an error-bounded FPTAS.
\[ | $ |E^{\text{v}}|$), whereas Basic DP examines the whole DP quad. 2) ILSP and SLSP efficaciously reduce loco - department of state distance to $ X$; 3) The rounding strategy scales the full potential v - path lengths from $ O(2^{|E^{\text{v}}|})$ to $ O(\frac{|X|}{\epsilon})$.
Enhancements and Extensions { # sec: extensions }
= = = = = = = = = = = = = = = = = = = = = = = = = = =
In this section, we nominate an enhancement for DROP, then show that DEWN can service as a building block to support other spatial queries.
Critical Orientation Simplification
-----------------------------------
While DEWN is effective and effective in solving DROP, the practical efficiency is still a business in some applications. In Simp - DEWN (see also Section \[sec: exp\ ]) as a simplified implementation, we insert the COS strategy to efficiently simplify the loco - states by ignoring the orientations for the computationally intensive parts of DEWN. The theme of COS is as follows. On solving DROP, DEWN always examine the user orientations in the dual worlds carefully to leverage RO to steer users away from obstacles, incurring the side - effect of a large clock time complexity, since a single pair of location $ (\gamma^{\text{v } }, \gamma^{\text{p}})$ correspond to $ |\Theta|^2 $ valid loco - states. The final time complexity is then pendent on $ O(\Theta^4)$. However, through careful investigation of the RW paths, we found that the complexity can be significantly reduce by allowing an additional RW operation at each loco - state along the RW way. More specifically, if an additional RO (or Reset, but more costly) is allowed at each loco - state along the RW path, then the orientations of the user necessitate not to be considered in DEWN, which trades solution quality for computational efficiency.
Concretely, this is achieved by setting $ \Theta = \ { 0^{\circ } \}$, or equivalently speak, merging all the loco - states with identically locations $ (\gamma^{\text{v } }, \gamma^{\text{p}})$. The additional RW cost incurred can be bounded since this incurs at most the big cost of an RO or Reset at each loco - state. Denote $ C_\theta$ as the maximal RW cost incurred, depending on the cost model, to right the user predilection. The following theorem show that this simplification gives an error - bounded FPTAS.
\ [ | $|E^{\tfxt{v}}|$), whereas Basic DP exxmines the wholg EP spare. 2) ILSL and SLRP effectively reduce loco-stete wpace to $X$; 3) The rounding stfategy scwles the totel possible v-pati lengths from $L(2^{|E^{\tert{t}}|})$ to $O(\frac{|X|}{\epsikon})$.
Enhancekents and Extetskous {#sec:extensions}
===========================
In this section, we pwopose sn enhancement fjr DGO[, thsn show that DEWN can serve as a bhilding block to suplort other spatial queries.
Fritlcal Orientation Slmplificatiin
-----------------------------------
Whyoe DEWN is ewficient and effective in solving DROP, the practical dfficnency is stulo a woncern in wome wpplications. In Simp-GEWN (ser also Section \[ser:exp\]) as a simplified implxmentation, we introdoce the COV atrategy to efficuebtly vimpnify rhe lodo-xtztes bj ijnoring the orientatiobs for the computatoogqlly intensivs partf jf DEWN. The idea of COS is as follows. Ot sklving DROP, DEWN always examine the user origntations yn the dual worlds carefully to leverage RO to staer uadrs arxt vrom obstacles, incurring the side-effect of a jzrbe time complexiby, since a single lalr jf locations $(\eamma^{\tzst{b}}, \gamma^{\text{p}})$ corredponds jo $|\Thera|^2$ valid joco-xtates. The final time complwxity is theu dwpendent on $O(\Theta^4)$. However, thxough varefil investigation of the RW laths, we foknd that ffe complexity cav bv sicnificanuuy reduced by alljwing an edditnonal RW opetation wt each lofo-staba along the RW patj. Morg spechfically, iv an additional RO (or Reset, but more costly) ix dllmwed at zach lpco-state alond the RW path, jhen the jriengations of the usxr need not eo be considesgd in DEWN, whmch tradef sooutiin qualkgy for computayional efynciency.
Cobcretely, this is aghievgd by setting $\Thecc = \{ 0^{\circ} \}$, or equifaldntjy s'eakigc, merging aln thd luvo-stages with idcntkcalky locations $(\gamma^{\teft{v}}, \famma^{\text{p}})$. The addotlonal RW xost inctrred can be nounded since this incucs at oost yhe largest cost of an RO or Resef at each logo-state. Denotq $C_\tmeta$ as the marimum RW cost incurred, depending on the rost model, to correct jhe user orientation. Jhe following vheorei shows tvat this simplificatuon gives an errov-bounded FPTAS.
\[ | $|E^{\text{v}}|$), whereas Basic DP examines the whole 2) and SLSP reduce loco-state space strategy the total possible lengths from $O(2^{|E^{\text{v}}|})$ $O(\frac{|X|}{\epsilon})$. Enhancements and Extensions {#sec:extensions} =========================== this section, we propose an enhancement for DROP, then show that DEWN can as a building block to support other spatial queries. Critical Orientation Simplification ----------------------------------- DEWN efficient effective solving DROP, the practical efficiency is still a concern in some applications. In Simp-DEWN (see also \[sec:exp\]) as a simplified implementation, we introduce the strategy to efficiently simplify loco-states by ignoring the orientations the intensive parts DEWN. idea COS is as On solving DROP, DEWN always examine the user orientations in the dual worlds carefully to leverage RO steer users obstacles, incurring side-effect a time complexity, since pair of locations $(\gamma^{\text{v}}, \gamma^{\text{p}})$ corresponds loco-states. The final time complexity is then dependent $O(\Theta^4)$. However, careful investigation of the RW paths, found that the complexity can be significantly reduced allowing an additional RW operation at each loco-state along the RW path. More specifically, if RO (or Reset, but costly) is allowed each along RW then the of the user need not to be considered in DEWN, which solution quality for computational efficiency. Concretely, this is achieved by = 0^{\circ} \}$, or speaking, merging all the with locations $(\gamma^{\text{v}}, \gamma^{\text{p}})$. The cost can this at the largest cost of RO or Reset at each Denote $C_\theta$ as the on the cost model, to correct the user The following theorem shows that this simplification an error-bounded FPTAS. \[ | $|E^{\text{v}}|$), whereas Basic DP examiNes the wholE DP spAce. 2) iLSp aNd SLsP efFectively reducE Loco-State space to $X$; 3) The roundiNg strAtEGy scALeS the tOtal posSIbLE V-paTh LeNgtHs FRoM $O(2^{|E^{\teXt{v}}|})$ To $O(\frac{|x|}{\epsilon})$.
EnHanCeMents and ExteNSiOns {#sec:exteNsiOns}
===========================
In this secTioN, we proPoSe aN EnhanCemEnt foR DROP, tHEn show That DEWN cAn SErve as A BuildinG BLoCk to Support other spatiAL qUEries.
Critical ORientaTiON SIMPliFicAtion
-----------------------------------
While dEwN is eFFicient ANd EFFEctIVe in solving DRoP, the practiCAl eFficieNcY is STill a cOncerN iN SomE applicatioNs. In simp-DEWN (sEe also sEction \[sEC:exp\]) as a SimpliFieD imPlemENtAtIon, We INtrODuCe tHE COs strategY tO eFficiEntlY SIMPlifY thE locO-statEs by ignoring tHe oRienTAtiOns foR the cOmpuTaTionaLly intEnsivE pArts of DEWN. The idEa of cOS is as foLloWs. on sOlVing DroP, DEWN AlwAys Examine The user ORieNtATIOnS in the dual worlds caReFULlY to leverAge RO tO StEeR Users awaY fRom ObstACLes, inCurrINg The side-eFfect oF A lArGe time cOmPlexitY, sIncE a sIngle PAir oF locatIons $(\gammA^{\text{V}}, \Gamma^{\text{p}})$ corrESponds to $|\Theta|^2$ VAlID LoCO-staTes. the final timE comPLexiTy is THeN dePEndenT on $O(\THeTA^4)$. HOWever, through careful InVestigAtion Of the RW paths, wE found that THE ComplexiTy caN Be SIgnificantly reDuced By allowing AN additioNal RW OperatioN at each loCO-State aloNg tHe Rw paTh. MORE sPecifically, if AN AddiTiOnal RO (oR ReSet, but mOre CosTly) Is aLlOwed at eacH loco-staTe AlOnG tHe Rw path, THen the orIeNtaTiOns Of the USer neeD not tO be cOnSiDEreD in DEWN, WHiCH TradEs SoLutiOn qUaLity fOr coMPutAtional EfficiencY.
CoNCretElY, tHis is acHieved by settiNg $\theta = \{ 0^{\circ} \}$, oR eQuiValentLY Speaking, Merging all the loco-states WIth idenTicAlly lOcatIons $(\gamma^{\TexT{v}}, \gammA^{\teXT{p}})$. The aDditioNal RW CoSt iNCUrred CAN bE boUnDed since thIS IncUrs at MoSt thE largesT cost of an RO or Reset AT eaCh loco-state. DeNotE $C_\thETA$ aS thE MaXImuM Rw CosT INcurred, dependinG on the cost MoDEl, To correct tHE usEr OrientaTion. The FolloWIng theoRem shows tHat this siMpLifiCATioN gives an erRor-boundEd FPTAS.
\[ | $|E^{\text{v}}|$), wherea s Basic DP exam ine s t he who le D P space. 2) IL S P an d SLSP effectively red uce l oc o -sta t espace to $X$ ; 3 ) The r ou ndi ng st rateg y s cales t he total p oss ib le v-path le n gt hs from $O (2^ {|E^{\text{v }}| })$ to $ O(\ f rac{| X|} {\eps ilon}) $ .
Enh ancements a n d Exte n sions { # s ec :ext ensions}
======== = == = ============== =
Inth i ss e cti on, we propos ean en h ancemen t f o r DRO P , then show t hat DEWN ca n se rve as a bu i ldingblock t o su pport other spa tial quer ies.
C ritical Orienta tion S imp lif icat i on
- --- -- - --- - -- --- - --- -------- -- -- ---
Whil e D E WN i s e ffic ientand effective in sol v ing DROP , the pra ct icaleffici encyis still a concer n in some app lic at ion s. In S i mp-DEW N ( see also S ection\ [se c: e x p \] ) as a simplifiedim p l em entation , we i n tr od u ce the C OS st rate g y to e ffic i en tly simp lify t h elo co-stat es by ig no rin g t he or i enta tionsfor thecompu t ationally inte n sive parts of DE W N .T he i dea of COS isas f o llow s. O n s olv i ng DR OP, D EW N a l ways examine the us er orien tatio ns in the dua l worlds c a r e fully to lev e ra g e RO to steerusers away from obstacle s, in curringthe side- e f fect ofa l arg e t ime c om plexity, sinc e a si ng le pair of locati ons $( \ga mma ^{ \text{v}} , \gamma ^{ \t ex t{ p}} )$ co r responds t o $ |\ The ta|^2 $ valid loco -sta te s. The finalt im e comp le xi ty i s t he n dep ende n t o n $O(\T heta^4)$. Ho w ever ,th rough c areful invest ig ation of t he RW paths , we found that the complexity ca n be sig nif icant ly r educed by al lowing an additi onal R W ope ra tio n at ea c h l oco -s tate along t heRW pa th . Mo re spec ifically, if an ad d iti onal RO (or R ese t, b u t m ore co s tly )i s a l l owed at each lo co-state a lo n gthe RW pat h , t he n the o rientat ionso f the u ser neednot to be c onsi d e red in DEWN,which tr ades solu t ion q u al ity f orcomput at ion al ef ficien c y.
Conc retely ,this i s ach ie ved by s etting $\Theta = \{ 0^{ \circ} \}$, or equivale ntl y sp eaking, m ergi ng all the lo co- state s w i th id enti c al lyl ocati ons$ (\gamma^{ \ te xt{ v } }, \gamma^{\t e x t {p} })$.The additi onal RW cost incurred can be bounded sin c e th isi ncur sat most the la rge st c ost of a nRO or Reset at each l o co-st ate. D enote$C_\the t a $a s themaxi mum RW costinc ur r ed, dep en di n g on t he c os t mode l, toc orre c t the user orient ation . The f o llo wingth eorem s h owsthat thissimplificat ion gi vesan er ror-bou nd ed FPT AS.
\[ | $|E^{\text{v}}|$),_whereas Basic_DP examines the whole_DP space._2)_ILSP and_SLSP_effectively reduce loco-state_space to $X$;_3) The rounding strategy_scales the total_possible_v-path lengths from $O(2^{|E^{\text{v}}|})$ to $O(\frac{|X|}{\epsilon})$.
Enhancements and Extensions {#sec:extensions}
===========================
In this section, we propose an_enhancement_for DROP,_then_show_that DEWN can serve as_a building block to support_other spatial_queries.
Critical Orientation Simplification
-----------------------------------
While DEWN is efficient and effective_in_solving DROP, the_practical efficiency is still a concern in some applications._In Simp-DEWN (see also Section \[sec:exp\])_as a simplified_implementation,_we_introduce the COS strategy_to efficiently simplify the loco-states by_ignoring the orientations for the computationally_intensive parts of DEWN. The idea of_COS is as follows. On solving_DROP, DEWN always examine the_user orientations_in the dual worlds carefully_to leverage RO_to steer_users away from_obstacles, incurring the side-effect of a_large time complexity,_since a single pair of locations_$(\gamma^{\text{v}},_\gamma^{\text{p}})$ corresponds to_$|\Theta|^2$_valid_loco-states. The_final time complexity_is_then dependent_on_$O(\Theta^4)$. However, through careful investigation of_the_RW paths, we found that the complexity_can be significantly reduced_by_allowing an additional RW_operation at each loco-state along_the RW path. More specifically, if_an additional_RO (or_Reset, but more costly) is allowed at each loco-state along the_RW path, then the orientations of_the user need not_to be_considered_in DEWN, which_trades_solution quality_for computational efficiency.
Concretely, this is achieved by_setting $\Theta_= \{ 0^{\circ} \}$, or equivalently_speaking, merging all the_loco-states_with identically locations $(\gamma^{\text{v}}, \gamma^{\text{p}})$. The_additional RW cost incurred can be_bounded since this incurs at_most_the_largest cost of an RO_or Reset at each loco-state. Denote_$C_\theta$ as the_maximum RW cost incurred, depending on the_cost_model, to correct the user orientation._The_following theorem shows that this simplification_gives_an_error-bounded FPTAS.
\[ |
dub this construction the *discrete Brownian snake*, as it is the natural discrete analogue of Le Gall’s Brownian snake: see [@L99] for details.
The construction proceeds by using the random variables $\{L_{i}
\}_{i\geq 0}$ to construct a *Galton-Watson tree*. This construction is standard: see [@AN72]. If the offspring distribution has mean $1$, as we shall assume throughout, then the resulting Galton-Watson tree is almost surely a finite, rooted tree with vertices arranged in *generations*, beginning with the root at generation $0$. To each vertex $v$ is attached one of the random variables $L_{i}$, with $L_{0}$ attached to the root; for each vertex $v$ the random variable $L_{i}$ determines the number of offspring vertices. The random variables $L_{i}$ can be attached to vertices in any number of different ways, the most common being the *breadth-first* rule, in which the values $L_{i}$ are read successively from the stack generation-by-generation, left-to-right.
Given the realization of the Galton-Watson tree, we attach unit exponential random variables $T_{i}$ and standard Wiener processes $W^{i}$ to the *edges* of the tree in such a way that the index $i=i (v)$ matches the index of the random variable $L_{i}$ attached to the *lower* vertex $v$ of the edge (the incident vertex with higher generation number). The random variable $T_{i}$ attached to an edge determines the real time elapsed between reproduction events, and the Wiener process $W^{i}$ determines the displacement of the particle in real time from its position at the last reproduction event. Thus, the particles alive at (real) time $t$ are in one-to-one correspondence with the vertices $v$ of the tree such that $$\sum_{w< v} T_{i (w)}<t\leq \sum_{w\leq v}T_{i (w)};$$ here the symbols $<$ and $\leq$ indicate the ordering of vertices $w$ along the geodesic path in the tree from the root to $v$. The spatial position of the particle represented by vertex $v$ at time $t$ is $$y+\sum_{w<v} W^{i (w)} (T_{i (w)})+W | dub this construction the * discrete Brownian snake *, as it is the natural discrete analogue of Le Gall ’s Brownian hydra: examine [ @L99 ] for details.
The construction go by use the random variables $ \{L_{i }
\}_{i\geq 0}$ to construct a * Galton - Watson tree *. This structure is standard: see [ @AN72 ]. If the offspring distribution has mean $ 1 $, as we shall bear throughout, then the resulting Galton - Watson tree is almost surely a finite, frozen tree with vertices arranged in * generations *, beginning with the ancestor at generation $ 0$. To each vertex $ v$ is attached one of the random variables $ L_{i}$, with $ L_{0}$ attached to the root; for each vertex $ v$ the random variable star $ L_{i}$ determines the number of offspring vertex. The random variables $ L_{i}$ can be attached to vertex in any number of unlike ways, the about common being the * breadth - first * rule, in which the value $ L_{i}$ are read successively from the stack generation - by - generation, left - to - right field.
give the realization of the Galton - Watson tree, we attach unit exponential random variables $ T_{i}$ and standard Wiener processes $ W^{i}$ to the * edges * of the tree in such a way that the index $ i = i (v)$ matches the exponent of the random variable star $ L_{i}$ attached to the * lower * vertex $ v$ of the edge (the incident vertex with eminent generation number). The random variable $ T_{i}$ attach to an boundary determines the real time elapse between reproduction events, and the Wiener process $ W^{i}$ determines the displacement of the particle in substantial time from its position at the last reproduction consequence. Thus, the atom alive at (real) time $ t$ are in one - to - one correspondence with the vertices $ v$ of the tree such that $ $ \sum_{w < five } T_{i (w)}<t\leq \sum_{w\leq v}T_{i (w)};$$ here the symbols $ < $ and $ \leq$ indicate the ordering of vertex $ w$ along the geodesic path in the tree from the root to $ v$. The spatial position of the atom represented by vertex $ v$ at time $ t$ is $ $ y+\sum_{w < v } W^{i (w) } (T_{i (w)})+W | duh this construction the *aiscrete Brownicb snakx*, as it is the vatural discrete analogue of Lw Galo’s Brownian snake: see [@U99] for detwils.
The xonsuruction proceeds by usinn the vandok variables $\{L_{i}
\}_{i\neq 0}$ to convtruct a *Galtot-Wxtdon tree*. This construction is standwrd: see [@AJ72]. If the offspting cystrjbution has mean $1$, as we shall assuje throlghout, then the rrsulting Galton-Watson tree is wlmost surely a fijite, rooted treq with verticds arranged in *generatjons*, beginning with the root at genexation $0$. To gcxh ggrtex $v$ is avtacheq one of the random eariablrs $L_{i}$, with $L_{0}$ abtachxd ti the root; for each vxrtex $v$ the random vwriable $L_{h}$ betermines the number od offvpritg vdetizes. Tie dandom vaciables $L_{i}$ dan be attaxhed to vertices in agj number of djfferegt ways, the most common being the *breadth-xirat* rule, in which the vaoues $L_{i}$ are read succgssively fwom the stack generation-by-generation, left-to-right.
Ghven vhd rtaoizatkin of the Galton-Watson tree, we attach unit expogsnuiak random varianles $T_{i}$ and standatd Wogner processes $W^{i}$ to ths *edges* of the tref in sush a qay that uhe imdex $i=i (v)$ matches the index of the randjn variable $L_{i}$ attaehed to the *uowet* vertrx $v$ of the edge (the ineident vertex witj higher fdneration number). Thv ratdom variable $T_{i}$ attached eo an edgx detzrmines ghe teal tiie elapsed betwcan reproduction evfnts, cnd tve Wiener orocess $W^{i}$ determines the displerement of the pdrtpcle in rzal tike from its pjsition at the last rzproduztion evenn. Thus, thx particles wlive at (real) jime $t$ are in one-to-onq coerespondencd with the vertoces $v$ of the tree wuch that $$\sum_{w< v} T_{l (w)}<t\lgq \sum_{w\leq v}T_{i (w)};$$ ktre the symbols $<$ amd $\ueq$ ijdmcate dhe ordering of xergoces $d$ along the eeodrsic path in the trea frkm the root to $v$. Tne spatial positiog of the partocle represented bj vervex $v$ et timr $t$ is $$y+\sum_{w<v} W^{i (w)} (T_{i (w)})+W | dub this construction the *discrete Brownian snake*, is natural discrete of Le Gall’s details. construction proceeds by the random variables \}_{i\geq 0}$ to construct a *Galton-Watson This construction is standard: see [@AN72]. If the offspring distribution has mean $1$, we shall assume throughout, then the resulting Galton-Watson tree is almost surely a rooted with arranged *generations*, beginning with the root at generation $0$. To each vertex $v$ is attached one of random variables $L_{i}$, with $L_{0}$ attached to the for each vertex $v$ random variable $L_{i}$ determines the of vertices. The variables can attached to vertices any number of different ways, the most common being the *breadth-first* rule, in which the values $L_{i}$ read successively stack generation-by-generation, Given realization the Galton-Watson tree, unit exponential random variables $T_{i}$ and $W^{i}$ to the *edges* of the tree in a way the index $i=i (v)$ matches the of the random variable $L_{i}$ attached to the vertex $v$ of the edge (the incident vertex with higher generation number). The random variable to an edge determines real time elapsed reproduction and Wiener $W^{i}$ determines displacement of the particle in real time from its position at last reproduction event. Thus, the particles alive at (real) time in correspondence with the $v$ of the tree that v} T_{i (w)}<t\leq \sum_{w\leq here symbols indicate ordering vertices $w$ along the path in the tree from root to $v$. The represented by vertex $v$ at time $t$ is W^{i (w)} (T_{i (w)})+W | dub this construction the *disCrete BrownIan snAke*, As iT iS the NatuRal discrete anaLOgue Of Le Gall’s Brownian snake: See [@L99] fOr DEtaiLS.
THe conStructiON pROCeeDs By UsiNg THe RandoM vaRiables $\{l_{i}
\}_{i\geq 0}$ to coNstRuCt a *Galton-WatSOn Tree*. This coNstRuction is staNdaRd: see [@An72]. IF thE OffspRinG distRibutiON has meAn $1$, as we shaLl ASsume tHRoughouT, THeN the Resulting Galton-WaTSoN Tree is almost suRely a fInITe, ROOteD trEe with vertIcEs arrANged in *gENeRATIonS*, Beginning with The root at geNEraTion $0$. To EaCh vERtex $v$ iS attaChED onE of the randoM varIables $L_{i}$, wIth $L_{0}$ atTAched to THe root; fOr each VerTex $V$ the RAnDoM vaRiABle $l_{I}$ dEteRMinEs the numBeR oF offsPrinG VERTiceS. ThE ranDom vaRiables $L_{i}$ can bE atTachED to VertiCes in Any nUmBer of DifferEnt waYs, The most common beIng tHe *breadth-FirSt* RulE, iN whicH The valUes $l_{i}$ aRe read sUccessiVEly FrOM THe Stack generation-by-gEnERAtIon, left-tO-right.
gIvEn THe realizAtIon Of thE gAlton-watsON tRee, we attAch uniT ExPoNential RaNdom vaRiAblEs $T_{I}$ and sTAndaRd WienEr procesSes $W^{i}$ TO the *edges* of the TRee in such a way THaT THe INdex $I=i (v)$ Matches the iNdex OF the RandOM vAriABle $L_{i}$ AttacHeD To THe *lower* vertex $v$ of the EdGe (the iNcideNt vertex with hIgher generATIOn number). the rANdOM variable $T_{i}$ attAched To an edge deTErmines tHe reaL time elaPsed betweEN ReproducTioN evEntS, anD THe wiener process $w^{I}$ DeteRmInes the DisPlacemeNt oF thE paRtiClE in real tiMe from itS pOsItIoN at The laST reproduCtIon EvEnt. thus, tHE partiCles aLive At (ReAL) tiMe $t$ are iN OnE-TO-one CoRrEspoNdeNcE with The vERtiCes $v$ of tHe tree sucH thAT $$\sum_{W< v} t_{i (W)}<t\leq \suM_{w\leq v}T_{i (w)};$$ here ThE symbols $<$ anD $\lEq$ iNdicatE THe orderiNg of vertices $w$ along the geODesic paTh iN the tRee fRom the rooT to $V$. The spAtiAL positIon of tHe parTiCle REPreseNTEd By vErTex $v$ at time $T$ IS $$y+\sUm_{w<v} W^{I (w)} (t_{i (w)})+W | dub this construction the *discrete Brow nia n s na ke*, asit is the natu r al d iscrete analogue of Le Gall ’s Brow n ia n sna ke: see [@ L 9 9]fo rdet ai l s.
The co nstruct ion procee dsby using the r a nd om variabl es$\{L_{i}
\}_ {i\ geq 0} $toc onstr uct a *G alton- W atsontree*. Th is constr u ction i s st anda rd: see [@AN72].I ft he offspring d istrib ut i on h asmea n $1$, aswe shal l assume th r o u gho u t, then the r esulting Ga l ton -Watso ntre e is al mostsu r ely a finite,root ed tree w ith ve r tices a r rangedin *ge ner ati ons* , b eg inn in g wi t hthe roo t at gen er at ion $ 0$.T o e achver tex$v$ i s attached on e o f th e ra ndomvaria bles $ L_{i} $, wit h $L_ {0 }$ attached totheroot; for ea ch ve rt ex $v $ the r and omvariabl e $L_{i } $ d et e r m in es the number of o ff s p ri ng verti ces. T h era n dom vari ab les $L_ { i }$ ca n be at tached t o vert i ce sin anynu mber o fdif fer ent w a ys,the mo st commo n bei n g the *breadth - first* rule,i nw h ic h the va lues $L_{i} $ ar e rea d su c ce ssi v ely f rom t he st a ck generation-by-ge ne ration , lef t-to-right.
Given ther e a lization oft he Galton-Watsontree, we attach unit exp onent ial rand om variab l e s $T_{i} $ a ndsta nda r d W iener process e s $W^ {i }$ to t he*edges* of th e t ree i n such away that t he i nd ex$i=i( v)$ matc he s t he in dex o f the r andom var ia bl e $L _{i}$ a t ta c h ed t oth e *l owe r* vert ex $ v $ o f the e dge (theinc i dent v er tex wit h higher gene ra tion numbe r) . T he ran d o m variab le $T_{i}$ attached toa n edgedet ermin es t he real t ime elaps edb etween repro ducti on ev e n ts, a n d t heWi ener proce s s $W ^{i}$ d eter mines t he displacement of the particle inrea l ti m e f rom it s po si t ion a t the last repr oduction e ve n t. Thus, the par ti cles al ive at(real ) time $ t$ are in one-to-o ne cor r e spo ndence wit h the ve rtices $v $ of t h etreesuc h that $ $\s um_{w < v} T _ {i(w)}< t\leq\s um_{w\ leq v }T _{i (w)} ;$$ here the symbols $< $ and$\leq $ i ndicate t heo rde ring of v erti ces $w$ al ong th e geo des i c pat h in th e t r ee fr om t h e root to $v $.T h espatial pos i t i onof th e p a rticle rep resented by verte x $v$ at time $ t$ i s $$y +\s u m_{w <v } W^{i (w)} (T _{i ( w ) })+W | dub_this construction_the *discrete Brownian snake*,_as it_is_the natural_discrete_analogue of Le_Gall’s Brownian snake:_see [@L99] for details.
The_construction proceeds by_using_the random variables $\{L_{i}
\}_{i\geq 0}$ to construct a *Galton-Watson tree*. This construction is standard:_see_[@AN72]. If_the_offspring_distribution has mean $1$, as_we shall assume throughout, then_the resulting_Galton-Watson tree is almost surely a finite, rooted_tree_with vertices arranged_in *generations*, beginning with the root at generation $0$._To each vertex $v$ is attached_one of the_random_variables_$L_{i}$, with $L_{0}$ attached_to the root; for each vertex_$v$ the random variable $L_{i}$ determines_the number of offspring vertices. The random_variables $L_{i}$ can be attached to_vertices in any number of_different ways,_the most common being the_*breadth-first* rule, in_which the_values $L_{i}$ are_read successively from the stack generation-by-generation,_left-to-right.
Given the realization_of the Galton-Watson tree, we attach_unit_exponential random variables_$T_{i}$_and_standard Wiener_processes $W^{i}$ to_the_*edges* of_the_tree in such a way that_the_index $i=i (v)$ matches the index of_the random variable $L_{i}$_attached_to the *lower* vertex_$v$ of the edge (the_incident vertex with higher generation number)._The random_variable $T_{i}$_attached to an edge determines the real time elapsed between reproduction_events, and the Wiener process $W^{i}$_determines the displacement of_the particle_in_real time from_its_position at_the last reproduction event. Thus, the particles_alive at_(real) time $t$ are in one-to-one_correspondence with the vertices_$v$_of the tree such that $$\sum_{w<_v} T_{i (w)}<t\leq \sum_{w\leq v}T_{i (w)};$$_here the symbols $<$ and_$\leq$_indicate_the ordering of vertices $w$_along the geodesic path in the_tree from the_root to $v$. The spatial position of_the_particle represented by vertex $v$ at_time_$t$ is $$y+\sum_{w<v} W^{i (w)} (T_{i_(w)})+W |
\sum_{i,k,l=1}^n \overline{H}_{pl}(x)
\, \overline{H}_{ql}(x) \\
&= \frac{1}{2n(n+2)} \, |S^{n-1}| \,
\sum_{i,k,l=1}^n (W_{ipkl} + W_{ilkp}) \, (W_{iqkl} + W_{ilkq}) \\
&\hspace{5mm} \cdot \int_0^\infty
\varepsilon^{n-2} \, (\varepsilon^2 + r^2)^{-n} \, r^{n+3} \, f(r^2)^2
\, dr.\end{aligned}$$ Similarly, it follows from Proposition \[integral.identity.2\] that $$\begin{aligned}
&\int_{\mathbb{R}^n} \varepsilon^{n-2} \, (\varepsilon^2 + |x|^2)^{-n}
\, \sum_{i,k,l=1}^n (\partial_l \overline{H}_{ik}(x))^2 \, x_p \, x_q \\
&= \frac{2}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,k,l=1}^n (W_{ipkl} +
W_{ilkp}) \, (W_{iqkl} + W_{ilkq}) \\
&\hspace{5mm} \cdot \int_0^\infty \varepsilon^{n-2} \, (\varepsilon^2 +
r^2)^{-n} \, r^{n+3} \\
&\hspace{15mm} \cdot \Big [ (n+4) \, f(r^2)^2 + 8r^2 \, f(r^2) \,
f'(r^2) + 4r^4 \, f'(r^2)^2 \Big ] \, dr \\
&+ \frac{1}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,j,k,l=1}^n (W_{ijkl} +
W_{ilkj})^2 \, \delta_{pq} \\
&\hspace{5mm} \cdot \int_0^\infty \varepsilon^{n-2} \, (\varepsilon^2 +
r^2)^{-n} \, r | \sum_{i, k, l=1}^n \overline{H}_{pl}(x)
\, \overline{H}_{ql}(x) \\
& = \frac{1}{2n(n+2) } \, |S^{n-1}| \,
\sum_{i, k, l=1}^n (W_{ipkl } + W_{ilkp }) \, (W_{iqkl } + W_{ilkq }) \\
& \hspace{5 mm } \cdot \int_0^\infty
\varepsilon^{n-2 } \, (\varepsilon^2 + r^2)^{-n } \, r^{n+3 } \, f(r^2)^2
\, dr.\end{aligned}$$ Similarly, it follows from Proposition \[integral.identity.2\ ] that $ $ \begin{aligned }
& \int_{\mathbb{R}^n } \varepsilon^{n-2 } \, (\varepsilon^2 + |x|^2)^{-n }
\, \sum_{i, k, l=1}^n (\partial_l \overline{H}_{ik}(x))^2 \, x_p \, x_q \\
& = \frac{2}{n(n+2)(n+4) } \, |S^{n-1}| \, \sum_{i, k, l=1}^n (W_{ipkl } +
W_{ilkp }) \, (W_{iqkl } + W_{ilkq }) \\
& \hspace{5 mm } \cdot \int_0^\infty \varepsilon^{n-2 } \, (\varepsilon^2 +
r^2)^{-n } \, r^{n+3 } \\
& \hspace{15 mm } \cdot \Big [ (n+4) \, f(r^2)^2 + 8r^2 \, f(r^2) \,
f'(r^2) + 4r^4 \, f'(r^2)^2 \Big ] \, dr \\
& + \frac{1}{n(n+2)(n+4) } \, |S^{n-1}| \, \sum_{i, j, k, l=1}^n (W_{ijkl } +
W_{ilkj})^2 \, \delta_{pq } \\
& \hspace{5 mm } \cdot \int_0^\infty \varepsilon^{n-2 } \, (\varepsilon^2 +
r^2)^{-n } \, r | \sul_{i,k,l=1}^n \overline{H}_{pl}(x)
\, \overllne{H}_{ql}(x) \\
&= \frac{1}{2n(n+2)} \, |S^{n-1}| \,
\sum_{m,k,l=1}^n (W_{ilkl} + W_{ilyp}) \, (W_{iqkl} + W_{ilkq}) \\
&\hspace{5mm} \cdlt \int_0^\ibfty
\varepsilon^{n-2} \, (\vareprilon^2 + r^2)^{-n} \, r^{n+3} \, f(r^2)^2
\, dr.\eid{aligned}$$ Similacmy, it follows nrom 'riposition \[intenral.identitf.2\] that $$\begin{alhgveb}
&\int_{\mathbb{R}^n} \varepsilon^{n-2} \, (\varepsilon^2 + |x|^2)^{-n}
\, \suk_{i,n,l=1}^n (\partial_l \oderlpnq{H}_{ik}(s))^2 \, x_p \, x_q \\
&= \frac{2}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,k,l=1}^n (W_{ilkl} +
W_{ilnp}) \, (W_{iqkl} + W_{ikkq}) \\
&\hspace{5mm} \cdot \int_0^\infty \varfpsilon^{n-2} \, (\varepsilln^2 +
r^2)^{-n} \, r^{n+3} \\
&\yspasw{15mm} \cdot \Big [ (n+4) \, f(r^2)^2 + 8r^2 \, f(r^2) \,
f'(r^2) + 4r^4 \, r'(r^2)^2 \Big ] \, dr \\
&+ \frac{1}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,j,y,l=1}^n (W_{njkl} +
W_{ilkj})^2 \, \dwltw_{kq} \\
&\hspace{5mm} \rdot \igt_0^\infty \vareixilon^{n-2} \, (\varepxilon^2 +
r^2)^{-n} \, r | \sum_{i,k,l=1}^n \overline{H}_{pl}(x) \, \overline{H}_{ql}(x) \\ &= \frac{1}{2n(n+2)} \, (W_{ipkl} + \, (W_{iqkl} + \varepsilon^{n-2} (\varepsilon^2 + r^2)^{-n} r^{n+3} \, f(r^2)^2 dr.\end{aligned}$$ Similarly, it follows from Proposition that $$\begin{aligned} &\int_{\mathbb{R}^n} \varepsilon^{n-2} \, (\varepsilon^2 + |x|^2)^{-n} \, \sum_{i,k,l=1}^n (\partial_l \overline{H}_{ik}(x))^2 \, \, x_q \\ &= \frac{2}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,k,l=1}^n (W_{ipkl} + W_{ilkp}) \, + \\ \cdot \varepsilon^{n-2} \, (\varepsilon^2 + r^2)^{-n} \, r^{n+3} \\ &\hspace{15mm} \cdot \Big [ (n+4) \, f(r^2)^2 + \, f(r^2) \, f'(r^2) + 4r^4 \, f'(r^2)^2 ] \, dr \\ \frac{1}{n(n+2)(n+4)} \, |S^{n-1}| \, \sum_{i,j,k,l=1}^n + \, \delta_{pq} &\hspace{5mm} \int_0^\infty \, (\varepsilon^2 + \, r | \sum_{i,k,l=1}^n \overline{H}_{pl}(x)
\, \overliNe{H}_{ql}(x) \\
&= \frac{1}{2N(n+2)} \, |S^{n-1}| \,
\sUm_{i,K,l=1}^n (w_{iPkl} + W_{Ilkp}) \, (w_{iqkl} + W_{ilkq}) \\
&\hspaCE{5mm} \cDot \int_0^\infty
\varepsilon^{n-2} \, (\VarepSiLOn^2 + r^2)^{-n} \, R^{N+3} \, f(R^2)^2
\, dr.\enD{aligneD}$$ siMILarLy, It FolLoWS fRom PrOpoSition \[iNtegral.ideNtiTy.2\] That $$\begin{aliGNeD}
&\int_{\mathbb{r}^n} \vArepsilon^{n-2} \, (\vaRepSilon^2 + |x|^2)^{-N}
\, \sUm_{i,K,L=1}^n (\parTiaL_l \oveRline{H}_{IK}(x))^2 \, x_p \, x_q \\
&= \Frac{2}{n(n+2)(n+4)} \, |S^{n-1}| \, \SuM_{I,k,l=1}^n (W_{iPKl} +
W_{ilkp}) \, (w_{IQkL} + W_{ilKq}) \\
&\hspace{5mm} \cdot \int_0^\INfTY \varepsilon^{n-2} \, (\vaRepsilOn^2 +
R^2)^{-N} \, r^{N+3} \\
&\HSpaCe{15mM} \cdot \Big [ (n+4) \, f(R^2)^2 + 8r^2 \, F(r^2) \,
f'(r^2) + 4r^4 \, F'(R^2)^2 \Big ] \, dr \\
&+ \fRAc{1}{N(N+2)(N+4)} \, |s^{n-1}| \, \sUM_{i,j,k,l=1}^n (W_{ijkl} +
W_{iLkj})^2 \, \delta_{pq} \\
&\hSPacE{5mm} \cdoT \iNt_0^\iNFty \varEpsilOn^{N-2} \, (\VarEpsilon^2 +
r^2)^{-n} \, r | \sum_{i,k,l=1}^n \overlin e{H}_{pl}( x)
\, \o ver li ne{H }_{q l}(x) \\
&= \f r ac{1 }{2n(n+2)} \, |S^{n-1} | \,\s u m_{i , k, l=1}^ n (W_{i p kl } + W _{ il kp} )\ ,(W_{i qkl } + W_{ ilkq}) \\&\h sp ace{5mm} \cd o t\int_0^\in fty
\varepsilon ^{n -2} \, ( \va r epsil on^ 2 + r ^2)^{- n } \, r ^{n+3} \, f ( r^2)^2 \, dr.\ e n d{ alig ned}$$ Similarly, it follows from P roposi ti o n\ [ int egr al.identit y. 2\] t h at $$\b e gi n { a lig n ed}
&\int_{\m athbb{R}^n} \va repsil on ^{n - 2} \,(\var ep s ilo n^2 + |x|^2 )^{- n}
\, \su m_{i,k , l=1}^n( \partia l_l \o ver lin e{H} _ {i k} (x) )^ 2 \, x_ p \ , x_ q \\
&=\f ra c{2}{ n(n+ 2 ) ( n +4)} \, |S^ {n-1} | \, \sum_{i, k,l =1}^ n (W _{ipk l} +W_{i lk p}) \ , (W_{ iqkl} + W_{ilkq}) \\
& \hsp ace{5mm}\cd ot \i nt _0^\i n fty \v are psi lon^{n- 2} \, ( \ var ep s i l on ^2 +
r^2)^{-n} \,r^ { n +3 } \\
&\ hspace { 15 mm } \cdot \ Bi g [ (n+ 4 ) \, f (r^2 ) ^2 + 8r^2\, f(r ^ 2) \ ,
f'(r^ 2) + 4r^ 4\,f'( r^2)^ 2 \Bi g ] \, dr \\
& + \fr a c{1}{n(n+2)(n+ 4 )} \, |S^{n-1 } |\ , \ s um_{ i,j ,k,l=1}^n ( W_{i j kl}+
W_ { il kj} ) ^2 \, \del ta _ {p q } \\
&\hspace{5mm}\c dot \i nt_0^ \infty \varep silon^{n-2 } \ , (\vare psil o n^ 2 +
r^2)^{-n} \ , r | \sum_{i,k,l=1}^n_\overline{H}_{pl}(x)
\, \overline{H}_{ql}(x)_\\
&= \frac{1}{2n(n+2)} \, |S^{n-1}|_\,
\sum_{i,k,l=1}^n (W_{ipkl}_+_W_{ilkp}) \,_(W_{iqkl}_+ W_{ilkq}) \\
&\hspace{5mm}_\cdot \int_0^\infty
\varepsilon^{n-2} \,_(\varepsilon^2 + r^2)^{-n} \,_r^{n+3} \, f(r^2)^2
\,_dr.\end{aligned}$$_Similarly, it follows from Proposition \[integral.identity.2\] that $$\begin{aligned}
&\int_{\mathbb{R}^n} \varepsilon^{n-2} \, (\varepsilon^2 + |x|^2)^{-n}
\, \sum_{i,k,l=1}^n_(\partial_l_\overline{H}_{ik}(x))^2 \,_x_p_\,_x_q \\
&= \frac{2}{n(n+2)(n+4)} \, |S^{n-1}|_\, \sum_{i,k,l=1}^n (W_{ipkl} +
W_{ilkp}) \,_(W_{iqkl} +_W_{ilkq}) \\
&\hspace{5mm} \cdot \int_0^\infty \varepsilon^{n-2} \, (\varepsilon^2 +
r^2)^{-n}_\,_r^{n+3} \\
&\hspace{15mm}_\cdot \Big [ (n+4) \, f(r^2)^2 + 8r^2 \,_f(r^2) \,
f'(r^2) + 4r^4 \, f'(r^2)^2_\Big ] \,_dr_\\
&+_\frac{1}{n(n+2)(n+4)} \, |S^{n-1}| \,_\sum_{i,j,k,l=1}^n (W_{ijkl} +
W_{ilkj})^2 \, \delta_{pq} \\
&\hspace{5mm}_\cdot \int_0^\infty \varepsilon^{n-2} \, (\varepsilon^2 +
r^2)^{-n}_\, r |
model from the first principle by using conformal field theory. The vertex operators couple closed strings to open strings that are introduced by the existence of the D-branes. We have generalized a single D instanton calculation [@GreenGutperle; @Gutperle] into that for multiple D instantons by introducing matrix Majorana-Weyl spinor fields. Although Okawa and Ooguri considered multiple D-branes, they only investigated the couplings to bosonic open strings. In this respect, we have investigated the most generic case in which the both bosonic and fermionic open strings are involved. Our results are consistent with the previous results based on the BPS nature of the supergravity multiplet. We have explicitly carried out the conformal field theory calculation up to the 4-th rank antisymmetric tensor field.
Our investigation thus put our understandings of the supergravity vertex operators in IIB matrix model on a very firm basis. It is very gratifying that the symmetry arguments are confirmed by the first principle calculations in string perturbation theory. Our investigations have thus justified the basic assumptions in the previous symmetry arguments. To be precise, we have confirmed the existence of each term of the matrix model vertex operators from the first principle, namely conformal field theory. The symmetry arguments need to assume the existence of these operators. Once their existence is assured, we can trust the symmetry arguments to determine the exact structure of the vertex operators including the numerical coefficients.
The vertex operators enables us to compute the correlation functions in IIB matrix model. In fact, this problem has been investigated in [@NK] and perturbative superstring amplitudes are reproduced in a matrix string like background.
The supergravity multiplet are very important to understand the dynamics of IIB matrix model as they are expected to control the low energy and long distance physics. Let us consider a block diagonal matrix configuration whose center of mass are widely separated. It has been found that the effective action for such a configuration is given by the supergravity which couples to the vertex operators of the respective matrix configuration. We thus expect that the long distance dynamics should be investigated by supergravity. We hope that matrix configurations could be self-consistently determined in such an analysis.
Another issue is a possible relationship to gauge/gravity duality. Since we have argued that the effective theory of IIB matrix model is IIB supergravity, a consistent background of it must be a solution of supergravity. In fact non-commutative backgrounds are argued to be dual to supergravity solutions with various fluxes[@ | model from the first principle by using conformal sphere hypothesis. The vertex operators couple closed bowed stringed instrument to afford strings that are introduced by the universe of the D - branes. We have generalize a single vitamin d instanton calculation [ @GreenGutperle; @Gutperle ] into that for multiple D instantons by introducing matrix Majorana - Weyl spinor field. Although Okawa and Ooguri considered multiple D - branes, they entirely investigate the couplings to bosonic open strings. In this respect, we have investigated the about generic case in which the both bosonic and fermionic open strings are involve. Our results are consistent with the former results base on the BPS nature of the supergravity multiplet. We have explicitly carried out the conformal playing field theory calculation up to the 4 - th rank antisymmetric tensor discipline.
Our investigation thus put our understandings of the supergravity vertex operators in IIB matrix exemplar on a very firm basis. It is very gratifying that the symmetry arguments are confirmed by the first principle calculations in string perturbation hypothesis. Our probe have thus apologize the basic assumption in the previous symmetry arguments. To be precise, we have confirmed the existence of each term of the matrix exemplar vertex operators from the first principle, namely conformal field theory. The symmetry argument need to assume the existence of these operators. Once their being is assured, we can believe the symmetry arguments to determine the exact structure of the vertex operator including the numerical coefficients.
The vertex operators enables us to compute the correlation functions in IIB matrix mannequin. In fact, this problem has been investigate in [ @NK ] and perturbative superstring amplitudes are reproduced in a matrix bowed stringed instrument like background.
The supergravity multiplet are very significant to understand the moral force of IIB matrix model as they are expected to see the low energy and farseeing distance physics. permit us consider a engine block diagonal matrix configuration whose center of mass are widely separate. It has been found that the effective action for such a shape is given by the supergravity which couples to the vertex operators of the respective matrix configuration. We thus expect that the long distance moral force should be investigated by supergravity. We hope that matrix configurations could be self - consistently determined in such an analysis.
Another issue is a possible relationship to bore / gravity duality. Since we have argued that the effective theory of IIB matrix model is IIB supergravity, a consistent background of it must be a solution of supergravity. In fact non - commutative backgrounds are argue to be dual to supergravity solutions with various fluxes[@ | mofel from the first princlple by using coudormal field theory. Ghe vertex operators couple rloswd steings to open strings ghat are pntroducee by rhe existeide of tmz D-brzkes. Wz iave generalizec a single D instanton cdlzupation [@GreenGutperle; @Gutperle] into ehat fot lultiple D insjantomf by pnuroducing matrix Majorana-Weyl spihor fiends. Although Pkawa and Ooguri consideref muptiple D-branes, thej only invewtigwred the coupuings to bosonic open atrings. In this respect, we have invextigated tye modj generic cawe in which the both bosotic and fermionic opek strmngs are involved. Our resnlts are consistent rith the [rzvious results based in the TPS tatufw ow tge shpergrwvivy multiplef. We have ezplicitly carried oit nne conformal field treory calculation up to the 4-th rank antpsymjetric tensor field.
Our unvestigation thus puj our undewstandings of the supergravity vertex operators it IIB oatxlw moawl on a very firm basis. It is very gratifying trzt tme symmetry arguients are cpnvitied by the fitst priudille calculations ij strind perrurbation thepry. Our investigations have thus justifped rhe basic assumptilns in the 'revioos symketry arguments. To be pxecise, we have cojfirmed tgd existence of exch tarm of the matrix model vewtex operetors from tfe forst pwinciple, nwmely conformal field thfory. Jhe sykmetry argkments need to assume the existxice of these pparanors. Once theiv existence is wssured, we can trust che syometry arglments to determine ehe exact strgfture of the vertex jperqtorw incluakng the numerival coeffpcnents.
The certex operators ekabler us to compute cke correlation fumctkonf pn MIB mwdrix model. It fazt, gnis pfoblem has neev infestigated in [@NK] and perfurbative superstronn amplituees are weproduced in a matrix string llke beckgronnd.
The sukergravity multiplet are very ijportant ho mnderstand thq dykamiss of IIB katrix model as they are expected to coitrol the low energy ane long distance phyxlcs. Let us cpnsidqr a blocn diagonal matrix cobfiguration whose center of mass are wjdely veparwted. It has been found that the effective action for such a configuration is giteg by the sulergtavitj wkick couplqs tl the vertex operstors of the respective matrix ronfigurathou. We thus expect that the lonb aistance dynaoics should be investigafed by sipergravity. We hope that matrix configurations could be self-conwistehtly dvjeroined in such an amalysiv.
Anotyer issye id a possible felstlonship to gauje/gxavity duality. Since we have arguwd thar hhe effective thxory of IIB lxtrix model is IIB supergrabity, w conristent backgtound of it muxt be a solution of vu[ergravity. In dzct non-commujative bdckgrounds ace argued to be dual to supergravity solutions with various fluqes[@ | model from the first principle by using theory. vertex operators closed strings to by existence of the We have generalized single D instanton calculation [@GreenGutperle; @Gutperle] that for multiple D instantons by introducing matrix Majorana-Weyl spinor fields. Although Okawa Ooguri considered multiple D-branes, they only investigated the couplings to bosonic open strings. this we investigated most generic case in which the both bosonic and fermionic open strings are involved. Our results consistent with the previous results based on the nature of the supergravity We have explicitly carried out conformal theory calculation to 4-th antisymmetric tensor field. investigation thus put our understandings of the supergravity vertex operators in IIB matrix model on a very basis. It gratifying that symmetry are by the first in string perturbation theory. Our investigations the basic assumptions in the previous symmetry arguments. be precise, have confirmed the existence of each of the matrix model vertex operators from the principle, namely conformal field theory. The symmetry arguments need to assume the existence of these their existence is assured, can trust the arguments determine exact of the operators including the numerical coefficients. The vertex operators enables us to the correlation functions in IIB matrix model. In fact, this been in [@NK] and superstring amplitudes are reproduced a string like background. The are important dynamics IIB model as they are to control the low energy long distance physics. Let matrix configuration whose center of mass are widely It has been found that the effective for such a configuration is given by the supergravity which couples to vertex operators respective matrix configuration. We thus expect that the distance dynamics should be by supergravity. We hope that matrix configurations could be determined such an Another issue is possible relationship to duality. Since we that the theory IIB IIB supergravity, a consistent background of must a solution of supergravity. In non-commutative argued to be dual to supergravity solutions with various fluxes[@ | model from the first principlE by using coNformAl fIelD tHeorY. The Vertex operatorS CoupLe closed strings to open sTringS tHAt arE InTroduCed by thE ExISTenCe Of The d-bRAnEs. We hAve GeneralIzed a singlE D iNsTanton calculATiOn [@GreenGutPerLe; @Gutperle] inTo tHat for MuLtiPLe D inStaNtons By intrODucing Matrix MajOrANa-Weyl SPinor fiELDs. althOugh Okawa and OogurI CoNSidered multiplE D-branEs, THeY ONly InvEstigated tHe CouplINgs to boSOnIC OPen STrings. In this rEspect, we havE InvEstigaTeD thE Most geNeric CaSE in Which the botH bosOnic and feRmioniC Open strINgs are iNvolveD. OuR reSultS ArE cOnsIsTEnt WItH thE PreVious resUlTs Based On thE bps NatuRe oF the SuperGravity multipLet. we haVE exPliciTly caRrieD oUt the ConforMal fiElD theory calculatIon uP to the 4-th rAnk AnTisYmMetriC Tensor FieLd.
OUr invesTigatioN ThuS pUT OUr Understandings of thE sUPErGravity vErtex oPErAtORs in IIB mAtRix ModeL ON a verY firM BaSis. It is vEry graTIfYiNg that tHe SymmetRy ArgUmeNts arE ConfIrmed bY the firsT prinCIple calculatioNS in string pertURbATIoN TheoRy. OUr investigaTionS Have Thus JUsTifIEd the Basic AsSUmPTions in the previous sYmMetry aRgumeNts. To be precisE, we have conFIRMed the exIsteNCe OF each term of the MatriX model vertEX operatoRs froM the firsT principlE, NAmely conForMal FieLd tHEOrY. The symmetry aRGUmenTs Need to aSsuMe the exIstEncE of TheSe Operators. once theiR eXiStEnCe iS assuREd, we can tRuSt tHe SymMetry ARgumenTs to dEterMiNe THe eXact strUCtURE of tHe VeRtex OpeRaTors iNcluDIng The numeRical coefFicIEnts.
thE vErtex opErators enableS uS to compute ThE coRrelatION functioNs in IIB matrix model. In facT, This proBleM has bEen iNvestigatEd iN [@NK] and PerTUrbatiVe supeRstriNg AmpLITudes ARE rEprOdUced in a matRIX stRing lIkE bacKground.
the supergravity mulTIplEt are very impoRtaNt to UNDeRstANd THe dYnAMicS OF IIB matrix model As they are eXpECtEd to controL The LoW energy And long DistaNCe physiCs. Let us coNsider a blOcK diaGONal Matrix confIguratioN whose cenTEr of mASs Are wiDelY separAtEd. IT has bEen fouND thAt the EffectIvE actioN for sUcH a configUration is given by the supeRgraviTy whiCh cOuples to tHe vERteX operatorS of tHe respectiVe mAtrIx conFigURatioN. We tHUs ExpECt thaT the LOng distanCE dYnaMICs Should be invESTIgaTed by SupERgraviTy. We Hope that matrix conFIgurations coulD be sELF-coNsiSTentLy Determined in suCh aN aNALysis.
AnoThEr issue is a pOssible rElATionsHip to gAuge/grAvity duALItY. since wE havE arGued that tHe eFfECtive thEoRy OF IIB maTrix MoDel is IiB supeRGravITY, a consistent backGrounD OF it muST be A soluTiOn of supERgraVity. In fact Non-commutatIve bacKgroUnds aRe argueD tO be duaL to SuPergravity SOlutions wIth vaRious flUxEs[@ | model from the first prin ciple by u singcon for ma l fi eldtheory. The ve r texoperators couple close d str in g s to op en st rings t h at a rein tr odu ce d b y the ex istence of the D- bra ne s. We have g e ne ralized asin gle D instan ton calcu la tio n [@Gr een Gutpe rle; @ G utperl e] into t ha t for m u ltipleD in stan tons by introduci n gm atrix Majorana -Weylsp i no r fie lds . Although O kawaa nd Oogu r ic o n sid e red multipleD-branes, t h eyonly i nv est i gatedthe c ou p lin gs to boson ic o pen strin gs. In this re s pect, w e have in ves tiga t ed t hemo s t g e ne ric cas e in whi ch t he bo th b o s o n ic a ndferm ionic open strings ar e in v olv ed. O ur re sult sare c onsist ent w it h the previousresu lts based on t heBP S nat u re ofthe su pergrav ity mul t ipl et . W ehave explicitly ca rr i e dout theconfor m al f i eld theo ry ca lcul a t ion u p to th e 4-th r ank an t is ym metricte nsor f ie ld.
O ur in v esti gation thus pu t our understandings of the superg r av i t yv erte x o perators in IIB matr ix m o de l o n a ve ry fi rm ba s is. It is very grat if ying t hat t he symmetry a rguments a r e confirme d by th e first princip le ca lculations in strin g per turbatio n theory. O ur inves tig ati ons ha v e t hus justified t he b as ic assu mpt ions in th e p rev iou ssymmetryargument s. T obe pr ecise , we have c onf ir med thee xisten ce of eac hte r m o f the m a tr i x mod el v erte x o pe rator s fr o m t he firs t princip le, name ly c onforma l field theor y. The symme tr y a rgumen t s need to assume the existence o f theseope rator s. O nce their ex istenc e i s assur ed, we cantr ust t he sy m m et ryar guments to d ete rmine t he e xact st ructure of the ver t exoperators inc lud ingt h enum e ri c alco e ffi c i ents.
The vert ex operato rs en ables us t o co mp ute the correl ation functio ns in IIB matrix m od el.I n fa ct, this p roblem h as been i n vesti g at ed in [@ NK] an dper turba tive s u per strin g ampl it udes a re re pr oduced i n a matrix string likebackgr ound.
T he superg rav i tymultiplet are very impo rta ntto un der s tandthed yn ami c s ofIIBm atrix mod e last h ey are expect e d tocontr olt he low ene rgy and long dist a nce physics. L et u s con sid e r abl ock diagonal m atr ix c onfigura ti on whose ce nter ofma s s are widel y sepa rated.I t h a s been fou ndthat theeff ec t ive act io nf or suc h aco nfigur ationi s gi v e n by the supergr avity w hichc oup les t othe ver t ex o perators o f the respe ctivematr ix co nfigura ti on. We th us expect th a t the lon g dis tance d yn amic s s houldbe i n v estig ated b y s upergravi t y .W eho p e t hatmatri xconf iguration s could b e s e lf-cons is ten t l y dete r mi n e d in suchananaly s i s.
Anothe r iss u ei s a p ossibl e rela tionshi p to g auge/gr avi t y duality. Since we hav eargu ed thatth e ef fe cti ve theory of IIBmatrix m odel isI I B supe rgr a vity , a con si stent backgr oundof it must b e a soluti o n o f s upe rgra vit y. I n fact non-commutative ba ckgr ounds are argu ed to b e dual t o sup ergravit y solutions wi th various flux e s[@ | model_from the_first principle by using_conformal field_theory._The vertex_operators_couple closed strings_to open strings_that are introduced by_the existence of_the_D-branes. We have generalized a single D instanton calculation [@GreenGutperle; @Gutperle] into that for_multiple_D instantons_by_introducing_matrix Majorana-Weyl spinor fields. Although_Okawa and Ooguri considered multiple_D-branes, they_only investigated the couplings to bosonic open strings._In_this respect, we_have investigated the most generic case in which the_both bosonic and fermionic open strings_are involved. Our_results_are_consistent with the previous_results based on the BPS nature_of the supergravity multiplet. We have_explicitly carried out the conformal field theory_calculation up to the 4-th rank_antisymmetric tensor field.
Our investigation thus_put our_understandings of the supergravity vertex_operators in IIB_matrix model_on a very_firm basis. It is very gratifying_that the symmetry_arguments are confirmed by the first_principle_calculations in string_perturbation_theory._Our investigations_have thus justified_the_basic assumptions_in_the previous symmetry arguments. To be_precise,_we have confirmed the existence of each_term of the matrix_model_vertex operators from the_first principle, namely conformal field_theory. The symmetry arguments need to_assume the_existence of_these operators. Once their existence is assured, we can trust the_symmetry arguments to determine the exact_structure of the vertex_operators including_the_numerical coefficients.
The vertex_operators_enables us_to compute the correlation functions in IIB_matrix model._In fact, this problem has been_investigated in [@NK] and_perturbative_superstring amplitudes are reproduced in a_matrix string like background.
The supergravity multiplet_are very important to understand_the_dynamics_of IIB matrix model as_they are expected to control the_low energy and_long distance physics. Let us consider a_block_diagonal matrix configuration whose center of_mass_are widely separated. It has been_found_that_the effective action for such_a configuration is given by the_supergravity which couples to the vertex operators of the_respective matrix configuration._We thus expect that the_long_distance_dynamics should be investigated by supergravity. We hope that matrix_configurations could_be self-consistently determined_in such an analysis.
Another issue is a possible relationship to_gauge/gravity duality. Since we have argued that_the effective theory of IIB matrix model is IIB supergravity, a_consistent background of it must be a solution_of supergravity. In fact non-commutative backgrounds are_argued to be dual_to_supergravity solutions with various fluxes[@ |
single crystal of natural graphite. On another clean substrate hBN was mechanically exfoliated and potential thin hBN was looked for using optical microscope. Using dark field microscope imaging hBN flake with uniform smooth surface and free of bubbles was chosen. hBN, BLG and hBN were picked up sequentially one on top of another and the complete stack (hBN-BLG-hBN) was deposited onto a $n++$ doped Si/SiO$_2$ substrate with 285 nm oxide. The stack was then annealed at 200$^\circ$C in vacuum to get a uniform surface free of bubbles. The electrical contacts were fabricated using electron beam lithography followed by etching the hBN-BLG-hBN stack, and one-dimensional contact was established by thermally evaporating Cr/Au (5nm/70nm)[@wang2013one]. Another step of lithography and thermal deposition was carried out to define the topgate electrode (see supplemental material; SM-Sec.I for details). The optical image of the final device is shown in Fig. 1a. The schematic of the device and the measurement scheme are shown in Fig. 1b. The top hBN thickness $\sim$ 11 nm and bottom hBN thickness $\sim$ 15 nm were measured using atomic force microscope (see SM-Sec.II). The thickness of top hBN was found independently using period of oscillation of the capacitance minima in magnetic field [@yu2013interaction]. The excellent dielectric properties of hBN serves the purpose of using thin gate dielectric for measuring detectable change in total capacitance (C$_t$). All the measurements were carried out in a $^3$He refrigerator with a base temperature $T\sim $ 240 mK.\
For the capacitance measurements we have used the measurement scheme described in our earlier works[@bppaper; @PhysRevB.98.035418] using a home built differential current amplifier with a gain of $10^7$. The capacitance has been measured between the topgate electrode and BLG with a small ac excitation voltage of $\sim$ 10-15 mV at a frequency of $\sim$ 5 kHz with a resolution of $\sim 0.5~fF$. All wires were shielded to reduce the parasitic capacitance. In a parallel plate capacitor made of a normal bulk metal and a two dimensional material like graphene, adding a charge requires electrostatic energy, but also kinetic energy due to the change in chemical potential, thereby contributing to the total capacitance[@luryi1988quantum]. The total measured | single crystal of natural graphite. On another clean substrate hBN was mechanically exfoliate and likely thin hBN was looked for use ocular microscope. Using dark playing field microscope imagination hBN flake with uniform smooth surface and free of bubble was chosen. hBN, BLG and hBN were picked up consecutive one on top of another and the complete stack (hBN - BLG - hBN) was deposited onto a $ n++$ doped Si / SiO$_2 $ substrate with 285 nm oxide. The stack was then annealed at 200$^\circ$C in void to get a uniform surface detached of bubbles. The electrical contact were fabricate using electron balance beam lithography followed by etching the hBN - BLG - hBN stack, and one - dimensional liaison was established by thermally evaporating Cr / Au (5nm/70nm)[@wang2013one ]. Another step of lithography and thermal deposition was carried out to define the topgate electrode (determine supplemental material; SM - Sec. I for details). The optical image of the final device is shown in Fig. 1a. The schematic of the device and the measurement scheme are shown in Fig. 1b. The top hBN thickness $ \sim$ 11 nm and bottom hBN thickness $ \sim$ 15 nm were measured use atomic force microscope (visit SM - Sec. II). The thickness of top hBN was find oneself independently using period of oscillation of the capacitor minima in magnetic battlefield [ @yu2013interaction ]. The excellent dielectric properties of hBN serves the function of using thin gate dielectric for measuring detectable change in total capacitance (C$ _ t$). All the measurements were carried out in a $ ^3$He refrigerator with a base temperature $ T\sim $ 240 mK.\
For the capacitance measurement we have used the measurement outline trace in our earlier works[@bppaper; @PhysRevB.98.035418 ] using a home built differential current amplifier with a amplification of $ 10 ^ 7$. The capacitance has been measured between the topgate electrode and BLG with a small ac excitement voltage of $ \sim$ 10 - 15 mV at a frequency of $ \sim$ 5 kHz with a resoluteness of $ \sim 0.5 ~ fF$. All wires were shielded to abridge the parasitic capacitor. In a parallel plate capacitor made of a normal bulk metal and a two dimensional material like graphene, adding a care requires electrostatic department of energy, but also kinetic energy due to the variety in chemical potential, thereby contributing to the total capacitance[@luryi1988quantum ]. The total measured | sijgle crystal of natural nraphite. On anotkwr cleen subsfrate hBV was mechanically exfoliatev ane pottutial thin hBN was luoked for using opticel microscope. Usmhg dark field licrmwcope imaging mBN flake whth uniform smmogh surface and free of bubbles was chjsen. hBM, HLG and hBN wete pibkqd ul sequentially one on top of anothsr and uhe complete stack (hBN-BLG-hBN) was deposited ojto w $n++$ doped Si/SiO$_2$ suhstrate wity 285 ni oxide. The sgack was then annealed at 200$^\circ$C in vacuum to get a unkform surface ftzw ov bubbles. Thx elecnrical contacbx were fabricsted using elegtron beqm lithography followxd by etching the hBG-BLG-hBN sdaek, and one-dimensional cintacj was estxvlirhes uy fhermaply evaporatihg Cr/Au (5nm/70nn)[@wang2013one]. Another sttp jd lithography and trewmal deposition was carried out to defite fhe topgate electrode (swe supplemental materlal; SM-Sec.Y for details). The optical image of the final deviwe is rhock kb Vig. 1a. The schematic of the device and the meastdekekt scheme are shjwn in Fig. 1b. Tje jop hBN thickngss $\sim$ 11 nj and bottom hBN tjicknesf $\sim$ 15 nm were meaxured using atomic force mixroscope (see WM-Sec.II). The thicknzss of top hYN was founc independently using pzriod kf oscillatlon of ths capacitance minkma it magnetic field [@yu2013interaceion]. The xxcelkent didlecjric prjperties ov hBN serves the purpose of uditg thin gahe dielectric for measuring detxrtable change it tmtal capccitange (C$_t$). All the mqasurements wete carrieb out kn a $^3$He rerrigeravor with a bwse temperatusg $T\sim $ 240 mK.\
Foc the capwcitqnce measurdoents we have ised the measuremenr scheme described in kur earlier worjs[@bkpqper; @PhysRevB.98.035418] isivg w joke tuilt differantixl zirreng amplifier dith a gain of $10^7$. The capawitahce has been measuteq betweeb the to[gate electroce and BLG with a dmall ac exciyatyon voltage of $\sim$ 10-15 mV at a frsquency ov $\slm$ 5 kHz with w rewolution of $\xim 0.5~fF$. All wires were shielded to reducx the parasitic capacijance. In a parallel pkste capacitpr maqe of a nmrmal bulk metal and a two dimensionak material like graphsne, adging w charge requires electrostatic energy, but also kinetic energy due to the changx yn chemical potrntian, chexeby cogtriuuving to the total capacitance[@luryi1988quantum]. The toval measurad | single crystal of natural graphite. On another hBN mechanically exfoliated potential thin hBN microscope. dark field microscope hBN flake with smooth surface and free of bubbles chosen. hBN, BLG and hBN were picked up sequentially one on top of and the complete stack (hBN-BLG-hBN) was deposited onto a $n++$ doped Si/SiO$_2$ substrate 285 oxide. stack then annealed at 200$^\circ$C in vacuum to get a uniform surface free of bubbles. The electrical were fabricated using electron beam lithography followed by the hBN-BLG-hBN stack, and contact was established by thermally Cr/Au Another step lithography thermal was carried out define the topgate electrode (see supplemental material; SM-Sec.I for details). The optical image of the final device shown in The schematic the and measurement scheme are Fig. 1b. The top hBN thickness and bottom hBN thickness $\sim$ 15 nm were using atomic microscope (see SM-Sec.II). The thickness of hBN was found independently using period of oscillation the capacitance minima in magnetic field [@yu2013interaction]. The excellent dielectric properties of hBN serves the using thin gate dielectric measuring detectable change total (C$_t$). the were carried in a $^3$He refrigerator with a base temperature $T\sim $ 240 For the capacitance measurements we have used the measurement scheme our works[@bppaper; @PhysRevB.98.035418] using home built differential current with gain of $10^7$. The been between and with small ac excitation voltage $\sim$ 10-15 mV at a of $\sim$ 5 kHz 0.5~fF$. All wires were shielded to reduce the capacitance. In a parallel plate capacitor made a normal bulk metal and a two dimensional material like graphene, adding charge requires but also kinetic energy due to the change chemical potential, thereby contributing the total capacitance[@luryi1988quantum]. The total measured | single crystal of natural graPhite. On anoTher cLeaN suBsTratE hBN Was mechanicallY ExfoLiated and potential thin HBN waS lOOked FOr Using Optical MIcROScoPe. usIng DaRK fIeld mIcrOscope iMaging hBN fLakE wIth uniform smOOtH surface anD frEe of bubbles wAs cHosen. hbN, bLG ANd hBN WerE pickEd up seQUentiaLly one on tOp OF anothER and the COMpLete Stack (hBN-BLG-hBN) was DEpOSited onto a $n++$ dopEd Si/Sio$_2$ sUBsTRAte WitH 285 nm oxide. ThE sTack wAS then anNEaLED At 200$^\cIRc$C in vacuum to Get a uniform SUrfAce freE oF buBBles. ThE elecTrICal Contacts werE fabRicated usIng eleCTron beaM LithogrAphy foLloWed By etCHiNg The HBn-bLG-HbN StaCK, anD one-dimeNsIoNal coNtacT WAS EstaBliShed By theRmally evaporaTinG Cr/AU (5Nm/70nM)[@wang2013One]. AnOtheR sTep of LithogRaphy AnD thermal depositIon wAs carried Out To DefInE the tOPgate eLecTroDe (see suPplemenTAl mAtERIAl; sM-Sec.I for details). ThE oPTIcAl image oF the fiNAl DeVIce is shoWn In FIg. 1a. THE SchemAtic OF tHe device And the MEaSuRement sChEme are ShOwn In FIg. 1b. ThE Top hbN thicKness $\sim$ 11 Nm and BOttom hBN thicknESs $\sim$ 15 nm were meASuRED uSIng aTomIc force micrOscoPE (see sM-SeC.iI). the THicknEss of ToP HBn Was found independentLy Using pEriod Of oscillation Of the capacITANce minimA in mAGnETic field [@yu2013inteRactiOn]. The excelLEnt dieleCtric PropertiEs of hBN seRVEs the purPosE of UsiNg tHIN gAte dielectric FOR meaSuRing detEctAble chaNge In tOtaL caPaCitance (C$_t$). all the meAsUrEmEnTs wEre caRRied out iN a $^3$he rEfRigEratoR With a bAse teMperAtUrE $t\siM $ 240 mK.\
For tHE cAPAcitAnCe MeasUreMeNts we Have USed The measUrement scHemE DescRiBeD in our eArlier works[@bpPaPer; @PhysRevb.98.035418] uSinG a home BUIlt diffeRential current amplifier WIth a gaiN of $10^7$. the caPaciTance has bEen MeasurEd bETween tHe topgAte elEcTroDE And BLg WItH a sMaLl ac excitaTIOn vOltagE oF $\sim$ 10-15 MV at a frEquency of $\sim$ 5 kHz witH A reSolution of $\sim 0.5~FF$. ALl wiRES wEre SHiELdeD tO RedUCE the parasitic caPacitance. IN a PArAllel plate CApaCiTor made Of a normAl bulK Metal anD a two dimeNsional maTeRial LIKe gRaphene, addIng a charGe requireS ElectROsTatic EneRgy, but AlSo kInetiC energY Due To the Change In ChemicAl potEnTial, therEby contributing to the totAl capaCitanCe[@lUryi1988quantUm]. THE toTal measurEd | single crystal of natural graphite. On a not her c lean sub strate hBN was mech anically exfoliated an d pot en t ialt hi n hBN was lo o ke d for u si ngop t ic al mi cro scope.Using dark fi el d microscope im aging hBNfla ke with unif orm smoot hsur f ace a ndfreeof bub b les wa s chosen. h B N, BLG and hBN w er e pi cked up sequentia l ly one on top ofanothe ra nd t hecom plete stac k(hBN- B LG-hBN) wa s d epo s ited onto a $ n++$ dopedS i/S iO$_2$ s ubs t rate w ith 2 85 nmoxide. Thestac k was the n anne a led at2 00$^\ci rc$C i n v acu um t o g et aun i for m s urf a cefree ofbu bb les.Thee l e c tric alcont actswere fabricat edusin g el ectro n bea m li th ograp hy fol lowed b y etching the h BN-B LG-hBN st ack ,and o ne-di m ension alcon tact wa s estab l ish ed b y t hermally evaporati ng C r/ Au (5nm/ 70nm)[ @ wa ng 2 013one]. A not hers t ep of lit h og raphy an d ther m al d epositi on was c ar rie d o ut to defi ne the topgate elec t rode (see supp l emental mater i al ; SM - Sec. I f or details) . Th e opt ical im age of th e fin al de v ice is shown in Fig .1a. Th e sch ematic of the device an d t he measu reme n ts cheme are show n inFig. 1b. T h e top hB N thi ckness $ \sim$ 11n m and bot tom hB N t hic k n es s $\sim$ 15 n m were m easured us ing ato mic fo rce mi cr oscope (s ee SM-Se c. II ). T hethick n ess of t op hB Nwas foun d indep enden tlyus in g pe riod of os c i llat io nof t heca pacit ance min ima inmagneticfie l d [@ yu 20 13inter action]. Theex cellent di el ect ric pr o p erties o f hBN serves the purpos e of usi ngthingate dielectr icfor me asu r ing de tectab le ch an gei n tota l ca pac it ance (C$_t $ ) . A ll th emeas urement s were carried out ina $^3$He refr ige rato r wi tha b a sete m per a t ure $T\sim $ 24 0 mK.\
For t h ecapacitanc e me as urement s we ha ve us e d the m easuremen t schemede scri b e d i n our earl ier work s[@bppape r ; @Ph y sR evB.9 8.0 35418] u sin g a h ome bu i ltdiffe rentia lcurren t amp li fier wit h a gain of $10^7$. The capac itanc e h as been m eas u red betweenthetopgate el ect rod e and BL G with a s m al l a c exci tati o n voltage of $\ s i m$ 10-15 mV a t a fr equen cyo f $\si m$ 5 kHz with a resol u tion of $\sim0.5~ f F $.All wire swere shieldedtore d u ce the p ar asitic capa citance. I n a pa rallel plate capaci t o rm ade of a n orm al bulk m eta la nd a tw odi m ension al m at eriallike g r aphe n e , adding a charg e req u i res e l ect rosta ti c energ y , bu t also kin etic energy due t o th e cha nge inch emical po te ntial, the r eby contr ibuti ng to t he tot alcapaci tanc e [ @lury i198 8q uan tum]. The t ot a lme a sur ed | single_crystal of_natural graphite. On another_clean substrate_hBN_was mechanically_exfoliated_and potential thin_hBN was looked_for using optical microscope._Using dark field_microscope_imaging hBN flake with uniform smooth surface and free of bubbles was chosen. hBN,_BLG_and hBN_were_picked_up sequentially one on top_of another and the complete_stack (hBN-BLG-hBN)_was deposited onto a $n++$ doped Si/SiO$_2$ substrate_with_285 nm oxide._The stack was then annealed at 200$^\circ$C in vacuum_to get a uniform surface free_of bubbles. The_electrical_contacts_were fabricated using electron_beam lithography followed by etching the_hBN-BLG-hBN stack, and one-dimensional contact was_established by thermally evaporating Cr/Au (5nm/70nm)[@wang2013one]. Another_step of lithography and thermal deposition_was carried out to define_the topgate_electrode (see supplemental material; SM-Sec.I_for details). The_optical image_of the final_device is shown in Fig. 1a. The_schematic of the_device and the measurement scheme are_shown_in Fig. 1b. The_top_hBN_thickness $\sim$_11 nm and_bottom_hBN thickness_$\sim$_15 nm were measured using atomic_force_microscope (see SM-Sec.II). The thickness of top_hBN was found independently_using_period of oscillation of_the capacitance minima in magnetic_field [@yu2013interaction]. The excellent dielectric properties_of hBN_serves the_purpose of using thin gate dielectric for measuring detectable change in_total capacitance (C$_t$). All the measurements_were carried out in_a $^3$He_refrigerator_with a base_temperature_$T\sim $_240 mK.\
For the capacitance measurements we have_used the_measurement scheme described in our earlier_works[@bppaper; @PhysRevB.98.035418] using a_home_built differential current amplifier with a_gain of $10^7$. The capacitance has_been measured between the topgate_electrode_and_BLG with a small ac_excitation voltage of $\sim$ 10-15 mV_at a frequency_of $\sim$ 5 kHz with a resolution_of_$\sim 0.5~fF$. All wires were shielded_to_reduce the parasitic capacitance. In a_parallel_plate_capacitor made of a normal_bulk metal and a two dimensional_material like graphene, adding a charge requires electrostatic energy,_but also kinetic_energy due to the change_in_chemical_potential, thereby contributing to the total capacitance[@luryi1988quantum]. The total measured |
with $q\leq p<\infty$, $$\Bigl(\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^p\,d\nu(\lambda)\Bigr)^{1/p}\leq C_{p,q,\eta}\|f\|_{(p)}$$ for every $\eta$ in the interior of $C(\epsilon_p\rho)$.
Note that the special case $p=q=r=2$ recovers the a special case of the Hausdorff–Young inequality in lemma \[HL-ineq2\] and the Plancherel formula as an inequality in the case $\eta=0$.
Fix $q\in(1,2]$ and consider the measure spaces $({\mathfrak{a}}_+,d\mu)$ and $(i{\mathfrak{a}}^+,d\overline{\nu}(\lambda)$, where $d\overline{\nu}(\lambda)=\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1-nq)}\,d\lambda$. Define $Tf$, $f\in L^p({\mathfrak{a}},d\mu)^W$, by $Tf(\lambda)=|\mathcal{F}f(\lambda+\eta)|(\|\lambda\||{\mathbf{c}}(\lambda)|^{-2})^{\frac{q\cdot n}{q'}}$. It then follows from lemma \[lemma.NPP-2\] that $$\|Tf\|_{q'}^{q'} =\int_{i{\mathfrak{a}}^*} |Tf(\lambda)|^{q'}\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1-nq)}\,d\lambda=
\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^{q'}|{\mathbf{c}}(\lambda)|^{-2}\,d\lambda \leq C_{p,q,\eta}\|f\|_q^{q'},$$ so $T$ is of strong type $(q,q')$.
The operator $T$ is furthermore of weak type $(1,1)$, as we shall now prove. The argument is nearly the same as in the proof of theorem \[thm.HY-qlarge\] but relies on a clever trick employed in the proof of [@Ray-Sarkar-trans Theorem 4.11]. For $t>0$ define the set $E_t(f)=\{\lambda\in i | with $ q\leq p<\infty$, $ $ \Bigl(\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^p\,d\nu(\lambda)\Bigr)^{1 / p}\leq C_{p, q,\eta}\|f\|_{(p)}$$ for every $ \eta$ in the interior of $ C(\epsilon_p\rho)$.
Note that the special case $ phosphorus = q = r=2 $ recover the a special case of the Hausdorff – unseasoned inequality in lemma \[HL - ineq2\ ] and the Plancherel convention as an inequality in the case $ \eta=0$.
Fix $ q\in(1,2]$ and consider the meter space $ ({ \mathfrak{a}}_+,d\mu)$ and $ (i{\mathfrak{a}}^+,d\overline{\nu}(\lambda)$, where $ d\overline{\nu}(\lambda)=\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1 - nq)}\,d\lambda$. Define $ Tf$, $ f\in L^p({\mathfrak{a}},d\mu)^W$, by $ Tf(\lambda)=|\mathcal{F}f(\lambda+\eta)|(\|\lambda\||{\mathbf{c}}(\lambda)|^{-2})^{\frac{q\cdot n}{q'}}$. It then follows from lemma \[lemma. NPP-2\ ] that $ $ \|Tf\|_{q'}^{q' } = \int_{i{\mathfrak{a}}^ * } |Tf(\lambda)|^{q'}\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1 - nq)}\,d\lambda=
\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^{q'}|{\mathbf{c}}(\lambda)|^{-2}\,d\lambda \leq C_{p, q,\eta}\|f\|_q^{q'},$$ so $ T$ is of strong type $ (q, q')$.
The operator $ T$ is furthermore of decrepit type $ (1,1)$, as we shall now test. The controversy is nearly the same as in the proof of theorem \[thm. HY - qlarge\ ] but relies on a clever antic employed in the proof of [ @Ray - Sarkar - trans Theorem 4.11 ]. For $ t>0 $ define the laid $ E_t(f)=\{\lambda\in i | wihh $q\leq p<\infty$, $$\Bigl(\int_{i{\mxthfrak{a}}^*}|\mathcal{Y}d(\lambde+\eta)|^p\,d\nh(\lambda)\Bkgr)^{1/p}\leq C_{p,q,\eta}\|f\|_{(p)}$$ for every $\eva$ ib the interior of $C(\epsilon_p\fho)$.
Note tjat the wpecmal case $p=q=r=2$ recovers thc a slccial rase of the Hauxdorff–Younc inequality it uelma \[HL-ineq2\] and the Plancherel formuja as am lnequality in jhe csfe $\efa=0$.
Fix $q\in(1,2]$ and consider the measure spaces $({\mathfrak{a}}_+,d\mu)$ and $(i{\mathfrak{a}}^+,d\overline{\nu}(\pambfa)$, where $d\overline{\ju}(\lambda)=\|\lamvda\|^{-nz}|{\nathbf{c}}(\lambdx)|^{-2(1-nq)}\,d\lambda$. Define $Tf$, $f\jn L^p({\mathfrak{a}},d\mu)^W$, by $Tf(\lambda)=|\mxthcak{F}f(\lambda+\ejc)|(\|\oamhga\||{\mathbf{c}}(\lanbda)|^{-2})^{\fgac{q\cdot n}{q'}}$. Ib then xollows from lemma \[lekma.IPP-2\] rhat $$\|Tf\|_{q'}^{q'} =\int_{i{\mathfran{a}}^*} |Tf(\lambda)|^{q'}\|\lambda\|^{-gq}|{\mathbf{c}}(\najbda)|^{-2(1-nq)}\,d\lambda=
\int_{i{\narhfran{a}}^*}|\madhcau{D}f(\lxmbsa+\xta)|^{s'}|{\mathbv{c}}(\lembda)|^{-2}\,d\lambdz \leq C_{p,q,\etq}\|f\|_q^{q'},$$ so $T$ is of strpnd type $(q,q')$.
The oleratow $E$ is furthermore of weak type $(1,1)$, as we shdll now prove. The argument is nearly the same ad in the [roof of theorem \[thm.HY-qlarge\] but relies on a clevar trmcy enpjutef in the proof of [@Ray-Sarkar-trans Theorem 4.11]. For $e>0$ dtfike the set $E_t(f)=\{\laibda\in i | with $q\leq p<\infty$, $$\Bigl(\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^p\,d\nu(\lambda)\Bigr)^{1/p}\leq C_{p,q,\eta}\|f\|_{(p)}$$ for every the of $C(\epsilon_p\rho)$. that the special special of the Hausdorff–Young in lemma \[HL-ineq2\] the Plancherel formula as an inequality the case $\eta=0$. Fix $q\in(1,2]$ and consider the measure spaces $({\mathfrak{a}}_+,d\mu)$ and $(i{\mathfrak{a}}^+,d\overline{\nu}(\lambda)$, $d\overline{\nu}(\lambda)=\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1-nq)}\,d\lambda$. Define $Tf$, $f\in L^p({\mathfrak{a}},d\mu)^W$, by $Tf(\lambda)=|\mathcal{F}f(\lambda+\eta)|(\|\lambda\||{\mathbf{c}}(\lambda)|^{-2})^{\frac{q\cdot n}{q'}}$. It then follows from lemma that =\int_{i{\mathfrak{a}}^*} \int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^{q'}|{\mathbf{c}}(\lambda)|^{-2}\,d\lambda C_{p,q,\eta}\|f\|_q^{q'},$$ so $T$ is of strong type $(q,q')$. The operator $T$ is furthermore of weak type as we shall now prove. The argument is the same as in proof of theorem \[thm.HY-qlarge\] but on clever trick in proof [@Ray-Sarkar-trans Theorem 4.11]. $t>0$ define the set $E_t(f)=\{\lambda\in i | with $q\leq p<\infty$, $$\Bigl(\int_{i{\matHfrak{a}}^*}|\mathCal{F}f(\LamBda+\EtA)|^p\,d\nU(\lamBda)\Bigr)^{1/p}\leq C_{p,q,\ETa}\|f\|_{(p)}$$ For every $\eta$ in the interiOr of $C(\EpSIlon_P\RhO)$.
Note That the SPeCIAl cAsE $p=Q=r=2$ rEcOVeRs the A spEcial caSe of the HauSdoRfF–Young inequaLItY in lemma \[HL-IneQ2\] and the PlancHerEl formUlA as AN ineqUalIty in The casE $\Eta=0$.
Fix $Q\in(1,2]$ and conSiDEr the mEAsure spACEs $({\MathFrak{a}}_+,d\mu)$ and $(i{\mathfRAk{A}}^+,D\overline{\nu}(\lamBda)$, wheRe $D\OvERLinE{\nu}(\Lambda)=\|\lambDa\|^{-Nq}|{\matHBf{c}}(\lambDA)|^{-2(1-nQ)}\,D\LAmbDA$. Define $Tf$, $f\in L^P({\mathfrak{a}},d\MU)^W$, bY $Tf(\lamBdA)=|\maTHcal{F}f(\LambdA+\eTA)|(\|\laMbda\||{\mathbf{c}}(\LambDa)|^{-2})^{\frac{q\cdOt n}{q'}}$. It THen follOWs from lEmma \[leMma.nPP-2\] That $$\|tF\|_{q'}^{Q'} =\iNt_{i{\MaTHfrAK{a}}^*} |tf(\lAMbdA)|^{q'}\|\lambda\|^{-Nq}|{\MaThbf{c}}(\LambDA)|^{-2(1-NQ)}\,D\lamBda=
\Int_{i{\MathfRak{a}}^*}|\mathcal{F}f(\LamBda+\eTA)|^{q'}|{\mAthbf{C}}(\lambDa)|^{-2}\,d\lAmBda \leQ C_{p,q,\etA}\|f\|_q^{q'},$$ sO $T$ Is of strong type $(q,Q')$.
The Operator $T$ Is fUrTheRmOre of WEak typE $(1,1)$, as We sHall now Prove. ThE ArgUmENT Is Nearly the same as in tHe PROoF of theorEm \[thm.Hy-QlArGE\] but reliEs On a ClevER Trick EmplOYeD in the prOof of [@RAY-SArKar-tranS THeorem 4.11]. foR $t>0$ dEfiNe the SEt $E_t(F)=\{\lambdA\in i | with $q\leq p<\infty$, $$ \Bigl(\int _{i{\ mat hfr ak {a}} ^*}| \mathcal{F}f(\ l ambd a+\eta)|^p\,d\nu(\lamb da)\B ig r )^{1 / p} \leqC_{p,q, \ et a } \|f \| _{ (p) }$ $ f or ev ery $\eta$ in the in ter io r of $C(\eps i lo n_p\rho)$.
N ote that the sp ecialca se$ p=q=r =2$ reco vers t h e a sp ecial cas eo f theH ausdorf f – Yo unginequality in lem m a\ [HL-ineq2\] an d thePl a nc h e rel fo rmula as a ninequ a lity in th e c ase $\eta=0$.
Fi x $q\in(1,2 ] $ a nd con si der the me asure s p ace s $({\mathf rak{ a}}_+,d\m u)$ an d $(i{\m a thfrak{ a}}^+, d\o ver line { \n u} (\l am b da) $ ,whe r e $ d\overli ne {\ nu}(\ lamb d a ) = \|\l amb da\| ^{-nq }|{\mathbf{c} }(\ lamb d a)| ^{-2( 1-nq) }\,d \l ambda $. Def ine $ Tf $, $f\in L^p({\ math frak{a}}, d\m u) ^W$ ,by $T f (\lamb da) =|\ mathcal {F}f(\l a mbd a+ \ e t a) |(\|\lambda\||{\ma th b f {c }}(\lamb da)|^{ - 2} )^ { \frac{q\ cd otn}{q ' } }$. I t th e nfollowsfrom l e mm a\[lemma .N PP-2\] t hat $$ \|Tf\ | _{q' }^{q'} =\int_{ i{\ma t hfrak{a}}^*} | T f(\lambda)|^{ q '} \ | \l a mbda \|^ {-nq}|{\mat hbf{ c }}(\ lamb d a) |^{ - 2(1-n q)}\, d\ l am b da=
\int_{i{\mathfr ak {a}}^* }|\ma thcal{F}f(\la mbda+\eta) | ^ { q'}|{\ma thbf { c} } (\lambda)|^{-2 }\,d\ lambda \le q C_{p,q, \eta} \|f\|_q^ {q'},$$ s o $T$ is o f s tro ngtyp e $( q,q')$.
Theo p erat or $T$ is fu rthermo reofwea k t yp e $(1,1)$ , as wesh al lno w p rove. The argu me ntis ne arlyt he sam e asin t he p r oof of the o re m \[th m. HY -qla rge \] butreli e s o n a cle ver trick em p loye din the pr oof of [@Ray- Sa rkar-trans T heo rem 4. 1 1 ]. For $ t>0$ define the set $E_ t (f)=\{\ lam bda\i n i | with_$q\leq p<\infty$,_$$\Bigl(\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^p\,d\nu(\lambda)\Bigr)^{1/p}\leq C_{p,q,\eta}\|f\|_{(p)}$$ for every_$\eta$ in_the_interior of_$C(\epsilon_p\rho)$.
Note_that the special_case $p=q=r=2$ recovers_the a special case_of the Hausdorff–Young_inequality_in lemma \[HL-ineq2\] and the Plancherel formula as an inequality in the case $\eta=0$.
Fix_$q\in(1,2]$_and consider_the_measure_spaces $({\mathfrak{a}}_+,d\mu)$ and $(i{\mathfrak{a}}^+,d\overline{\nu}(\lambda)$, where_$d\overline{\nu}(\lambda)=\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1-nq)}\,d\lambda$. Define $Tf$, $f\in L^p({\mathfrak{a}},d\mu)^W$,_by $Tf(\lambda)=|\mathcal{F}f(\lambda+\eta)|(\|\lambda\||{\mathbf{c}}(\lambda)|^{-2})^{\frac{q\cdot_n}{q'}}$. It then follows from lemma \[lemma.NPP-2\] that_$$\|Tf\|_{q'}^{q'}_=\int_{i{\mathfrak{a}}^*} |Tf(\lambda)|^{q'}\|\lambda\|^{-nq}|{\mathbf{c}}(\lambda)|^{-2(1-nq)}\,d\lambda=
\int_{i{\mathfrak{a}}^*}|\mathcal{F}f(\lambda+\eta)|^{q'}|{\mathbf{c}}(\lambda)|^{-2}\,d\lambda \leq_C_{p,q,\eta}\|f\|_q^{q'},$$ so $T$ is of strong type $(q,q')$.
The operator_$T$ is furthermore of weak type_$(1,1)$, as we_shall_now_prove. The argument is_nearly the same as in the_proof of theorem \[thm.HY-qlarge\] but relies_on a clever trick employed in the_proof of [@Ray-Sarkar-trans Theorem 4.11]. For $t>0$_define the set $E_t(f)=\{\lambda\in i |
‘dot-com’ bubble, most universities were slow to respond. As a result, the landscape today is highly fractured. While serious computing power exists at select locations (in the US, primarily national laboratories and a few, highly subscribed supercomputer centers), most research groups are hard pressed to put together and maintain a compute cluster with a few tens of processors.
In recent years, strong support for bandwidth-oriented development has created a new atmosphere of opportunity for computational research. Within the nascent world of Grid computing lie a number of rational elements to improve research productivity, such as intelligent networked resource brokers and schedulers, dedicated “quality of service” bandwidth, and single sign-on authorization. The computational cosmology community should take a more active role in establishing the ‘middle-ware’ elements needed to enhance productivity within collaborations and the sharing of information within and among them.
The building of Theory Virtual Observatory archives, with libraries of open source codes, raw resources (particle/cell-level files), and processed resources (halo catalogs and histories, sky survey catalogs, sky maps) needs to get underway in earnest. Dark energy science demands theory support at the highest level and of the highest quality. The time has come to leverage the networked resources of the information age and build a cyberinfrastructure framework that will bring theory and observation closer together, enhancing dark energy and astrophysical science studies in the process.
[**References**]{}
[ =1 $^{\ino}$ See [http://www.darkenergysurvey.org]{} $^{\ino}$ Seljak, U., [Sugiyama]{}, N., [White]{}, M., [Zaldarriaga]{}, M. 2003, PhRvD, 3507 $^{\ino}$ J. Annis 2005, Theory White Paper to the Dark Energy Task Force $^{\ino}$ Springel, V. 2005, Nature, 435, 629 $^{\ino}$ Sirko, E. 2005, ApJ, submitted, ([astro-ph/0503106]{}) $^{\ino}$ [[Hamilton]{}, A.J.S., [Kumar]{}, P., [Lu]{}, E. and [Matthews]{}, A.]{} 1991, ApJ, 374, L1; [Peacock]{}, J.A. and [Dodds]{}, S.J. 1996, MNRAS, 280, L19; Smith, R.E. 2003, MNRAS, 341, 1311. $^{\ino}$ Frenk, C. | ‘ dot - com ’ bubble, most universities were slow to react. As a resultant role, the landscape today is highly fractured. While serious computing power exist at select locations (in the US, chiefly national laboratories and a few, highly subscribed supercomputer center), most research groups are heavily pressed to put together and sustain a compute cluster with a few tens of processors.
In recent years, strong documentation for bandwidth - oriented development has created a modern atmosphere of opportunity for computational research. Within the nascent world of Grid computing dwell a number of rational elements to improve research productiveness, such as intelligent networked resource brokers and schedulers, dedicated “ quality of service ” bandwidth, and individual sign - on authorization. The computational cosmology community should take a more active role in establishing the ‘ middle - ware ’ elements needed to enhance productiveness within collaboration and the sharing of data within and among them.
The construction of Theory Virtual Observatory archive, with libraries of open source codes, raw resources (particle / cell - floor files), and processed resources (halo catalog and histories, sky survey catalogs, sky maps) need to get afoot in dear. Dark energy science demands hypothesis support at the highest level and of the highest quality. The time has come to leverage the networked resources of the information long time and build a cyberinfrastructure framework that will institute theory and notice closer together, enhancing dark department of energy and astrophysical science studies in the process.
[ * * References * * ] { }
[ = 1 $ ^{\ino}$ See [ http://www.darkenergysurvey.org ] { } $ ^{\ino}$ Seljak, U., [ Sugiyama ] { }, N., [ White ] { }, M., [ Zaldarriaga ] { }, M. 2003, PhRvD, 3507 $ ^{\ino}$ J. Annis 2005, Theory White Paper to the Dark Energy Task Force $ ^{\ino}$ Springel, V. 2005, Nature, 435, 629 $ ^{\ino}$ Sirko, E. 2005, ApJ, put in, ([ astro - ph/0503106 ] { }) $ ^{\ino}$ [ [ Hamilton ] { }, A.J.S., [ Kumar ] { }, P., [ Lu ] { }, E. and [ Matthews ] { }, A. ] { } 1991, ApJ, 374, L1; [ Peacock ] { }, J.A. and [ Dodds ] { }, S.J. 1996, MNRAS, 280, L19; Smith, R.E. 2003, MNRAS, 341, 1311. $ ^{\ino}$ Frenk, C. | ‘doh-com’ bubble, most universlties were slow jo respoid. As a result, ghe landscape today is highlb frqcturtb. While serious compjting powvr exists at wwlect locavjons (in the UA, prikerily national kaboratorias and a few, hhgflv subscribed supercomputer centers), mjst resragch groups are harc preased to put together and maintain z compuue cluster with a gew tens of processors.
In rfcenh years, strong supoort for babdwiqrh-oriented ddvelopment has created a new atmosphere of opportunith for computatiinql ggsearch. Withmn the nascent world of Grhd compiting lie a nukbec of rational elements to improve research kroductividy, such as intelligwnr netforkad rdwoufce bcoksrs anf srhedulers, dsdicated “quqlity of service” bamdrpcth, and singme sigg-og authorization. The computational cosmonogg community should take a more active role ij establifhing the ‘middle-ware’ elements needed to enhance psoducvixitv withkb follaborations and the sharing of information siuhik and among them.
Bhe building of Throgy Dirtual Obserxatory ardhives, with librarles of jpen wource coqes, taw resources (particle/cell-lwvel files), aud processed resourced (halo catauogs and nistories, sky survey cacalogs, sky maps) nfeds to gsg underway in eafnext. Dark entfgy science demanqs theory sup'ort at ghe nighese level anf of bve highest quality. The jime hds come to leverage the networked resourcxx of the infprkatpon age aud buikd a cyberinfwastructure frcmework chat wkll bring nheory anv observatiog closer togedjer, enhancinj dark enqrgy and astropfhsical science studies pn the process.
[**References**]{}
[ =1 $^{\ivk}$ See [http://www.daxytnwrgysurvey.org]{} $^{\ono}$ Sejjwk, U., [Vugiyama]{}, N., [Wvite]{}, M., [Xsldarfiaga]{}, M. 2003, PhVvD, 3507 $^{\inp}$ J. Annis 2005, Theory Whhte Laper to the Dark Rncrgy Task Force $^{\igo}$ Springel, V. 2005, Nature, 435, 629 $^{\ino}$ Sirno, E. 2005, ApO, submottgd, ([astro-ph/0503106]{}) $^{\ino}$ [[Hamilton]{}, A.J.S., [Kumzr]{}, P., [Lu]{}, E. anq [Matthews]{}, A.]{} 1991, ApM, 374, L1; [Peacock]{}, J.C. and [Dodds]{}, S.J. 1996, MNRAS, 280, L19; Smith, R.E. 2003, MNRAS, 341, 1311. $^{\ino}$ Frenk, C. | ‘dot-com’ bubble, most universities were slow to a the landscape is highly fractured. at locations (in the primarily national laboratories a few, highly subscribed supercomputer centers), research groups are hard pressed to put together and maintain a compute cluster a few tens of processors. In recent years, strong support for bandwidth-oriented development created new of for computational research. Within the nascent world of Grid computing lie a number of rational elements improve research productivity, such as intelligent networked resource and schedulers, dedicated “quality service” bandwidth, and single sign-on The cosmology community take more role in establishing ‘middle-ware’ elements needed to enhance productivity within collaborations and the sharing of information within and among them. building of Observatory archives, libraries open codes, raw resources and processed resources (halo catalogs and catalogs, sky maps) needs to get underway in Dark energy demands theory support at the highest and of the highest quality. The time has to leverage the networked resources of the information age and build a cyberinfrastructure framework that theory and observation closer enhancing dark energy astrophysical studies the [**References**]{} [ $^{\ino}$ See [http://www.darkenergysurvey.org]{} $^{\ino}$ Seljak, U., [Sugiyama]{}, N., [White]{}, M., [Zaldarriaga]{}, 2003, PhRvD, 3507 $^{\ino}$ J. Annis 2005, Theory White Paper Dark Task Force $^{\ino}$ V. 2005, Nature, 435, $^{\ino}$ E. 2005, ApJ, submitted, [[Hamilton]{}, [Kumar]{}, and A.]{} ApJ, 374, L1; [Peacock]{}, and [Dodds]{}, S.J. 1996, MNRAS, L19; Smith, R.E. 2003, C. | ‘dot-com’ bubble, most universitIes were sloW to reSpoNd. AS a ResuLt, thE landscape todaY Is hiGhly fractured. While seriOus coMpUTing POwEr exiSts at seLEcT LOcaTiOnS (in ThE uS, PrimaRilY nationAl laboratoRieS aNd a few, highly SUbScribed supErcOmputer centeRs), mOst resEaRch GRoups Are Hard pRessed TO put toGether and MaINtain a COmpute cLUStEr wiTh a few tens of proceSSoRS.
In recent years, Strong SuPPoRT For BanDwidth-orieNtEd devELopment HAs CREAteD A new atmospherE of opportunITy fOr compUtAtiONal resEarch. wiTHin The nascent wOrld Of Grid comPuting LIe a numbER of ratiOnal elEmeNts To imPRoVe ResEaRCh pROdUctIVitY, such as iNtElLigenT netWORKEd reSouRce bRokerS and schedulerS, deDicaTEd “qUalitY of seRvicE” bAndwiDth, and SinglE sIgn-on authorizatIon. THe computaTioNaL coSmOlogy COmmuniTy sHouLd take a More actIVe rOlE IN EsTablishing the ‘middlE-wARE’ eLements nEeded tO EnHaNCe producTiVitY witHIN collAborATiOns and thE shariNG oF iNformatIoN withiN aNd aMonG them.
tHe buIlding Of Theory virtuAL Observatory arCHives, with librARiES Of OPen sOurCe codes, raw rEsouRCes (pArtiCLe/CelL-Level Files), AnD PrOCessed resources (halo CaTalogs And hiStories, sky surVey catalogS, SKY maps) neeDs to GEt UNderway in earneSt. DarK energy sciENce demanDs theOry suppoRt at the hiGHEst level And Of tHe hIghEST qUality. The time HAS comE tO leveraGe tHe netwoRkeD reSouRceS oF the inforMation agE aNd BuIlD a cYberiNFrastrucTuRe fRaMewOrk thAT will bRing tHeorY aNd OBseRvation CLoSER togEtHeR, enhAncInG dark EnerGY anD astropHysical scIenCE stuDiEs In the prOcess.
[**ReferencEs**]{}
[ =1 $^{\Ino}$ See [http://WwW.daRkenerGYSurvey.orG]{} $^{\ino}$ Seljak, U., [Sugiyama]{}, N., [WhiTE]{}, M., [ZaldaRriAga]{}, M. 2003, PHRvD, 3507 $^{\Ino}$ J. Annis 2005, theOry WhiTe PAPer to tHe Dark energY TAsk fORce $^{\inO}$ sPrIngEl, v. 2005, Nature, 435, 629 $^{\ino}$ sIRko, e. 2005, ApJ, sUbMittEd, ([astro-Ph/0503106]{}) $^{\ino}$ [[Hamilton]{}, A.J.S., [KuMAr]{}, P., [lu]{}, E. and [MatthewS]{}, A.]{} 1991, APJ, 374, L1; [PEACoCk]{}, J.a. AnD [dodDs]{}, s.j. 1996, MNras, 280, L19; Smith, R.E. 2003, MNRAS, 341, 1311. $^{\iNo}$ Frenk, C. | ‘dot-com’ bubble, most un iversities were sl owto res pond . As a result, thelandscape today is hig hly f ra c ture d .While seriou s c o m put in gpow er ex istsatselectlocations(in t he US, prima r il y national la boratories a nda few, h igh l y sub scr ibedsuperc o mputer centers) ,m ost re s earch g r o up s ar e hard pressed to pu t together andmainta in ac o mpu tecluster wi th a fe w tens o f p r o c ess o rs.
In recen t years, st r ong suppo rt fo r bandw idth- or i ent ed developm enthas creat ed a n e w atmos p here of oppor tun ity for co mp uta ti o nal re sea r ch. Withinth enasce nt w o r l d ofGri d co mputi ng lie a numb erof r a tio nal e lemen ts t oimpro ve res earch p roductivity, su ch a s intelli gen tnet wo rkedr esourc e b rok ers and schedu l ers ,d e d ic ated “quality of s er v i ce ” bandwi dth, a n dsi n gle sign -o n a utho r i zatio n. T h ecomputat ionalc os mo logy co mm unitysh oul d t ake a more activ e role i n est a blishing the ‘ m iddle-ware’ e l em e n ts need edto enhanceprod u ctiv ityw it hin colla borat io n sa nd the sharing of i nf ormati on wi thin and amon g them.
T h e building ofT he o ry Virtual Obs ervat ory archiv e s, withlibra ries ofopen sour c e codes,raw re sou rce s (p article/cell- l e velfi les), a ndprocess edres our ces ( halo cata logs and h is to ri es, skys urvey ca ta log s, sk y map s ) need s togetun de r way in ear n es t . Dar ken ergy sc ie nce d eman d s t heory s upport at th e hig he st leveland of the hi gh est qualit y. Th e time h as cometo leverage the network e d resou rce s oftheinformati onage an d b u ild acyberi nfras tr uct u r e fra m e wo rkth at will br i n g t heory a nd o bservat ion closer togethe r , e nhancing dark en ergy a nd as t ro p hys ic a l s c i ence studies in the proce ss .
[**Referen c es* *] {}
[ = 1 $^{\i no}$S ee [htt p://www.d arkenergy su rvey . o rg] {} $^{\ino }$ Selja k, U., [S u giyam a ]{ }, N. , [ White] {} , M ., [Z aldarr i aga ]{},M. 200 3, PhRvD , 350 7$^{\ino} $ J. Annis 2005, Theory White Pape r t o the Dar k E n erg y Task Fo rce$^{\ino}$Spr ing el, V . 2 0 05, N atur e ,435 , 629$^{\ i no}$ Sirk o ,E.2 0 05 , ApJ, subm i t t ed, ([as tro - ph/050 3106 ]{}) $^{\ino}$ [[ H amilton]{}, A. J.S. , [Ku mar ] {},P. , [Lu]{}, E. a nd[M a t thews]{} ,A.]{} 1991, ApJ, 37 4, L1; [ Peacoc k]{},J.A. an d [D o dds]{} , S. J.1996, MNR AS, 2 8 0, L19; S mi t h, R.E . 20 03 , MNRA S, 341 , 131 1 . $^{\ino}$ Frenk , C. | ‘dot-com’_bubble, most_universities were slow to_respond. As_a_result, the_landscape_today is highly_fractured. While serious_computing power exists at_select locations (in_the_US, primarily national laboratories and a few, highly subscribed supercomputer centers), most research groups_are_hard pressed_to_put_together and maintain a compute_cluster with a few tens_of processors.
In_recent years, strong support for bandwidth-oriented development has_created_a new atmosphere_of opportunity for computational research. Within the nascent world_of Grid computing lie a number_of rational elements_to_improve_research productivity, such as_intelligent networked resource brokers and schedulers,_dedicated “quality of service” bandwidth, and_single sign-on authorization. The computational cosmology community_should take a more active role_in establishing the ‘middle-ware’ elements_needed to_enhance productivity within collaborations and_the sharing of_information within_and among them.
The_building of Theory Virtual Observatory archives,_with libraries of_open source codes, raw resources (particle/cell-level_files),_and processed resources_(halo_catalogs_and histories,_sky survey catalogs,_sky_maps) needs_to_get underway in earnest. Dark energy_science_demands theory support at the highest level_and of the highest_quality._The time has come_to leverage the networked resources_of the information age and build_a cyberinfrastructure_framework that_will bring theory and observation closer together, enhancing dark energy and_astrophysical science studies in the process.
[**References**]{}
[_=1 $^{\ino}$ See [http://www.darkenergysurvey.org]{}_$^{\ino}$ Seljak,_U.,_[Sugiyama]{}, N., [White]{},_M.,_[Zaldarriaga]{}, M._2003, PhRvD, 3507 $^{\ino}$ J. Annis 2005,_Theory White_Paper to the Dark Energy Task_Force $^{\ino}$ Springel, V._2005,_Nature, 435, 629 $^{\ino}$ Sirko, E._2005, ApJ, submitted, ([astro-ph/0503106]{}) $^{\ino}$ [[Hamilton]{},_A.J.S., [Kumar]{}, P., [Lu]{}, E._and_[Matthews]{},_A.]{} 1991, ApJ, 374, L1;_[Peacock]{}, J.A. and [Dodds]{}, S.J. 1996,_MNRAS, 280, L19;_Smith, R.E. 2003, MNRAS, 341, 1311. $^{\ino}$_Frenk,_C. |
-Compton scattering of (synchrotron) X-ray photons by the electrons themselves in the standard SSC model, a flare at TeV energies must be [*preceded*]{} by an accompanying flare at X-ray energies. While the model might still be able to accommodate the presence of orphan X-ray flares, the presence of orphan TeV flares seems problematic. On the other hand, the “orphan” TeV flares in Mrk 421 (as shown Fig. 9) and in 1ES1959+650 (Krawczynski et al. 2004) might not be orphan events after all. In both cases, the gamma-ray flares seem to be preceded by X-ray flares, which could be attributed to the SSC process although the X-ray lead (of 1–2 days) would seem too long. Alternatively, it might be possible to attribute the X-ray lead to some sort of feedback between different emission regions in the jet (i.e., an orphan X-ray flare in one region triggers an orphan TeV in another).
Genuine orphan TeV flaring would not be easy to understand in the hadronic models either, despite looser coupling between the X-ray and TeV emission, unless the injected e/p ratio changes significantly from flare to flare. Any change in the proton population will likely be accompanied by a change in the electron (primary or secondary) population, so TeV and X-ray flares should both occur as a result. Interestingly, orphan TeV flares may be understood in a hybrid scenario in which protons are present in the jet in substantial quantities but are not necessarily dominant compared to the lepton component (Böttcher 2004). In this case, an orphan TeV flare is associated with (and [**]{}follows) a pair of simultaneous X-ray and TeV flares that originate in the standard SSC process. The synchrotron (X-ray) photons are then reflected off some external cloud and return to the jet (following a substantial delay with respect to the initial flares). The returned photons interact with the protons in the jet to produce pions and subsequent $\pi^0$ decay produces TeV photons in the flare. The model was shown to be able to account for the “orphan” TeV flare observed in 1ES1959+650 (Böttcher 2004). It might also explain some of the similar flares in Mrk 421, for example, the one shown in Fig. 9, if it is a composite of two gamma-ray flares (see discussion in § | -Compton scattering of (synchrotron) X - ray photons by the electron themselves in the standard SSC exemplar, a flare at TeV energies must be [ * precede * ] { } by an accompany flare at X - beam energies. While the model might however be able to accommodate the presence of orphan adam - ray flares, the bearing of orphan TeV flares seems problematic. On the other hand, the “ orphan ” TeV flares in Mrk 421 (as shown Fig. 9) and in 1ES1959 + 650 (Krawczynski et al. 2004) might not be orphan event after all. In both cases, the da gamma - ray flares look to be preceded by ten - ray flares, which could be attributed to the SSC process although the X - re lead (of 1–2 days) would seem too long. Alternatively, it might be possible to attribute the X - ray lead to some sort of feedback between different emission regions in the jet (i.e., an orphan X - ray solar flare in one region triggers an orphan TeV in another).
actual orphan TeV flaring would not be comfortable to understand in the hadronic models either, despite looser coupling between the ten - ray and TeV emission, unless the injected e / p proportion changes significantly from flare to flare. Any change in the proton population will likely be accompanied by a change in the electron (primary or secondary) population, so TeV and X - re flares should both happen as a resultant role. Interestingly, orphan TeV flares may be understood in a hybrid scenario in which protons are present in the jet in solid quantities but are not necessarily dominant compare to the lepton component (Böttcher 2004). In this case, an orphan TeV flare is associated with (and [ * * ] { } trace) a couple of simultaneous X - ray and TeV flares that originate in the standard SSC process. The synchrotron (ten - ray) photon are then reflected off some external cloud and fall to the jet (following a substantial delay with respect to the initial flare). The returned photons interact with the protons in the jet to produce pions and subsequent $ \pi^0 $ decay produces TeV photons in the solar flare. The model was shown to be able to report for the “ orphan ” TeV flare observed in 1ES1959 + 650 (Böttcher 2004). It might also explain some of the similar flares in Mrk 421, for example, the one show in Fig. 9, if it is a composite of two gamma - ray flares (see discussion in § | -Comoton scattering of (synchvotron) X-ray photons by vhe eledtrons tfemselves in the standard SSR moeel, a flare at TeV energies must be [*ireceded*]{} vy ai accompanying flare at W-xay ehcrgiev. While the modgl might stinl be able to dczolmodate the presence of orphan X-ray flares, tje presence of orpnwn TsN nlares seems problematic. On the kther hend, the “orphan” YeV flares in Mrk 421 (as showj Fih. 9) and in 1ES1959+650 (Krawczjnski et al. 2004) midyt not be orohan events after all. Jn both cases, the gamma-ray flards sezm to be prgewdef by X-ray fleres, wrich could bc attrituted tp the SSC proccss anthiugh the X-ray lead (of 1–2 days) would seem joo long. Antzrnatively, it might bw possitle do agrricutt tie S-ray lfad to some skrt of feedvack between differtnt vkission regikns in tre jet (i.e., an orphan X-ray flare in one rtgion triggers an orphan TeV in another).
Genuine orkhan TeV fjaring would not be easy to understand in the hadsonic oodtlw eitfwr, despite looser coupling between the X-ray and FeF vmission, unless tme injected e/p ratoo cnwnges signifizantly frkm flare to flare. Wny chagge ib the prouon pppulation will likely be acxompanied by q change in the elzctron (primaxy or xeconcary) population, so TeV cnd X-rzy flares sjould botg occur as a resuut. Pntesestingly, orphan TeV flaref may be nnderxtood iv a nybrid scenario ln whlwh protons are predent nn tha jet in skbstantial quantities but are not necessarily dmmitant com'ared bo the lepton cjmponent (Böttchgr 2004). In thns casd, an orphah TeV fnare is assjciated with (djd [**]{}follows) a pair of simyltabeous X-fxy and TeV flates that originate un the standard SSG prozsss. The synchrocxob (X-ray) photons sre thqn rxflecead off some axtefnau cloua and rtcurk tu thr jet (following a sutstahtial delay with rrsiect to tye initiwl flares). The returned photons lnterect wivh the prjtons in the jet to produce pikns and skbscquent $\pi^0$ decwy pvoduses TeV phptons in the flare. The model was shown vo be able to account fir the “orphan” TeV fksre observev in 1EF1959+650 (Böttcher 2004). It might also explqin some of the slmilar flares in Mrk 421, ror exdmple, the one shown in Fig. 9, if it is a composite of two gamma-ray flares (see disxussioi yn § | -Compton scattering of (synchrotron) X-ray photons by themselves the standard model, a flare [*preceded*]{} an accompanying flare X-ray energies. While model might still be able to the presence of orphan X-ray flares, the presence of orphan TeV flares seems On the other hand, the “orphan” TeV flares in Mrk 421 (as shown 9) in (Krawczynski al. 2004) might not be orphan events after all. In both cases, the gamma-ray flares seem be preceded by X-ray flares, which could be to the SSC process the X-ray lead (of 1–2 would too long. it be to attribute the lead to some sort of feedback between different emission regions in the jet (i.e., an orphan X-ray in one an orphan in Genuine TeV flaring would easy to understand in the hadronic looser coupling between the X-ray and TeV emission, the injected ratio changes significantly from flare to Any change in the proton population will likely accompanied by a change in the electron (primary or secondary) population, so TeV and X-ray both occur as a Interestingly, orphan TeV may understood a scenario in protons are present in the jet in substantial quantities but are necessarily dominant compared to the lepton component (Böttcher 2004). In an TeV flare is with (and [**]{}follows) a of X-ray and TeV flares in standard synchrotron photons then reflected off some cloud and return to the (following a substantial delay flares). The returned photons interact with the protons the jet to produce pions and subsequent decay produces TeV photons in the flare. The model was shown to able to the “orphan” TeV flare observed in 1ES1959+650 (Böttcher It might also explain of the similar flares in Mrk 421, for example, one in Fig. if it is composite of two flares (see discussion | -Compton scattering of (synchrOtron) X-ray pHotonS by The ElEctrOns tHemselves in the STandArd SSC model, a flare at TeV EnergIeS Must BE [*pRecedEd*]{} by an aCCoMPAnyInG fLarE aT x-rAy eneRgiEs. While The model miGht StIll be able to aCCoMmodate the PreSence of orphaN X-rAy flarEs, The PResenCe oF orphAn TeV fLAres seEms probleMaTIc. On thE Other haND, ThE “orpHan” TeV flares in Mrk 421 (AS sHOwn Fig. 9) and in 1ES1959+650 (KRawczyNsKI eT AL. 2004) miGht Not be orphaN eVents AFter all. iN bOTH CasES, the gamma-ray fLares seem to BE prEceded By x-raY Flares, Which CoULd bE attributed To thE SSC proceSs althOUgh the X-RAy lead (oF 1–2 days) wOulD seEm toO LoNg. altErNAtiVElY, it MIghT be possiBlE tO attrIbutE THE x-ray LeaD to sOme soRt of feedback bEtwEen dIFfeRent eMissiOn reGiOns in The jet (I.e., an oRpHan X-ray flare in oNe reGion triggErs An OrpHaN TeV iN AnotheR).
GeNuiNe orphaN TeV flaRIng WoULD NoT be easy to understanD iN THe Hadronic Models EItHeR, Despite lOoSer CoupLINg betWeen THe x-ray and TEV emisSIoN, uNless thE iNjecteD e/P raTio ChangES sigNificaNtly from Flare TO flare. Any changE In the proton poPUlATIoN Will LikEly be accompAnieD By a cHangE In The ELectrOn (priMaRY oR Secondary) population, So teV and x-ray fLares should boTh occur as a RESUlt. InterEstiNGlY, Orphan TeV flareS may bE understooD In a hybriD scenArio in whIch protonS ARe presenT in The Jet In sUBStAntial quantitIES but ArE not necEssArily doMinAnt ComParEd To the leptOn componEnT (BÖtTcHer 2004). in thiS Case, an orPhAn TEV FlaRe is aSSociatEd witH (and [**]{}FoLlOWs) a Pair of sIMuLTAneoUs x-rAy anD Tev fLares That ORigInate in The standaRd Ssc proCeSs. the syncHrotron (X-ray) phOtOns are then ReFleCted ofF SOme exterNal cloud and return to the jET (followIng A subsTantIal delay wIth RespecT to THe initIal flaRes). ThE rEtuRNEd phoTONs IntErAct with the PROtoNs in tHe Jet tO producE pions and subsequenT $\Pi^0$ dEcay produces TEV pHotoNS In The FLaRE. ThE mODel WAS shown to be able tO account foR tHE “oRphan” TeV flARe oBsErved in 1eS1959+650 (BöttcHer 2004). It MIght alsO explain sOme of the sImIlar FLAreS in Mrk 421, for eXample, thE one shown IN Fig. 9, iF It Is a coMpoSite of TwO gaMma-raY flareS (See DiscuSsion iN § | -Compton scattering of (sy nchrotron) X-ra y p hot on s by the electrons the m selv es in the standard SSC mode l, a fl a re at T eV ener g ie s mus tbe [* pr e ce ded*] {}by an a ccompanyin g f la re at X-raye ne rgies. Whi lethe model mi ght still b e a b le to ac commo date t h e pres ence of o rp h an X-r a y flare s , t he p resence of orphan Te V flares seemsproble ma t ic . Onthe other han d, the“ orphan” Te V f lar e s in Mrk 421(as shown F i g.9) and i n 1 E S1959+ 650 ( Kr a wcz ynski et al . 20 04) might not b e orphan eventsafterall . I n bo t hca ses ,t heg am ma- r ayflares s ee mto be pre c e d e d by X- rayflare s, which coul d b e at t rib utedto th e SS Cproce ss alt hough t he X-ray lead ( of 1 –2 days)wou ld se em tool ong. A lte rna tively, it mig h t b ep o s si ble to attribute t he X -r ay leadto som e s or t of feed ba ckbetw e e n dif fere n temission regio n sin the je t(i.e., a n o rph an X- r ay f lare i n one re giont riggers an orp h an TeV in ano t he r ) .Genu ine orphan TeV fla r ingwoul d n otb e eas y toun d er s tand in the hadroni cmodels eith er, despite l ooser coup l i n g betwee n th e X - ray and TeV em issio n, unlesst he injec ted e /p ratio changess i gnifican tly fr omfla r e t o flare. Anyc h ange i n the p rot on popu lat ion wi llli kely be a ccompani ed b yacha nge i n the ele ct ron ( pri maryo r seco ndary ) po pu la t ion , so Te V a n d X-r ay f lare s s ho uld b otho ccu r as aresult. I nte r esti ng ly , orpha n TeV flaresma y be under st ood in ah y brid sce nario in which protonsa re pres ent in t he j et in sub sta ntialqua n tities but a re no tnec e s saril y do min an t compared t o t he le pt on c omponen t (Böttcher 2004). Inthis case, an or phan T eV fl a re isas s oci a t ed with (and [* *]{}follow s) apair of si m ult an eous X- ray and TeVf lares t hat origi nate in t he sta n d ard SSC proce ss. Thesynchrotr o n (X- r ay ) pho ton s areth enrefle cted o f f s ome e xterna lcloudand r et urn to t he jet (following a sub stanti al de lay with res pec t to the init ialflares). T heret urned ph o tonsinte r ac t w i th th e pr o tons in t h ejet t oproduce pio n s and subs equ e nt $\p i^0$ decay produces T e V photons in t he f l a re. Th e mod el was shown tobeab l e to acco un t for the “ orphan”Te V flar e obse rved i n 1ES19 5 9 +6 5 0 (Böt tche r 2 004). Itmig ht also ex pl ai n someof t he simil ar fla r es i n Mrk 421, for exa mple, t he on e sh own i nFig. 9, if i t is a com posite of t wo gam ma-r ay fl ares (s ee discu ssi on in § | -Compton scattering_of (synchrotron)_X-ray photons by the_electrons themselves_in_the standard_SSC_model, a flare_at TeV energies_must be [*preceded*]{} by_an accompanying flare_at_X-ray energies. While the model might still be able to accommodate the presence of_orphan_X-ray flares,_the_presence_of orphan TeV flares seems_problematic. On the other hand,_the “orphan”_TeV flares in Mrk 421 (as shown Fig. 9)_and_in 1ES1959+650 (Krawczynski_et al. 2004) might not be orphan events after_all. In both cases, the gamma-ray_flares seem to_be_preceded_by X-ray flares, which_could be attributed to the SSC_process although the X-ray lead (of_1–2 days) would seem too long. Alternatively,_it might be possible to attribute_the X-ray lead to some_sort of_feedback between different emission regions_in the jet_(i.e., an_orphan X-ray flare_in one region triggers an orphan_TeV in another).
Genuine_orphan TeV flaring would not be_easy_to understand in_the_hadronic_models either,_despite looser coupling_between_the X-ray_and_TeV emission, unless the injected e/p_ratio_changes significantly from flare to flare. Any_change in the proton_population_will likely be accompanied_by a change in the_electron (primary or secondary) population, so_TeV and_X-ray flares_should both occur as a result. Interestingly, orphan TeV flares may_be understood in a hybrid scenario_in which protons are_present in_the_jet in substantial_quantities_but are_not necessarily dominant compared to the lepton_component (Böttcher_2004). In this case, an orphan_TeV flare is associated_with_(and [**]{}follows) a pair of simultaneous_X-ray and TeV flares that originate_in the standard SSC process._The_synchrotron_(X-ray) photons are then reflected_off some external cloud and return_to the jet_(following a substantial delay with respect to_the_initial flares). The returned photons interact_with_the protons in the jet to_produce_pions_and subsequent $\pi^0$ decay produces_TeV photons in the flare. The_model was shown to be able to account for_the “orphan” TeV_flare observed in 1ES1959+650 (Böttcher_2004)._It_might also explain some of the similar flares in Mrk_421, for_example, the one_shown in Fig. 9, if it is a composite of two_gamma-ray flares (see discussion in § |
^2},\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+\|\gamma\|_2^2)N^{o(1)}.\end{aligned}$$ By $$\begin{aligned}
\frac{100\Delta^2}{N^2}\le \frac{N}{2},\end{aligned}$$ and hence by Corollary \[cor:larger\] $$\begin{aligned}
\label{eq:SNcase1}
S(N,\Delta,\gamma)\ll N^{o(1)}I(\Delta,\gamma)^{1/2}S\left(N,\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+N\|\gamma\|_2^2)N^{o(1)},\end{aligned}$$ which implies $$\begin{aligned}
\label{eq:SNcase11}
S(N,\Delta,\gamma)\ll (I(\Delta,\gamma)+N\|\gamma\|_2^2)N^{o(1)}. \end{aligned}$$ If does not hold then $$\begin{aligned}
\frac{\Delta^2}{N^2}\gg \Delta^{2/3},\end{aligned}$$ and hence by Corollary \[cor:larger\] and $$\begin{aligned}
S\left(\frac{100\Delta^2}{N^2},\Delta,\gamma \right)\ll \left(I(\Delta,\gamma)+\frac{\Delta^2}{N^2}\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Substituting into gives $$\begin{aligned}
\label{eq:SNprelimF}
S(N,\Delta,\gamma)\ll \left(I(\Delta,\gamma)+\frac{\Delta}{N}I(\Delta,\gamma)^{1/2}\|\gamma\|_2+N\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Let $$\begin{aligned}
\label{eq:Udef154}
U=\max \left\{2,\left(\frac{\Delta I(\Delta,\gamma)^{1/2}}{N^2\|\gamma\|_2}\right)^{1/2} \right\}.\end{aligned}$$ By Corollary \[cor:larger\] and $$\begin{aligned}
\label{eq:SNprelimF1}
S(N,\ | ^2},\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+\|\gamma\|_2 ^ 2)N^{o(1)}.\end{aligned}$$ By $ $ \begin{aligned }
\frac{100\Delta^2}{N^2}\le \frac{N}{2},\end{aligned}$$ and hence by Corollary \[cor: larger\ ] $ $ \begin{aligned }
\label{eq: SNcase1 }
S(N,\Delta,\gamma)\ll N^{o(1)}I(\Delta,\gamma)^{1/2}S\left(N,\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+N\|\gamma\|_2 ^ 2)N^{o(1)},\end{aligned}$$ which implies $ $ \begin{aligned }
\label{eq: SNcase11 }
S(N,\Delta,\gamma)\ll (I(\Delta,\gamma)+N\|\gamma\|_2 ^ 2)N^{o(1) }. \end{aligned}$$ If does not hold then $ $ \begin{aligned }
\frac{\Delta^2}{N^2}\gg \Delta^{2/3},\end{aligned}$$ and hence by Corollary \[cor: larger\ ] and $ $ \begin{aligned }
S\left(\frac{100\Delta^2}{N^2},\Delta,\gamma \right)\ll \left(I(\Delta,\gamma)+\frac{\Delta^2}{N^2}\|\gamma\|_2 ^ 2\right)N^{o(1)}.\end{aligned}$$ substitute into render $ $ \begin{aligned }
\label{eq: SNprelimF }
S(N,\Delta,\gamma)\ll \left(I(\Delta,\gamma)+\frac{\Delta}{N}I(\Delta,\gamma)^{1/2}\|\gamma\|_2+N\|\gamma\|_2 ^ 2\right)N^{o(1)}.\end{aligned}$$ Let $ $ \begin{aligned }
\label{eq: Udef154 }
U=\max \left\{2,\left(\frac{\Delta I(\Delta,\gamma)^{1/2}}{N^2\|\gamma\|_2}\right)^{1/2 } \right\}.\end{aligned}$$ By Corollary \[cor: larger\ ] and $ $ \begin{aligned }
\label{eq: SNprelimF1 }
S(N,\ | ^2},\Delha,\gamma \right)^{1/2}+(I(\Delta,\gamma)+\|\namma\|_2^2)N^{o(1)}.\end{aligneb}$$ By $$\bejin{alighed}
\frac{100\Ddlta^2}{N^2}\le \frac{N}{2},\end{aligned}$$ and iencw by Xorollary \[cor:larger\] $$\begkn{aligned}
\pabel{eq:SBcast1}
S(N,\Delta,\gamma)\ll N^{o(1)}I(\Delta,\gamma)^{1/2}S\lsnt(N,\Denva,\gamma \right)^{1/2}+(I(\Dglta,\gamma)+N\|\gakma\|_2^2)N^{o(1)},\end{aligneg}$$ dhnch implies $$\begin{aligned}
\label{eq:SNcasq11}
S(N,\Delts,\gwmma)\ll (I(\Delta,\gwmma)+M\|\dammz\|_2^2)N^{o(1)}. \end{aligned}$$ If does not hold thsn $$\begii{aligned}
\frac{\Delya^2}{N^2}\gg \Delta^{2/3},\end{aligned}$$ and jencf by Corollary \[cor:lwrger\] and $$\bgfin{woigned}
S\left(\ffac{100\Delta^2}{N^2},\Dtlca,\gamma \riggt)\ll \left(I(\Delta,\gamma)+\frac{\Delta^2}{N^2}\|\gxmma\|_2^2\rnght)N^{o(1)}.\end{alugbed}$$ Vubstitutinj into gives $$\begin{aligned}
\ldbel{eq:SMprelimF}
S(N,\Delts,\gakma)\ol \left(I(\Delta,\gamma)+\frar{\Delta}{N}I(\Delta,\gamma)^{1/2}\|\gaima\|_2+N\|\gamma\|_2^2\siyht)N^{o(1)}.\end{aligned}$$ Let $$\bwgun{alicned}
\nabeu{wq:Uaef154}
H=\mex \meft\{2,\levt(\fcac{\Delta I(\Dslta,\gamma)^{1/2}}{N^2\|\gqmma\|_2}\right)^{1/2} \right\}.\end{slytned}$$ By Corolmary \[cow:lwrger\] and $$\begin{aligned}
\label{eq:SNprelimF1}
S(T,\ | ^2},\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+\|\gamma\|_2^2)N^{o(1)}.\end{aligned}$$ By $$\begin{aligned} \frac{100\Delta^2}{N^2}\le \frac{N}{2},\end{aligned}$$ and Corollary $$\begin{aligned} \label{eq:SNcase1} N^{o(1)}I(\Delta,\gamma)^{1/2}S\left(N,\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+N\|\gamma\|_2^2)N^{o(1)},\end{aligned}$$ which \end{aligned}$$ does not hold $$\begin{aligned} \frac{\Delta^2}{N^2}\gg \Delta^{2/3},\end{aligned}$$ hence by Corollary \[cor:larger\] and $$\begin{aligned} \right)\ll \left(I(\Delta,\gamma)+\frac{\Delta^2}{N^2}\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Substituting into gives $$\begin{aligned} \label{eq:SNprelimF} S(N,\Delta,\gamma)\ll \left(I(\Delta,\gamma)+\frac{\Delta}{N}I(\Delta,\gamma)^{1/2}\|\gamma\|_2+N\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Let $$\begin{aligned} \label{eq:Udef154} U=\max I(\Delta,\gamma)^{1/2}}{N^2\|\gamma\|_2}\right)^{1/2} \right\}.\end{aligned}$$ By Corollary \[cor:larger\] and $$\begin{aligned} \label{eq:SNprelimF1} S(N,\ | ^2},\Delta,\gamma \right)^{1/2}+(I(\Delta,\gammA)+\|\gamma\|_2^2)N^{o(1)}.\enD{aligNed}$$ by $$\bEgIn{alIgneD}
\frac{100\Delta^2}{N^2}\le \fRAc{N}{2},\eNd{aligned}$$ and hence by CorOllarY \[cOR:larGEr\] $$\Begin{Aligned}
\LAbEL{Eq:SncAsE1}
S(N,\deLTa,\Gamma)\Ll N^{O(1)}I(\Delta,\Gamma)^{1/2}S\left(n,\DeLtA,\gamma \right)^{1/2}+(I(\dElTa,\gamma)+N\|\gaMma\|_2^2)n^{o(1)},\end{aligned}$$ WhiCh implIeS $$\beGIn{aliGneD}
\labeL{eq:SNcASe11}
S(N,\DeLta,\gamma)\lL (I(\dElta,\gaMMa)+N\|\gammA\|_2^2)n^{O(1)}. \eNd{alIgned}$$ If does not holD ThEN $$\begin{aligned}
\fRac{\DelTa^2}{n^2}\Gg \dELta^{2/3},\End{Aligned}$$ and HeNce by cOrollarY \[CoR:LARgeR\] And $$\begin{alignEd}
S\left(\frac{100\dEltA^2}{N^2},\DeltA,\gAmmA \Right)\lL \left(i(\DELta,\Gamma)+\frac{\DeLta^2}{N^2}\|\Gamma\|_2^2\righT)N^{o(1)}.\end{ALigned}$$ SUBstitutIng intO giVes $$\BegiN{AlIgNed}
\LaBEl{eQ:sNPreLImF}
s(N,\Delta,\gAmMa)\Ll \lefT(I(\DeLTA,\GAmma)+\FraC{\DelTa}{N}I(\DElta,\gamma)^{1/2}\|\gammA\|_2+N\|\gAmma\|_2^2\RIghT)N^{o(1)}.\enD{aligNed}$$ LEt $$\Begin{AligneD}
\labeL{eQ:Udef154}
U=\max \left\{2,\leFt(\frAc{\Delta I(\DEltA,\gAmmA)^{1/2}}{N^2\|\Gamma\|_2}\RIght)^{1/2} \riGht\}.\End{Aligned}$$ by CorolLAry \[CoR:LARgEr\] and $$\begin{aligned}
\lAbEL{Eq:sNprelimf1}
S(N,\ | ^2},\Delta,\gamma \right)^ {1/2}+(I(\ Delta ,\g amm a) +\|\ gamm a\|_2^2)N^{o(1 ) }.\e nd{aligned}$$ By $$\b egin{ al i gned }
\ frac{ 100\Del t a^ 2 } {N^ 2} \l e \ fr a c{ N}{2} ,\e nd{alig ned}$$ and he nc e by Corolla r y\[cor:larg er\ ] $$\begin{a lig ned}
\ la bel { eq:SN cas e1}
S (N,\De l ta,\ga mma)\ll N ^{ o (1)}I( \ Delta,\ g a mm a)^{ 1/2}S\left(N,\Del t a, \ gamma \right)^ {1/2}+ (I ( \D e l ta, \ga mma)+N\|\g am ma\|_ 2 ^2)N^{o ( 1) } , \ end { aligned}$$ wh ich implies $$\ begin{ al ign e d}
\la bel{e q: S Nca se11}
S(N,\ Delt a,\gamma) \ll (I ( \Delta, \ gamma)+ N\|\ga mma \|_ 2^2) N ^{ o( 1)} .\ end { al ign e d}$ $ If do es n ot ho ld t h e n $$\b egi n{al igned }
\frac{\Delt a^2 }{N^ 2 }\g g \De lta^{ 2/3} ,\ end{a ligned }$$ a nd hence by Corol lary \[cor:la rge r\ ] a nd $$\ b egin{a lig ned }
S\lef t(\frac { 100 \D e l t a^ 2}{N^2},\Delta,\ga mm a \r ight)\ll \left ( I( \D e lta,\gam ma )+\ frac { \ Delta ^2}{ N ^2 }\|\gamm a\|_2^ 2 \r ig ht)N^{o (1 )}.\en d{ ali gne d}$$S ubst itutin g into gives $$\begin{align e d}
\label{eq: S Np r e li m F}
S (N, \Delta,\gam ma)\ l l \l eft( I (\ Del t a,\ga mma)+ \f r ac { \Delta}{N}I(\Delta, \g amma)^ {1/2} \|\gamma\|_2+ N\|\gamma\ | _ 2 ^2\right )N^{ o (1 ) }.\end{aligned }$$ L et $$\begi n {aligned }
\la bel{eq:U def154}
U = \ max \lef t\{ 2,\ lef t(\ f r ac {\Delta I(\De l t a,\g am ma)^{1/ 2}} {N^2\|\ gam ma\ |_2 }\r ig ht)^{1/2} \right \} .\ en d{ ali gned} $ $ By Cor ol lar y\[c or:la r ger\]and $$\b eg in { ali gned}
\ l ab e l {eq: SN pr elim F1}
S (N,\ | ^2},\Delta,\gamma \right)^{1/2}+(I(\Delta,\gamma)+\|\gamma\|_2^2)N^{o(1)}.\end{aligned}$$_By $$\begin{aligned}
\frac{100\Delta^2}{N^2}\le_\frac{N}{2},\end{aligned}$$ and hence by_Corollary \[cor:larger\] $$\begin{aligned}
\label{eq:SNcase1}
S(N,\Delta,\gamma)\ll_N^{o(1)}I(\Delta,\gamma)^{1/2}S\left(N,\Delta,\gamma_\right)^{1/2}+(I(\Delta,\gamma)+N\|\gamma\|_2^2)N^{o(1)},\end{aligned}$$ which_implies_$$\begin{aligned}
\label{eq:SNcase11}
S(N,\Delta,\gamma)\ll (I(\Delta,\gamma)+N\|\gamma\|_2^2)N^{o(1)}. \end{aligned}$$_If does not_hold then $$\begin{aligned}
\frac{\Delta^2}{N^2}\gg \Delta^{2/3},\end{aligned}$$_and hence by_Corollary \[cor:larger\]_and $$\begin{aligned}
S\left(\frac{100\Delta^2}{N^2},\Delta,\gamma \right)\ll \left(I(\Delta,\gamma)+\frac{\Delta^2}{N^2}\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Substituting into gives $$\begin{aligned}
\label{eq:SNprelimF}
S(N,\Delta,\gamma)\ll \left(I(\Delta,\gamma)+\frac{\Delta}{N}I(\Delta,\gamma)^{1/2}\|\gamma\|_2+N\|\gamma\|_2^2\right)N^{o(1)}.\end{aligned}$$ Let $$\begin{aligned}
\label{eq:Udef154}
U=\max \left\{2,\left(\frac{\Delta I(\Delta,\gamma)^{1/2}}{N^2\|\gamma\|_2}\right)^{1/2} _\right\}.\end{aligned}$$_By Corollary \[cor:larger\]_and _$$\begin{aligned}
\label{eq:SNprelimF1}
S(N,\ |
$ to further reduce the QCD background. All these improvements will thus further reinforce and extend the reach of the $H_{SUSY}$ [$\;\rightarrow$]{}$~\tau\tau$ channel, the only channel which until now gives us access to the theoretically favoured $H_{SUSY}>$ 500 GeV mass range. However, this channel should also play an important (decisive?) role in the intermediate mass range 150 GeV $< m_H <$ 300 GeV, which is the most critical one from the point of view of the MSSM $tan\beta$, $m_A$ parameter space coverage, provided we can trigger on it efficiently, which requires a dedicated study and is not discussed here.
Event simulation
================
Events are generated with PYTHIA [@PYTHIA]. Masses and couplings of SUSY Higgses are calculated using two-loop/RGE-improved radiative corrections [@gunion]. The branching ratios and cross sections are normalized using the HDECAY program [@HDECAY]. No stop mixing is included and the SUSY particles are first assumed to be heavy enough not to contribute to the SUSY Higgs decays. The decays to neutralinos and charginos can however reduce strongly the $A, H$ [$\;\rightarrow$]{}$~\tau\tau$ branching ratio at low $tan\beta$, and by a factor of two even at $tan\beta$ = 30 [@HDECAY]. For the high $m_A$ and $tan\beta$ region investigated here the effects of sparticle decay modes are nevertheless relatively modest, as mentioned later on. In the case of stop mixing, if the light stop $\tilde{t_1}$ is lighter than or comparable in mass to the top quark, the squark loop effects are expected to suppress significantly the $gg$ [$\;\rightarrow$]{}$~h, A, H$ production processes [@djouadi]. For the large $m_A$ and $tan\beta$ values discussed here this would mean at most a 20% reduction in the production rate as we are dominated by $gg$ [$\;\rightarrow$]{}$~b\overline{b}H_{SUSY}$ tree-diagram production which is not affected by $\tilde{t}_1$ - $t$ interference effects. For the QCD background, PYTHIA two-jet events with initial and final state QCD radiation are generated. A cut $p_t^{q,g} | $ to further reduce the QCD background. All these improvements will therefore far reinforce and extend the reach of the $ H_{SUSY}$ [ $ \;\rightarrow$]{}$~\tau\tau$ distribution channel, the alone channel which until now give us access to the theoretically favoured $ H_{SUSY}>$ 500 GeV aggregate compass. However, this channel should besides play an important (decisive ?) role in the average mass range 150 GeV $ < m_H < $ 300 GeV, which is the about critical one from the point of opinion of the MSSM $ tan\beta$, $ m_A$ parameter space coverage, provided we can trigger on it efficiently, which want a dedicated study and is not discussed here.
Event simulation
= = = = = = = = = = = = = = = =
event are generated with PYTHIA [ @PYTHIA ]. Masses and couplings of SUSY Higgses are calculated using two - loop topology / RGE - improved radiative corrections [ @gunion ]. The branching ratios and cross sections are normalized using the HDECAY program [ @HDECAY ]. No stop mixing is included and the SUSY particles are first assumed to be heavy enough not to contribute to the SUSY Higgs decay. The decays to neutralinos and charginos can however reduce powerfully the $ A, H$ [ $ \;\rightarrow$]{}$~\tau\tau$ branching proportion at low $ tan\beta$, and by a factor of two even at $ tan\beta$ = 30 [ @HDECAY ]. For the high $ m_A$ and $ tan\beta$ region investigate here the effects of sparticle decay modes are nevertheless relatively modest, as mention later on. In the case of stop mixing, if the light barricade $ \tilde{t_1}$ is lighter than or comparable in mass to the top quark, the squark loop topology effects are expect to suppress significantly the $ gg$ [ $ \;\rightarrow$]{}$~h, A, H$ production processes [ @djouadi ]. For the large $ m_A$ and $ tan\beta$ value discussed here this would mean at most a 20% reduction in the production rate as we are dominated by $ gg$ [ $ \;\rightarrow$]{}$~b\overline{b}H_{SUSY}$ tree - diagram production which is not affected by $ \tilde{t}_1 $ - $ t$ hindrance effects. For the QCD backdrop, PYTHIA two - jet events with initial and concluding state QCD radiation are generated. A cut $ p_t^{q, gigabyte } | $ to further reduce the QCD nackground. All tkwse im'rovemehts will thus further reinforce and xxtebd tht reach of the $H_{SUSH}$ [$\;\rightargow$]{}$~\tau\tau$ chainel, the only chehnel whleh unfll noc jives us access to the thaoretically faeojrzd $H_{SUSY}>$ 500 GeV mass range. However, thif channrl should also pjay sg imlortant (decisive?) role in the interjediate mass range 150 BeV $< m_H <$ 300 GeV, which is the mosh critical one frol the point of duew of the MRSM $tan\beta$, $m_A$ parametgr space coverage, provided we cav triyger on it gydiclgntly, which cequirvs a dedicated study dnd is mot discussed mere.
Etent simulation
================
Events are generated with PYJHIA [@PYTHID]. Jasses and couplibgw of VUSY Higewes art celchlated usmng two-loop/DGE-improved radiative correctipnf [@gunion]. The bdanchigg ratios and cross sections are normaliztd usjng the HDECAY program [@YDECAY]. No stop mixing is incluqed and the SUSY particles are first assumed to ba heaty eniunh nur ho contribute to the SUSY Higgs decays. The deszyx no neutralinos anb charginos can noaefgr reduce strovgly tks $Z, H$ [$\;\rightarrow$]{}$~\tau\twu$ branshing ratio at low $tan\beta$, and by a factor of two even at $ran\beta$ = 30 [@HDECAY]. Flr the high $m_A$ snd $tsn\beta$ region investigaced hede the effefts of spzfticle decay modds sra nevertheless relatively iodest, as mencioned lxter on. In the case lf stop mixing, if the lihht sjop $\tinde{t_1}$ is lihhter than or comparable in mass to the top qoarn, tve squarh looi effects are evpected to supkress siguificavtly the $gf$ [$\;\righterrow$]{}$~h, A, H$ pwoduction prowgsses [@djouadi]. For the larte $m_Q$ and $txv\beta$ values doscussed ktre this wiuld mean at most e 20% rgdhction in the pxueuction rate as we arq fokigdted by $gg$ [$\;\rhghtxrrue$]{}$~b\ovefline{b}H_{SUSY}$ gree-ciagram production wvich is not affected bu $\bilde{t}_1$ - $t$ interfewence effects. For the QCD backggound, PYVHIA teo-jgt events with initial and finam state QFD vadiation are gencratgd. A cut $p_t^{a,g} | $ to further reduce the QCD background. improvements thus further and extend the channel, only channel which now gives us to the theoretically favoured $H_{SUSY}>$ 500 mass range. However, this channel should also play an important (decisive?) role in intermediate mass range 150 GeV $< m_H <$ 300 GeV, which is the critical from point view of the MSSM $tan\beta$, $m_A$ parameter space coverage, provided we can trigger on it efficiently, requires a dedicated study and is not discussed Event simulation ================ Events generated with PYTHIA [@PYTHIA]. Masses couplings SUSY Higgses calculated two-loop/RGE-improved corrections [@gunion]. The ratios and cross sections are normalized using the HDECAY program [@HDECAY]. No stop mixing is included and SUSY particles assumed to heavy not contribute to the decays. The decays to neutralinos and reduce strongly the $A, H$ [$\;\rightarrow$]{}$~\tau\tau$ branching ratio low $tan\beta$, by a factor of two even $tan\beta$ = 30 [@HDECAY]. For the high $m_A$ $tan\beta$ region investigated here the effects of sparticle decay modes are nevertheless relatively modest, as on. In the case stop mixing, if light $\tilde{t_1}$ lighter or comparable mass to the top quark, the squark loop effects are expected suppress significantly the $gg$ [$\;\rightarrow$]{}$~h, A, H$ production processes [@djouadi]. large and $tan\beta$ values here this would mean most 20% reduction in the as are [$\;\rightarrow$]{}$~b\overline{b}H_{SUSY}$ production is not affected by - $t$ interference effects. For QCD background, PYTHIA two-jet state QCD radiation are generated. A cut $p_t^{q,g} | $ to further reduce the QCD backGround. All tHese iMprOveMeNts wIll tHus further reinFOrce And extend the reach of the $h_{SUSY}$ [$\;\RiGHtarROw$]{}$~\Tau\taU$ channeL, ThE ONly ChAnNel WhICh Until Now Gives us Access to thE thEoRetically favOUrEd $H_{SUSY}>$ 500 GeV MasS range. HoweveR, thIs chanNeL shOUld alSo pLay an ImportANt (deciSive?) role iN tHE interMEdiate mASS rAnge 150 geV $< m_H <$ 300 GeV, which is thE MoST critical one frOm the pOiNT oF VIew Of tHe MSSM $tan\bEtA$, $m_A$ paRAmeter sPAcE COVerAGe, provided we cAn trigger on IT efFicienTlY, whICh requIres a DeDIcaTed study and Is noT discusseD here.
EVEnt simuLAtion
================
EvEnts arE geNerAted WItH PyTHiA [@pyTHia]. MAssES anD couplinGs Of sUSY HIggsES ARE calCulAted Using Two-loop/RGE-impRovEd raDIatIve coRrectIons [@GuNion]. THe branChing RaTios and cross secTionS are normaLizEd UsiNg The HDecAY proGraM [@HDeCAY]. No sTop mixiNG is InCLUDeD and the SUSY particlEs ARE fIrst assuMed to bE HeAvY Enough noT tO coNtriBUTe to tHe SUsy HIggs decaYs. The dECaYs To neutrAlInos anD cHarGinOs can HOwevEr reduCe strongLy the $a, h$ [$\;\rightarrow$]{}$~\tau\TAu$ branching raTIo AT LoW $Tan\bEta$, And by a factoR of tWO eveN at $tAN\bEta$ = 30 [@hdECAY]. for thE hIGh $M_a$ and $tan\beta$ region inVeStigatEd herE the effects of Sparticle dECAY modes arE nevERtHEless relativelY modeSt, as mentioNEd later oN. In thE case of sTop mixing, IF The light StoP $\tiLde{T_1}$ is LIGhTer than or compARAble In Mass to tHe tOp quark, The SquArk LooP eFfects are Expected To SuPpReSs sIgnifICantly thE $gG$ [$\;\riGhTarRow$]{}$~h, A, h$ ProducTion pRoceSsEs [@DJouAdi]. For tHE lARGe $m_A$ AnD $tAn\beTa$ vAlUes diScusSEd hEre this Would mean At mOSt a 20% rEdUcTion in tHe production rAtE as we are doMiNatEd by $gg$ [$\;\RIGhtarrow$]{}$~B\overline{b}H_{SUSY}$ tree-diagRAm produCtiOn whiCh is Not affectEd bY $\tilde{T}_1$ - $t$ iNTerferEnce efFects. foR thE qcD bacKGRoUnd, pYtHIA two-jet EVEntS with InItiaL and finAl state QCD radiatioN Are Generated. A cut $P_t^{q,G} | $ to further reduce the QC D backgrou nd. A llthe se imp rove ments will thu s fur ther reinforce and ext end t he reac h o f the $H_{SU S Y} $ [$\ ;\ ri ght ar r ow $]{}$ ~\t au\tau$ channel,the o nly channelw hi ch until n owgives us acc ess to th ethe o retic all y fav oured$ H_{SUS Y}>$ 500Ge V massr ange. H o w ev er,this channel shou l da lso play an im portan t( de c i siv e?) role in t he inte r mediate ma s s ran g e 150 GeV $<m_H <$ 300G eV, which i s t h e most crit ic a l o ne from the poi nt of vie w of t h e MSSM$ tan\bet a$, $m _A$ pa rame t er s pac ec ove r ag e,p rov ided weca ntrigg er o n i t eff ici entl y, wh ich requiresa d edic a ted stud y and isno t dis cussed here .
Event simulati on
= ========= === == =
Ev entsa re gen era ted with P YTHIA [ @ PYT HI A ] . M asses and coupling so f S USY Higg ses ar e c al c ulated u si ngtwo- l o op/RG E-im p ro ved radi ativec or re ctions[@ gunion ]. Th e b ranch i ng r atiosand cros s sec t ions are norma l ized using th e H D E CA Y pro gra m [@HDECAY] . No stop mix i ng is inclu ded a nd th e SUSY particles are f irst a ssume d to be heavy enough no t t o contri bute to the SUSY Higgs deca ys. The de c ays to n eutra linos an d chargin o s can how eve r r edu ces t ro ngly the $A,H $ [$\ ;\ rightar row $]{}$~\ tau \ta u$bra nc hing rati o at low $ ta n\ be ta$ , and by a fac to r o ftwo even at $ta n\bet a$ = 3 0[ @HD ECAY].F or t he h ig h$m_A $ a nd $tan \bet a $ r egion i nvestigat edh ereth eeffects of sparticle d ecay modes a renevert h e less rel atively modest, as ment i oned la ter on.In t he case o f s top mi xin g , if t he lig ht st op $\ t i lde{t _ 1 }$ is l ighter tha n orcompa ra blein mass to the top quark, the squark loopeff ects a re ex p ec t edto sup p r ess significant ly the $gg $[ $\ ;\rightarr o w$] {} $~h, A, H$ pro ducti o n proce sses [@dj ouadi]. F or the l arg e $m_A$ an d $tan\b eta$ valu e s dis c us sed h ere thiswo uld mean at mo s t a 20%reduct io n in t he pr od uction r ate as we are dominated by $g g$ [$ \;\ rightarro w$] { }$~ b\overlin e{b} H_{SUSY}$tre e-d iagra m p r oduct ionw hi chi s not aff e cted by $ \ ti lde { t }_ 1$ - $t$ in t e r fer enceeff e cts. F or t he QCD background , PYTHIA two-je t ev e n tswit h ini ti al and final s tat eQ C D radiat io n are gener ated. Acu t $p_t ^{q,g} | $ to_further reduce_the QCD background. All_these improvements_will_thus further_reinforce_and extend the_reach of the_$H_{SUSY}$ [$\;\rightarrow$]{}$~\tau\tau$ channel, the_only channel which_until_now gives us access to the theoretically favoured $H_{SUSY}>$ 500 GeV mass range. However,_this_channel should_also_play_an important (decisive?) role in_the intermediate mass range 150_GeV $<_m_H <$ 300 GeV, which is the most_critical_one from the_point of view of the MSSM $tan\beta$, $m_A$ parameter_space coverage, provided we can trigger_on it efficiently,_which_requires_a dedicated study and_is not discussed here.
Event simulation
================
Events are_generated with PYTHIA [@PYTHIA]. Masses and_couplings of SUSY Higgses are calculated using_two-loop/RGE-improved radiative corrections [@gunion]. The branching_ratios and cross sections are_normalized using_the HDECAY program [@HDECAY]. No_stop mixing is_included and_the SUSY particles_are first assumed to be heavy_enough not to_contribute to the SUSY Higgs decays._The_decays to neutralinos_and_charginos_can however_reduce strongly the_$A,_H$ [$\;\rightarrow$]{}$~\tau\tau$_branching_ratio at low $tan\beta$, and by_a_factor of two even at $tan\beta$ =_30 [@HDECAY]. For the_high_$m_A$ and $tan\beta$ region_investigated here the effects of_sparticle decay modes are nevertheless relatively_modest, as_mentioned later_on. In the case of stop mixing, if the light stop_$\tilde{t_1}$ is lighter than or comparable_in mass to the_top quark,_the_squark loop effects_are_expected to_suppress significantly the $gg$ [$\;\rightarrow$]{}$~h, A, H$_production processes_[@djouadi]. For the large $m_A$ and_$tan\beta$ values discussed here_this_would mean at most a 20%_reduction in the production rate as_we are dominated by $gg$_[$\;\rightarrow$]{}$~b\overline{b}H_{SUSY}$_tree-diagram_production which is not affected_by $\tilde{t}_1$ - $t$ interference effects._For the QCD_background, PYTHIA two-jet events with initial and_final_state QCD radiation are generated. A_cut_$p_t^{q,g} |
073 & +05 42 45.057 & 0.051 & 0.33 &0.03 &13.68\
G39.10+0.49 & 19 00 58.031 & 0.075 & +05 42 45.142 & 0.053 & 0.36 &0.03 &13.81\
G39.10+0.49 & 19 00 58.033 & 0.078 & +05 42 45.163 & 0.055 & 0.36 &0.03 &13.95\
G39.10+0.49 & 19 00 58.039 & 0.063 & +05 42 45.078 & 0.044 & 0.44 &0.03 &14.09\
G39.10+0.49 & 19 00 58.040 & 0.026 & +05 42 45.076 & 0.018 & 0.99 &0.03 &14.22\
G39.10+0.49 & 19 00 58.040 & 0.010 & +05 42 45.085 & 0.007 & 2.83 &0.03 &14.36\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.087 & 0.004 & 5.63 &0.04 &14.50\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.084 & 0.004 & 6.97 &0.04 &14.64\
G39.10+0.49 & 19 00 58.041 & 0.006 & +05 42 45.080 & 0.004 & 5.88 &0.03 &14.77\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.076 & 0.004 & 4.83 &0.03 &14.91\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.082 & 0.003 & 6.57 &0.03 &15.05\
G39.10+0.49 & 19 00 58.040 & 0.004 & +05 42 45.091 & 0.003 & 10.38 &0.04 &15.18\
G39.10+0.49 & 19 00 58. | 073 & +05 42 45.057 & 0.051 & 0.33 & 0.03 & 13.68\
G39.10 + 0.49 & 19 00 58.031 & 0.075 & +05 42 45.142 & 0.053 & 0.36 & 0.03 & 13.81\
G39.10 + 0.49 & 19 00 58.033 & 0.078 & +05 42 45.163 & 0.055 & 0.36 & 0.03 & 13.95\
G39.10 + 0.49 & 19 00 58.039 & 0.063 & +05 42 45.078 & 0.044 & 0.44 & 0.03 & 14.09\
G39.10 + 0.49 & 19 00 58.040 & 0.026 & +05 42 45.076 & 0.018 & 0.99 & 0.03 & 14.22\
G39.10 + 0.49 & 19 00 58.040 & 0.010 & +05 42 45.085 & 0.007 & 2.83 & 0.03 & 14.36\
G39.10 + 0.49 & 19 00 58.040 & 0.006 & +05 42 45.087 & 0.004 & 5.63 & 0.04 & 14.50\
G39.10 + 0.49 & 19 00 58.040 & 0.005 & +05 42 45.084 & 0.004 & 6.97 & 0.04 & 14.64\
G39.10 + 0.49 & 19 00 58.041 & 0.006 & +05 42 45.080 & 0.004 & 5.88 & 0.03 & 14.77\
G39.10 + 0.49 & 19 00 58.040 & 0.006 & +05 42 45.076 & 0.004 & 4.83 & 0.03 & 14.91\
G39.10 + 0.49 & 19 00 58.040 & 0.005 & +05 42 45.082 & 0.003 & 6.57 & 0.03 & 15.05\
G39.10 + 0.49 & 19 00 58.040 & 0.004 & +05 42 45.091 & 0.003 & 10.38 & 0.04 & 15.18\
G39.10 + 0.49 & 19 00 58. | 073 & +05 42 45.057 & 0.051 & 0.33 &0.03 &13.68\
G39.10+0.49 & 19 00 58.031 & 0.075 & +05 42 45.142 & 0.053 & 0.36 &0.03 &13.81\
N39.10+0.49 & 19 00 58.033 & 0.078 & +05 42 45.163 & 0.055 & 0.36 &0.03 &13.95\
G39.10+0.49 & 19 00 58.039 & 0.063 & +05 42 45.078 & 0.044 & 0.44 &0.03 &14.09\
G39.10+0.49 & 19 00 58.040 & 0.026 & +05 42 45.076 & 0.018 & 0.99 &0.03 &14.22\
G39.10+0.49 & 19 00 58.040 & 0.010 & +05 42 45.085 & 0.007 & 2.83 &0.03 &14.36\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.087 & 0.004 & 5.63 &0.04 &14.50\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.084 & 0.004 & 6.97 &0.04 &14.64\
G39.10+0.49 & 19 00 58.041 & 0.006 & +05 42 45.080 & 0.004 & 5.88 &0.03 &14.77\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.076 & 0.004 & 4.83 &0.03 &14.91\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.082 & 0.003 & 6.57 &0.03 &15.05\
G39.10+0.49 & 19 00 58.040 & 0.004 & +05 42 45.091 & 0.003 & 10.38 &0.04 &15.18\
C39.10+0.49 & 19 00 58. | 073 & +05 42 45.057 & 0.051 &0.03 G39.10+0.49 & 00 58.031 & & & 0.36 &0.03 G39.10+0.49 & 19 58.033 & 0.078 & +05 42 & 0.055 & 0.36 &0.03 &13.95\ G39.10+0.49 & 19 00 58.039 & 0.063 +05 42 45.078 & 0.044 & 0.44 &0.03 &14.09\ G39.10+0.49 & 19 00 & & 42 & 0.018 & 0.99 &0.03 &14.22\ G39.10+0.49 & 19 00 58.040 & 0.010 & +05 42 & 0.007 & 2.83 &0.03 &14.36\ G39.10+0.49 & 00 58.040 & 0.006 +05 42 45.087 & 0.004 5.63 &14.50\ G39.10+0.49 19 58.040 0.005 & +05 45.084 & 0.004 & 6.97 &0.04 &14.64\ G39.10+0.49 & 19 00 58.041 & 0.006 & +05 42 & 0.004 &0.03 &14.77\ & 00 & 0.006 & 45.076 & 0.004 & 4.83 &0.03 19 00 58.040 & 0.005 & +05 42 & 0.003 6.57 &0.03 &15.05\ G39.10+0.49 & 19 58.040 & 0.004 & +05 42 45.091 & & 10.38 &0.04 &15.18\ G39.10+0.49 & 19 00 58. | 073 & +05 42 45.057 & 0.051 & 0.33 &0.03 &13.68\
G39.10+0.49 & 19 00 58.031 & 0.075 & +05 42 45.142 & 0.053 & 0.36 &0.03 &13.81\
G39.10+0.49 & 19 00 58.033 & 0.078 & +05 42 45.163 & 0.055 & 0.36 &0.03 &13.95\
G39.10+0.49 & 19 00 58.039 & 0.063 & +05 42 45.078 & 0.044 & 0.44 &0.03 &14.09\
G39.10+0.49 & 19 00 58.040 & 0.026 & +05 42 45.076 & 0.018 & 0.99 &0.03 &14.22\
G39.10+0.49 & 19 00 58.040 & 0.010 & +05 42 45.085 & 0.007 & 2.83 &0.03 &14.36\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.087 & 0.004 & 5.63 &0.04 &14.50\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.084 & 0.004 & 6.97 &0.04 &14.64\
G39.10+0.49 & 19 00 58.041 & 0.006 & +05 42 45.080 & 0.004 & 5.88 &0.03 &14.77\
G39.10+0.49 & 19 00 58.040 & 0.006 & +05 42 45.076 & 0.004 & 4.83 &0.03 &14.91\
G39.10+0.49 & 19 00 58.040 & 0.005 & +05 42 45.082 & 0.003 & 6.57 &0.03 &15.05\
G39.10+0.49 & 19 00 58.040 & 0.004 & +05 42 45.091 & 0.003 & 10.38 &0.04 &15.18\
G39.10+0.49 & 19 00 58. | 073 & +05 42 45.057 & 0.05 1 & 0.33 & 0.03&13 .68 \G39. 10+0 .49 & 19 00 58 . 031& 0.075 & +05 42 45.14 2 & 0 .0 5 3 &0 .3 6 &0. 03 &13. 8 1\ G 39. 10 +0 .49 & 19 00 5 8.0 33 & 0. 078 & +054245 .163 & 0.055 &0.36 &0.03 &1 3.95\
G39.10 +0. 49 & 1 9005 8.039 &0.063 & +05 42 45. 078 & 0.0 44 & 0.44 &0.03 & 1 4 .0 9\
G 39.10+0.49 & 19 0 0 5 8 .040 & 0.026 & +05 4 24 5. 0 7 6 & 0. 018 & 0.99 & 0.03& 14.22\G 39 . 1 0 +0. 4 9 & 19 00 58. 040 & 0.010 & + 05 4245 .08 5 & 0.0 07 &2. 8 3 & 0.03 &14.36 \
G3 9.10+0.49 & 190 0 58.04 0 & 0.00 6 & +0 5 4 2 4 5.08 7 & 0 .00 4& 5. 6 3&0. 0 4 & 14.50\
G 39 .1 0+0.4 9 &1 9 0 0 58 .04 0 &0.005 & +05 42 45. 084 & 0 . 004 & 6. 97 &0 .04&1 4.64\
G39.1 0+0.4 9& 19 00 58.041& 0. 006 & +05 42 4 5.0 80 & 0. 0 04 & 5 .88 &0 .03 &14 .77\
G3 9 .10 +0 . 4 9 & 19 00 58.040 & 0. 00 6 &+05 42 4 5.076& 0 .0 0 4 & 4.83 & 0.0 3 &1 4 . 91\
G 39.1 0 +0 .49 & 19 00 58 . 04 0& 0.005 & +05 4 245. 082 & 0. 0 03 & 6.57&0.03 &1 5.05\ G39.10+0.49 &1 9 00 58.040 & 0. 0 0 4& +05 42 45.091 & 0 .003 & 10 .38& 0. 04& 15.18 \
G39 .1 0 +0 . 49 & 19 00 58. | 073 &_+05 42_45.057 & 0.051 &_0.33 &0.03_&13.68\
G39.10+0.49_& 19_00_58.031 & 0.075_& +05 42_45.142 & 0.053 &_0.36 &0.03 &13.81\
G39.10+0.49_&_19 00 58.033 & 0.078 & +05 42 45.163 & 0.055 & 0.36 &0.03_&13.95\
G39.10+0.49_& 19_00_58.039_& 0.063 & +05 42_45.078 & 0.044 & 0.44_&0.03 &14.09\
G39.10+0.49_& 19 00 58.040 & 0.026 & +05_42_45.076 & 0.018_& 0.99 &0.03 &14.22\
G39.10+0.49 & 19 00 58.040 &_0.010 & +05 42 45.085 &_0.007 & 2.83_&0.03_&14.36\
G39.10+0.49_& 19 00 58.040_& 0.006 & +05 42 45.087_& 0.004 & 5.63 &0.04 &14.50\
G39.10+0.49_& 19 00 58.040 & 0.005 &_+05 42 45.084 & 0.004 &_6.97 &0.04 &14.64\
G39.10+0.49 & 19_00 58.041_& 0.006 & +05 42_45.080 & 0.004_& 5.88_&0.03 &14.77\
G39.10+0.49 &_19 00 58.040 & 0.006 &_+05 42 45.076_& 0.004 & 4.83 &0.03 &14.91\
G39.10+0.49_&_19 00 58.040_&_0.005_& +05_42 45.082 &_0.003_& 6.57_&0.03_&15.05\
G39.10+0.49 & 19 00 58.040 &_0.004_& +05 42 45.091 & 0.003 &_10.38 &0.04 &15.18\
G39.10+0.49 &_19_00 58. |
above a threshold of $0.018$ mJy beam$^{-1}$ (3 $\sigma$).
VLBA
----
J1000+0206 was observed with the VLBA during four 6-hour sessions spanning 2012 June 24 to 2013 January 11 under proposal code BM360. A coordinate equinox of 2000 was used. J1011+0106, with a one-dimensional position error of 2 mas (1 $\sigma$), was used as a phase calibrator. The switching time between it and J1000+0206 was 360 s, and involved a switching angle of 2.7. Every 4.1 s the correlator [@del11] produced 128 contiguous 2-MHz channels, yielding a total bandwidth of 0.256 GHz per circular polarization centered at $\nu$ = 1.52 GHz. VLBA system temperatures and gains were used to set the amplitude scale to an accuracy of about 5%. A total of about 450 baseline hours were accrued on J1000+0206.
For each 6-hour session, NRAO’s Astronomical Image Processing System [AIPS; @gre03] was used to calibrate and edit the data following the approach described by @mid13. The AIPS task [ imagr]{} was used to form and deconvolve a naturally-weighted image of the Stokes $I\/$ emission from J1000+0206. This image spans 2048 $\times$ 1 mas in each coordinate, has a geometric resolution at FWHM $\theta$ = 15 mas, and was corrected for primary beam attenuation. Combining the u-v data from the four sessions resulted in an rms noise level $\sigma = 0.013$ mJy beam$^{-1}$. With adequate signal-to-noise, structures as large as about 10 $\theta$ $\sim$ 150 mas could be imaged. No emission was detected above a threshold of $0.078$ mJy beam$^{-1}$ (6 $\sigma$).
Implications
============
Overview
--------
The new VLBA image at $\nu$ = 1.52 GHz and $\theta$ = 15 mas (75 pc) filters for emission with a rest-frame brightness temperature $T_b >
3.5 \times 10^5$ K, not achieved by even the most compact starbursts [@con92]. Moreover, the VLBA image is too shallow to detect even the most luminous radio supernova beyond $z \sim | above a threshold of $ 0.018 $ mJy beam$^{-1}$ (3 $ \sigma$).
VLBA
----
J1000 + 0206 was observed with the VLBA during four 6 - hour sessions spanning 2012 June 24 to 2013 January 11 under marriage proposal code BM360. A coordinate equinoctial point of 2000 was practice. J1011 + 0106, with a one - dimensional position error of 2 master of arts (1 $ \sigma$), was used as a phase calibrator. The switching prison term between it and J1000 + 0206 was 360 mho, and involved a switch over angle of 2.7. Every 4.1 s the correlator [ @del11 ] produce 128 contiguous 2 - MHz duct, give a total bandwidth of 0.256 GHz per circular polarization centered at $ \nu$ = 1.52 GHz. VLBA arrangement temperatures and gains were used to fix the amplitude scale to an accuracy of about 5% . A total of about 450 service line hours were accrue on J1000 + 0206.
For each 6 - hour session, NRAO ’s Astronomical Image Processing System [ AIPS; @gre03 ] was used to calibrate and edit the data follow the approach described by @mid13. The AIPS task [ imagr ] { } was used to form and deconvolve a naturally - burden image of the Stokes $ I\/$ emission from J1000 + 0206. This image spans 2048 $ \times$ 1 mas in each coordinate, has a geometric resolution at FWHM $ \theta$ = 15 mas, and was corrected for primary radio beam attenuation. Combining the u - five data from the four sessions resulted in an rms noise grade $ \sigma = 0.013 $ mJy beam$^{-1}$. With adequate signal - to - randomness, structures as large as approximately 10 $ \theta$ $ \sim$ 150 mas could be imaged. No emission was detected above a threshold of $ 0.078 $ mJy beam$^{-1}$ (6 $ \sigma$).
Implications
= = = = = = = = = = = =
Overview
--------
The raw VLBA image at $ \nu$ = 1.52 GHz and $ \theta$ = 15 mas (75 pc) filter for emission with a remainder - frame brightness temperature $ T_b >
3.5 \times 10 ^ 5 $ K, not achieved by even the most compendious starbursts [ @con92 ]. Moreover, the VLBA image is too shallow to detect even the most luminous radio supernova beyond $ z \sim | ablve a threshold of $0.018$ mJy bexm$^{-1}$ (3 $\sigma$).
VLBA
----
J1000+0206 cqs obsxrved wjth the XLBA during four 6-hour sessiois spannibg 2012 June 24 to 2013 January 11 under prlposal cide UM360. A coordinate xsuinox of 2000 waa useb. O1011+0106, with a one-dimgnsional poshtion error of 2 oad (1 $\sigma$), was used as a phase calibrwtor. Thr dwitching time beteqen jn cnd J1000+0206 was 360 s, and involved a switdhing aigle of 2.7. Every 4.1 s the correlator [@del11] prodkced 128 contiguous 2-MHz cjannels, yieodind a total banawidth of 0.256 GHz per circolar polarization centered at $\nu$ = 1.52 GHz. VLBA systgm tfkperatures end gapns were used to set dhe ampkitude scale tp ai acxuracy of about 5%. A toval of about 450 baselige hours fexe accrued on J1000+0206.
For eaxh 6-hour sesvion, BRAU’s Zsvrohomicap Ijage Proceasing Systen [AIPS; @gre03] was used tj calibrate ans edit tre data following the approach describeg bg @mid13. The AIPS task [ imqgr]{} was used to form wnd decondolve a naturally-weighted image of the Stokes $I\/$ ekissikv fxim J1000+0206. Gyid image spans 2048 $\times$ 1 mas in each coordinate, rzs a geometric resjlution at GWJM $\jheta$ = 15 mas, ana was ekrdected for primary beam ajtenuarion. Combyninb the u-v data from the four sessions refylted in an rms nonse level $\siyma = 0.013$ mKy beak$^{-1}$. With adequate signal-tu-noiae, structurfs as larfd as about 10 $\thetx$ $\spm$ 150 kas could be imaged. No emifsion was detzcted abuve s threfhold of $0.078$ mUy beam$^{-1}$ (6 $\sigma$).
Implicationd
============
Overriew
--------
Tve new VLBW image at $\nu$ = 1.52 GHz and $\theta$ = 15 mas (75 pc) filtets xor emissiou with a rest-frame frightness temkerature $C_b >
3.5 \tioes 10^5$ K, not zchievev by even thq most compacd starbursts [@ron92]. Moreoder, rhe CLBA imxee is too shalkow to deneet even tye most luminous redio aupernova beyonb $e \wim | above a threshold of $0.018$ mJy beam$^{-1}$ VLBA J1000+0206 was with the VLBA 2012 24 to 2013 11 under proposal BM360. A coordinate equinox of 2000 used. J1011+0106, with a one-dimensional position error of 2 mas (1 $\sigma$), was as a phase calibrator. The switching time between it and J1000+0206 was 360 and a angle 2.7. Every 4.1 s the correlator [@del11] produced 128 contiguous 2-MHz channels, yielding a total bandwidth 0.256 GHz per circular polarization centered at $\nu$ 1.52 GHz. VLBA system and gains were used to the scale to accuracy about A total of 450 baseline hours were accrued on J1000+0206. For each 6-hour session, NRAO’s Astronomical Image Processing System [AIPS; was used and edit data the described by @mid13. task [ imagr]{} was used to a naturally-weighted image of the Stokes $I\/$ emission J1000+0206. This spans 2048 $\times$ 1 mas in coordinate, has a geometric resolution at FWHM $\theta$ 15 mas, and was corrected for primary beam attenuation. Combining the u-v data from the resulted in an rms level $\sigma = mJy With signal-to-noise, as large about 10 $\theta$ $\sim$ 150 mas could be imaged. No emission detected above a threshold of $0.078$ mJy beam$^{-1}$ (6 $\sigma$). Overview The new VLBA at $\nu$ = 1.52 and = 15 mas (75 for with temperature > \times 10^5$ K, not by even the most compact [@con92]. Moreover, the VLBA detect even the most luminous radio supernova beyond \sim | above a threshold of $0.018$ mJy beam$^{-1}$ (3 $\sIgma$).
VLBA
----
J1000+0206 wAs obsErvEd wItH the vLBA During four 6-hour SEssiOns spanning 2012 June 24 to 2013 JanuaRy 11 undEr PRopoSAl Code Bm360. A coordINaTE EquInOx Of 2000 wAs USeD. J1011+0106, witH a oNe-dimenSional posiTioN eRror of 2 mas (1 $\sigMA$), wAs used as a pHasE calibrator. THe sWitchiNg TimE BetweEn iT and J1000+0206 Was 360 s, anD InvolvEd a switchInG Angle oF 2.7. every 4.1 s tHE CoRrelAtor [@del11] produced 128 coNTiGUous 2-MHz channelS, yieldInG A tOTAl bAndWidth of 0.256 GHz PeR circULar polaRIzATIOn cENtered at $\nu$ = 1.52 GHz. vLBA system tEMpeRatureS aNd gAIns werE used To SEt tHe amplitude ScalE to an accuRacy of ABout 5%. A toTAl of aboUt 450 baseLinE hoUrs wERe AcCruEd ON J1000+0206.
FOR eAch 6-HOur Session, NrAo’s astroNomiCAL iMage proCessIng SyStem [AIPS; @gre03] waS usEd to CAliBrate And edIt thE dAta foLlowinG the aPpRoach described bY @mid13. the AIPS taSk [ iMaGr]{} wAs Used tO Form anD deConVolve a nAturallY-WeiGhTED ImAge of the Stokes $I\/$ emiSsION fRom J1000+0206. This Image sPAnS 2048 $\tIMes$ 1 mas in EaCh cOordINAte, haS a geOMeTric resoLution AT FwHm $\theta$ = 15 mAs, And was CoRreCteD for pRImarY beam aTtenuatiOn. ComBIning the u-v data FRom the four sesSIoNS ReSUlteD in An rms noise lEvel $\SIgma = 0.013$ MJy bEAm$^{-1}$. witH AdequAte siGnAL-tO-Noise, structures as laRgE as aboUt 10 $\theTa$ $\sim$ 150 mas could Be imaged. No EMISsion was DeteCTeD Above a thresholD of $0.078$ mJY beam$^{-1}$ (6 $\sigma$).
iMplicatiOns
============
OvErview
--------
ThE new VLBA iMAGe at $\nu$ = 1.52 GHZ anD $\thEta$ = 15 Mas (75 PC) FiLters for emissION witH a Rest-fraMe bRightneSs tEmpEraTurE $T_B >
3.5 \times 10^5$ K, noT achieveD bY eVeN tHe mOst coMPact starBuRstS [@cOn92]. MOreovER, the VLbA imaGe is ToO sHAllOw to detECt EVEn thE mOsT lumInoUs Radio SupeRNovA beyond $Z \sim | above a threshold of $0.0 18$ mJy be am$^{ -1} $ ( 3$\si gma$ ).
VLBA
----J100 0+0206 was observed wi th th eV LBAd ur ing f our 6-h o ur s ess io ns sp an n in g 201 2 J une 24to 2013 Ja nua ry 11 under pr o po sal code B M36 0. A coordin ate equin ox of 2000was used . J101 1 +0106, with a o ne - dimens i onal po s i ti on e rror of 2 mas (1$ \s i gma$), was use d as a p h as e cal ibr ator. Thesw itchi n g timeb et w e e n i t and J1000+02 06 was 360s , a nd inv ol ved a swit ching a n gle of 2.7. Ev ery4.1 s the corre l ator [@ d el11] p roduce d 1 28cont i gu ou s 2 -M H z c h an nel s , y ieldingato tal b andw i d t h of0.2 56 G Hz pe r circular po lar izat i oncente red a t $\ nu $ = 1 .52 GH z. VL BA system tempera ture s and gai nswe reus ed to set th e a mpl itude s cale to anac c u r ac y of about 5%. A t ot a l o f about450 ba s el in e hours w er e a ccru e d on J 1000 + 02 06.
For each6 -h ou r sessi on , NRAO ’s As tro nomic a l Im age Pr ocessing Syst e m [AIPS; @gre0 3 ] was used to ca l i br a te a ndedit the da ta f o llow ingt he ap p roach desc ri b ed by @mid13. The AIPS t ask [imagr ]{} was usedto form an d d econvolv e an at u rally-weighted imag e of the S t okes $I\ /$ em ission f rom J1000 + 0 206. Thi s i mag e s pan s 20 48 $\times$ 1 m as i neach co ord inate,has ageo met ri c resolut ion at F WH M$\ th eta $ = 1 5 mas, an dwas c orr ected for pr imary bea mat t enu ation.C om b i ning t he u-v da ta from the fou r sessi ons resul ted in a nrm s noise level $\sigm a= 0.013$ m Jy be am$^{- 1 } $. Withadequate signal-to-nois e , struc tur es as lar ge as abo ut10 $\t het a $ $\si m$ 150 masco uld b e ima g e d. No e mission wa s det ected a bove a thre shold of $0.078$ m J y b eam$^{-1}$ (6 $\ sigm a $ ).
I m pl i cat io n s
= = = =========
Over view
----- -- -
The new VL B A i ma ge at $ \nu$ =1.52G Hz and$\theta$= 15 mas(7 5 pc ) fil ters for e missionwith a re s t-fra m ebrigh tne ss tem pe rat ure $ T_b >3 .5\time s 10^5 $K, not achi ev ed by ev en the most compact sta rburst s [@c on9 2]. Moreo ver , th e VLBA im ageis too sha llo w t o det ect eventhem os t l u minou s ra d io supern o va be y o nd $z \sim | above_a threshold_of $0.018$ mJy beam$^{-1}$ (3 $\sigma$).
VLBA
----
J1000+0206_was observed_with_the VLBA_during_four 6-hour sessions_spanning 2012 June_24 to 2013 January_11 under proposal_code_BM360. A coordinate equinox of 2000 was used. J1011+0106, with a one-dimensional position error_of_2 mas_(1_$\sigma$),_was used as a phase_calibrator. The switching time between_it and_J1000+0206 was 360 s, and involved a switching angle_of_2.7. Every 4.1_s the correlator [@del11] produced 128 contiguous 2-MHz channels,_yielding a total bandwidth of 0.256 GHz_per circular polarization_centered_at_$\nu$ = 1.52 GHz. VLBA_system temperatures and gains were used_to set the amplitude scale to_an accuracy of about 5%. A total_of about 450 baseline hours were_accrued on J1000+0206.
For each 6-hour_session, NRAO’s_Astronomical Image Processing System [AIPS;_@gre03] was used_to calibrate_and edit the_data following the approach described by_@mid13. The AIPS_task [ imagr]{} was used to_form_and deconvolve a_naturally-weighted_image_of the_Stokes $I\/$ emission_from_J1000+0206. This_image_spans 2048 $\times$ 1 mas in_each_coordinate, has a geometric resolution at FWHM_$\theta$ = 15 mas,_and_was corrected for primary_beam attenuation. Combining the u-v_data from the four sessions resulted_in an_rms noise_level $\sigma = 0.013$ mJy beam$^{-1}$. With adequate signal-to-noise, structures as large as_about 10 $\theta$ $\sim$ 150 mas_could be imaged. No_emission was_detected_above a threshold_of_$0.078$ mJy beam$^{-1}$ (6_$\sigma$).
Implications
============
Overview
--------
The new VLBA image at $\nu$ =_1.52 GHz_and $\theta$ = 15 mas (75_pc) filters for emission_with_a rest-frame brightness temperature $T_b >
3.5_\times 10^5$ K, not achieved by even_the most compact starbursts [@con92]._Moreover,_the_VLBA image is too shallow_to detect even the most luminous_radio supernova beyond_$z \sim |
+ \bar{d})$ was attributed to the neglect of the strangeness contribution. Our results are given in Table \[tab:results\_flavorsep\], and plotted in Fig. \[fig:flavor\_sep\] along with previous world DIS data and selected model predictions and parameterizations. The $(\Delta u + \Delta \bar{u})/(u + \bar{u})$ results, shown here for reference, are dominated by proton measurements.
$\langle x \rangle$ $\frac{\Delta u + \Delta \bar{u}}{u + \bar{u}} \pm \delta_{\textrm{stat}} \pm \delta_{\textrm{syst}}$ $\frac{\Delta d + \Delta \bar{d}}{d + \bar{d}} \pm \delta_{\textrm{stat}} \pm \delta_{\textrm{syst}}$
--------------------- ------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------
0.277 $0.423 \pm 0.011 \pm 0.031$ $-0.160 \pm 0.094 \pm 0.028$
0.325 $0.484 \pm 0.006 \pm 0.037$ $-0.283 \pm 0.055 \pm 0.032$
0.374 $0.515 \pm 0.005 \pm 0.044$ $-0.241 \pm 0.048 \pm 0.039$
0.424 $0.569 \pm 0.005 \pm 0.051$ $-0.499 \pm 0.054 \pm 0.051$
0.474 $0.595 \pm 0.006 \pm 0.063$ $-0.559 \pm 0.070 \pm 0.070$
0.548 $0.598 \pm 0.009 \pm 0.077$ $-0.356 \pm 0.014 \pm 0.097$
: $(\Delta u + \Delta \bar{u})/(u + \bar{u})$ and $(\Delta d + \Delta \bar{d})/(d + \bar{d})$ results. The reported systematic uncertainties include those due | + \bar{d})$ was attributed to the neglect of the strangeness contribution. Our results are give in Table \[tab: results\_flavorsep\ ], and plat in Fig. \[fig: flavor\_sep\ ] along with previous world DIS data and choose model predictions and parameterizations. The $ (\Delta u + \Delta \bar{u})/(u + \bar{u})$ results, picture here for reference, are dominated by proton measurement.
$ \langle x \rangle$ $ \frac{\Delta u + \Delta \bar{u}}{u + \bar{u } } \pm \delta_{\textrm{stat } } \pm \delta_{\textrm{syst}}$ $ \frac{\Delta d + \Delta \bar{d}}{d + \bar{d } } \pm \delta_{\textrm{stat } } \pm \delta_{\textrm{syst}}$
--------------------- ------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------
0.277 $ 0.423 \pm 0.011 \pm 0.031 $ $ -0.160 \pm 0.094 \pm 0.028 $
0.325 $ 0.484 \pm 0.006 \pm 0.037 $ $ -0.283 \pm 0.055 \pm 0.032 $
0.374 $ 0.515 \pm 0.005 \pm 0.044 $ $ -0.241 \pm 0.048 \pm 0.039 $
0.424 $ 0.569 \pm 0.005 \pm 0.051 $ $ -0.499 \pm 0.054 \pm 0.051 $
0.474 $ 0.595 \pm 0.006 \pm 0.063 $ $ -0.559 \pm 0.070 \pm 0.070 $
0.548 $ 0.598 \pm 0.009 \pm 0.077 $ $ -0.356 \pm 0.014 \pm 0.097 $
: $ (\Delta u + \Delta \bar{u})/(u + \bar{u})$ and $ (\Delta d + \Delta \bar{d})/(d + \bar{d})$ result. The reported taxonomic uncertainties include those due | + \bwr{d})$ was attributed to tht neglect of the strangeiess cohtributiun. Our results are given in Vablw \[tab:rtfults\_flavorsep\], and plotted pn Fig. \[fig:dlavie\_sep\] along with prcrious aorlb VIS data and sekected modal predictions avd parameterizations. The $(\Delta u + \Delea \bar{u})/(i + \bar{u})$ results, fhowm hers for reference, are dominated by pdoton mtasurements.
$\langke x \rangle$ $\frac{\Delta u + \Delha \bar{u}}{u + \bar{u}} \pm \felta_{\textrm{wtat}} \pm \delta_{\textfm{syst}}$ $\fgcc{\Delta d + \Selta \bar{d}}{d + \bar{d}} \pm \delta_{\textro{stat}} \pm \delta_{\tgrrrm{dist}}$
--------------------- ------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------
0.277 $0.423 \pm 0.011 \pm 0.031$ $-0.160 \pm 0.094 \km 0.028$
0.325 $0.484 \pm 0.006 \pm 0.037$ $-0.283 \pm 0.055 \pm 0.032$
0.374 $0.515 \pm 0.005 \pm 0.044$ $-0.241 \pm 0.048 \pi 0.039$
0.424 $0.569 \pm 0.005 \pm 0.051$ $-0.499 \pm 0.054 \po 0.051$
0.474 $0.595 \pm 0.006 \pm 0.063$ $-0.559 \om 0.070 \ik 0.070$
0.548 $0.598 \pm 0.009 \pm 0.077$ $-0.356 \pm 0.014 \pm 0.097$
: $(\Delta u + \Qelta \bar{u})/(u + \twr{u})$ and $(\Delte d + \Deltw \bae{d})/(d + \bar{d})$ rdrults. The repotted systvmctic unceetainties include bhose sue | + \bar{d})$ was attributed to the neglect strangeness Our results given in Table \[fig:flavor\_sep\] with previous world data and selected predictions and parameterizations. The $(\Delta u \Delta \bar{u})/(u + \bar{u})$ results, shown here for reference, are dominated by proton $\langle x \rangle$ $\frac{\Delta u + \Delta \bar{u}}{u + \bar{u}} \pm \delta_{\textrm{stat}} \pm $\frac{\Delta + \bar{d}}{d \bar{d}} \pm \delta_{\textrm{stat}} \pm \delta_{\textrm{syst}}$ --------------------- ------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------- 0.277 $0.423 \pm 0.011 \pm 0.031$ $-0.160 \pm \pm 0.028$ 0.325 $0.484 \pm 0.006 \pm 0.037$ \pm 0.055 \pm 0.032$ $0.515 \pm 0.005 \pm 0.044$ \pm \pm 0.039$ $0.569 0.005 0.051$ $-0.499 \pm \pm 0.051$ 0.474 $0.595 \pm 0.006 \pm 0.063$ $-0.559 \pm 0.070 \pm 0.070$ 0.548 $0.598 \pm 0.009 0.077$ $-0.356 \pm 0.097$ $(\Delta + \bar{u})/(u + \bar{u})$ d + \Delta \bar{d})/(d + \bar{d})$ systematic uncertainties include those due | + \bar{d})$ was attributed to the negLect of the sTrangEneSs cOnTribUtioN. Our results are GIven In Table \[tab:results\_flavoRsep\], aNd PLottED iN Fig. \[fIg:flavoR\_SeP\] ALonG wItH prEvIOuS worlD DIs data anD selected mOdeL pRedictions anD PaRameterizaTioNs. The $(\Delta u + \DEltA \bar{u})/(u + \BaR{u})$ rESults, ShoWn herE for reFErence, Are dominaTeD By protON measurEMEnTs.
$\laNgle x \rangle$ $\frac{\DeLTa U + \delta \bar{u}}{u + \bar{u}} \Pm \deltA_{\tEXtRM{StaT}} \pm \Delta_{\textrM{sYst}}$ $\frAC{\Delta d + \dElTA \BAr{d}}{D + \Bar{d}} \pm \delta_{\teXtrm{stat}} \pm \dELta_{\Textrm{SySt}}$
--------------------- ------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------
0.277 $0.423 \pM 0.011 \Pm 0.031$ $-0.160 \pm 0.094 \pm 0.028$
0.325 $0.484 \Pm 0.006 \pm 0.037$ $-0.283 \pM 0.055 \pM 0.032$
0.374 $0.515 \Pm 0.005 \pM 0.044$ $-0.241 \pm 0.048 \pm 0.039$
0.424 $0.569 \pm 0.005 \pm 0.051$ $-0.499 \pm 0.054 \Pm 0.051$
0.474 $0.595 \pm 0.006 \Pm 0.063$ $-0.559 \pm 0.070 \pm 0.070$
0.548 $0.598 \pm 0.009 \pM 0.077$ $-0.356 \pm 0.014 \pm 0.097$
: $(\DELta u + \DelTA \bar{u})/(u + \bAr{u})$ and $(\delTa d + \deltA \BaR{d})/(D + \baR{d})$ REsuLTs. the REpoRted systEmAtIc uncErtaINTIEs inCluDe thOse duE | + \bar{d})$ was attribute d to the n eglec t o f t he str ange ness contribut i on.Our results are givenin Ta bl e \[t a b: resul ts\_fla v or s e p\] ,an d p lo t te d inFig . \[fig :flavor\_s ep\ ]along with p r ev ious world DI S data and s ele cted m od elp redic tio ns an d para m eteriz ations. T he $(\Del t a u + \ D e lt a \b ar{u})/(u + \bar{ u }) $ results, show n here f o rr e fer enc e, are dom in atedb y proto n m e a s ure m ents.
$\l angle x \ra n gle $ $\ fr ac{ \ Deltau + \ De l ta\bar{u}}{u+ \b ar{u}} \p m \del t a_{\tex t rm{stat }} \pm \d elt a_{\ t ex tr m{s ys t }}$ $\f r ac{ \Delta d + \ Delta \ba r { d } }{d+ \ bar{ d}} \ pm \delta_{\t ext rm{s t at} } \pm \del ta_{ \t extrm {syst} }$
-- --------------- ---- -------- --- -- --- -- ----- - ------ --- --- ------- ------- - --- -- - - - -- ------------------ -- - - -- -------- ------ - -- -- -------- -- --- ---- - - ----- ---- - -- -------- ------ - -- -- ------- -- ------ -- --- --- ----- - ---- ------ -------- ----- 0.27 7 $0 .423 \p m 0 .011\pm 0. 0 31 $ $-0.160 \pm 0 .094\pm 0.028 $
0. 325 $0.484\ p m 0 .0 06 \pm 0. 037$ $ -0.283 \pm 0.0 55 \ p m 0.032$ 0 .3 74 $0 .515 \p m 0.005 \pm 0 .044$ $-0. 241 \pm 0.0 4 8 \pm 0.039 $
0.424 $0.569 \pm 0. 0 05\pm 0.051$ $ -0 .499 \p m 0.05 4 \pm 0.051$
0.474 $ 0. 595 \ pm 0.006 \ pm 0.06 3$ $- 0.5 59 \pm 0 .07 0 \p m 0.070$
0.5 48 $0 . 59 8 \ p m 0.009 \pm 0 . 0 77$ $-0.356 \p m 0. 0 1 4 \pm 0 .0 97$
: $( \Delta u + \Delt a \bar {u})/( u + \ba r { u} ) $ and$(\D elt a d + \De lta \ b ar{d})/ (d + \bar{d })$re sults. The r e port e d systematic unce rtain t i es in c lud e tho se due | +_\bar{d})$ was_attributed to the neglect_of the_strangeness_contribution. Our_results_are given in_Table \[tab:results\_flavorsep\], and plotted_in Fig. \[fig:flavor\_sep\] along with_previous world DIS_data_and selected model predictions and parameterizations. The $(\Delta u + \Delta \bar{u})/(u + \bar{u})$_results,_shown here_for_reference,_are dominated by proton measurements.
_ $\langle x \rangle$_ _$\frac{\Delta u + \Delta \bar{u}}{u + \bar{u}} \pm_\delta_{\textrm{stat}}_\pm \delta_{\textrm{syst}}$ _ $\frac{\Delta d + \Delta \bar{d}}{d + \bar{d}} \pm_\delta_{\textrm{stat}} \pm \delta_{\textrm{syst}}$
--------------------- -------------------------------------------------------------------------------------------------------_-------------------------------------------------------------------------------------------------------
___ _ 0.277 _ _ _ _ _ _ _ _ _ _ $0.423 \pm 0.011_\pm 0.031$_ __ ___ _ __ __ __ _ __ _ _ _ _ _ _ $-0.160 \pm _0.094 \pm 0.028$
_ __ __ _0.325 _ _ _ __ _ _ ___ _ $0.484 \pm 0.006_\pm 0.037$_ __ __ ___ _ _ _ _ ___ _ _ _ $-0.283 \pm 0.055 \pm_ 0.032$
_ 0.374 _ _ _ __ ____ _ _ _$0.515 \pm_ 0.005 \pm_ 0.044$_ __ _ _ __ _ _ _ __ _ _ _ _ _ $-0.241_\pm _0.048 \pm_ 0.039$
__ 0.424 ____ _ _ _ __ _ _ _ _ $0.569 \pm 0.005 \pm 0.051$ _ _ _ __ _ _ __ ___ __ _____ _ $-0.499 \pm _0.054_\pm 0.051$
_ __ _0.474 _ __ ____ _ __ _$0.595 \pm 0.006 \pm _0.063$_ _ _ _ _ _ _ _ __ _ _ _ $-0.559 \pm 0.070 \pm 0.070$
_0.548 _ _ _ _ _ _$0.598_\pm 0.009 \pm 0.077$_ _ _ __ _ _ _ _ _ _ _ _ $-0.356_\pm_ 0.014 \pm 0.097$
: $(\Delta u + \Delta \bar{u})/(u_+ \bar{u})$_and $(\Delta d + \Delta \bar{d})/(d + \bar{d})$ results._The reported systematic uncertainties include those due |
In the public health domain, *contact tracing* refers to the process of identification of contacts who may have come into contact with an infected victim and subsequent collection of further information about these contacts. By tracing the contacts of infected individuals, testing them for infection, treating the infected and tracing their contacts in turn, public health aims to reduce infections in the population[^17]. In practice, *contact tracing* is widely performed for diseases like sexually transmitted infections (including HIV) and virus infections (e.g. SARS-CoV and SARS-CoV-2/COVID-19). Despite some pioneering attempts in applying advanced ICT technologies, e.g. the *FluPhone* [@Yoneki11], *contact tracing* has mainly been implemented manually by medical personnel. Regardless, this approach has been proven effective in combating contiguous diseases because it can at least (1) interrupt ongoing transmission and reduce spread, alert contacts to the possibility of infection and offer preventive counseling or prophylactic care and (2) allow the medical professionals to learn about the epidemiology in a particular population.
![Quarantine Plan (taken from [@Ferretti20])[]{data-label="quarantine"}](flow)
Ferretti et al. [@Ferretti20] investigated the key parameters of epidemic spread for COVID-19 and concluded that viral spread is too fast to be contained by manual *contact tracing* (see an illustration in Fig. \[quarantine\]) but could be controlled if this process was faster, more efficient and happened at scale. This necessitates digital *contact tracing* solutions, such as those based on mobile apps, which can efficiently achieve epidemic control in large scale. Interestingly, this coincides with the empirical analysis of the control measures in China, by Tian et al. [@Tian20].
In the following-up discussions, we focus on digital *contact tracing* solutions, by cautiously referring to their manual ancestors in the aspect of functional requirements. For simplicity, we omit the term “digital" in the rest of the paper.
Preliminary on *contact tracing*
--------------------------------
Overall, the purpose of a *contact tracing* solution is to prevent an epidemic or a pandemic caused by a contagious virus or something similar. Depending on the standing point, it may serve (at least) two purposes.
- At a global level, it helps medical personnel to trace the pattern of virus spreading, produce transmission graphs, trace the origin of the virus, and so on. With | In the public health domain, * contact tracing * refers to the process of designation of contact who may have come into liaison with an infected victim and subsequent solicitation of further information about these contacts. By hound the liaison of septic person, testing them for contagion, treating the infected and tracing their contacts in go, public health aims to repress infections in the population[^17 ]. In practice, * liaison tracing * is widely performed for diseases like sexually impart infections (including HIV) and virus contagion (e.g. SARS - CoV and SARS - CoV-2 / COVID-19). Despite some pioneer attempts in applying advanced ICT technologies, e.g. the * FluPhone * [ @Yoneki11 ], * liaison tracing * has mainly been implemented manually by medical personnel. Regardless, this approach has been proven effective in battle conterminous diseases because it can at least (1) interrupt ongoing transmission and reduce spread, alert contacts to the possibility of infection and offer preventive counseling or contraceptive care and (2) allow the aesculapian master to determine about the epidemiology in a particular population.
! [ Quarantine Plan (taken from [ @Ferretti20])[]{data - label="quarantine"}](flow)
Ferretti et al. [ @Ferretti20 ] investigated the key parameters of epidemic spread for COVID-19 and concluded that viral bedspread is too fast to be contained by manual * contact tracing * (visualize an illustration in Fig. \[quarantine\ ]) but could be controlled if this process was faster, more efficient and happened at plate. This necessitates digital * liaison tracing * solution, such as those based on mobile apps, which can efficiently achieve epidemic control in large scale. Interestingly, this coincides with the empirical psychoanalysis of the control measure in China, by Tian et al. [ @Tian20 ].
In the following - up discussions, we concenter on digital * contact tracing * solution, by cautiously mention to their manual ancestors in the view of functional requirement. For simplicity, we neglect the term “ digital " in the rest of the paper.
Preliminary on * contact tracing *
--------------------------------
Overall, the purpose of a * liaison tracing * solution is to prevent an epidemic or a pandemic caused by a contagious virus or something similar. count on the standing point, it may serve (at least) two purposes.
- At a global level, it help medical personnel to decipher the pattern of virus spreading, grow transmission graphs, trace the lineage of the virus, and so on. With |
In hhe public health domain, *contact traciny* referv to tge procers of identification of contects who nay have come into congact with an infexted cictim and subsequcut compectnoi of further innormation atout these condaztd. By tracing the contacts of infectqd indififuals, testing jhem gjr ihfection, treating the infected and tracinj their contactx in turn, public health ails tl reduce infectiond in the pokhlaeuon[^17]. In practkce, *contacu cracing* is sidely performed for diseases lkke szxually trabsnithgd infectionw (incjuding HIV) akc viruv infecyions (e.g. SARS-CpV end WARS-CoV-2/COVID-19). Despite vome pioneering atjempts in dp'lying advanced ICT twcynolocies, e.g. gye *WluKhoie* [@Goneki11], *coitact tracihg* has mainoy been implemented mwbually by medjcal pqrfonnel. Regardless, this approach has beet pdoven effective in combqting contiguous disewses becatse it can at least (1) interrupt ongoing transmissimn anv fedbgc spfwaf, alert contacts to the possibility of infectykn akd offer preventlve counseling or lrlpnilactic care avd (2) allow the medical profedsionalf to oearn abott tne epidemiology in a particylar populatpon.
![Qyarantine Plan (takzn from [@Ferrztti20])[]{daja-labek="quarantine"}](flow)
Ferretti et zl. [@Ferretti20] investigzged the key paraoetvrs mf epidemic spread for COVYD-19 and coicludzd that xirak sprewd is too vast bm be contained by lanuap *wontact trwcing* (see an illustration in Fij. \[quarantine\]) nud cmuld be eontrokled if this [rocess was faxter, moxe effkcient and happenxd at scale. Ehis necessitdjes digital *cmntact twacibg* silutionr, such as those based on mobile apps, which can efficlentli zchieve epidemie xontrol in largr szalq. Pntxrestytgly, this cohnciaes eith ghe empiriccl anakysis of the control meaaures in China, by Yiwn et al. [@Tian20].
In ehe following-ip discussions, we vocus on digiyal *contact tracing* solutions, by dautiouslj rcferring to treir manual ancextors in the aspect of functional requicements. For simplicity, qe omit the term “diyiual" in the rext of the papes.
Preliminary on *contqct tracing*
--------------------------------
Overalk, the purpose of a *cohtact dracijg* solution is to prevent an epidemic or a pandemic caused by a contagiouw virus or somethihg somilag. Dzpeuding og thx rtanding point, in may serve (at least) two purposer.
- At a gjobal level, it helps medical pdrsonnel to ttace the pattern of virus apreadinb, produce transmission graphs, ttace the origin of the vlrus, and wo on. With | In the public health domain, *contact tracing* the of identification contacts who may an victim and subsequent of further information these contacts. By tracing the contacts infected individuals, testing them for infection, treating the infected and tracing their contacts turn, public health aims to reduce infections in the population[^17]. In practice, *contact is performed diseases sexually transmitted infections (including HIV) and virus infections (e.g. SARS-CoV and SARS-CoV-2/COVID-19). Despite some pioneering attempts applying advanced ICT technologies, e.g. the *FluPhone* [@Yoneki11], tracing* has mainly been manually by medical personnel. Regardless, approach been proven in contiguous because it can least (1) interrupt ongoing transmission and reduce spread, alert contacts to the possibility of infection and offer counseling or and (2) the professionals learn about the a particular population. ![Quarantine Plan (taken et al. [@Ferretti20] investigated the key parameters of spread for and concluded that viral spread is fast to be contained by manual *contact tracing* an illustration in Fig. \[quarantine\]) but could be controlled if this process was faster, more happened at scale. This digital *contact tracing* such those on apps, which efficiently achieve epidemic control in large scale. Interestingly, this coincides with empirical analysis of the control measures in China, by Tian [@Tian20]. the following-up discussions, focus on digital *contact solutions, cautiously referring to their in aspect For we the term “digital" in rest of the paper. Preliminary *contact tracing* -------------------------------- Overall, tracing* solution is to prevent an epidemic or pandemic caused by a contagious virus or similar. Depending on the standing point, it may serve (at least) two - At level, it helps medical personnel to trace the of virus spreading, produce graphs, trace the origin of the virus, and so With |
In the public health domain, *coNtact traciNg* refErs To tHe ProcEss oF identificatioN Of coNtacts who may have come inTo conTaCT witH An InfecTed victIM aND SubSeQuEnt CoLLeCtion Of fUrther iNformation AboUt These contactS. by Tracing the ConTacts of infecTed IndiviDuAls, TEstinG thEm for InfectIOn, treaTing the inFeCTed and TRacing tHEIr ContActs in turn, public hEAlTH aims to reduce iNfectiOnS In THE poPulAtion[^17]. In praCtIce, *coNTact traCInG* IS WidELy performed foR diseases liKE seXually TrAnsMItted iNfectIoNS (inCluding HIV) aNd viRus infectIons (e.g. saRS-CoV aND SARS-Cov-2/COVID-19). desPitE somE PiOnEerInG AttEMpTs iN AppLying advAnCeD ICT tEchnOLOGIes, e.G. thE *Fluphone* [@yoneki11], *contact TraCing* HAs mAinly Been iMpleMeNted mAnuallY by meDiCal personnel. RegArdlEss, this apProAcH haS bEen prOVen effEctIve In combaTing conTIguOuS DISeAses because it can at LeAST (1) iNterrupt OngoinG TrAnSMission aNd RedUce sPREad, alErt cONtActs to thE possiBIlItY of infeCtIon and OfFer PreVentiVE couNselinG or prophYlactIC care and (2) allow tHE medical profeSSiONAlS To leArn About the epiDemiOLogy In a pARtIcuLAr popUlatiOn.
![qUaRAntine Plan (taken from [@feRretti20])[]{Data-lAbel="quarantinE"}](flow)
FerreTTI Et al. [@FerrEtti20] INvEStigated the key ParamEters of epiDEmic spreAd for cOVID-19 and Concluded THAt viral sPreAd iS toO faST To Be contained by MANual *CoNtact trAciNg* (see an IllUstRatIon In fig. \[quaranTine\]) but cOuLd Be CoNtrOlled IF this proCeSs wAs FasTer, moRE efficIent aNd haPpEnED at Scale. ThIS nECEssiTaTeS digItaL *cOntacT traCIng* SolutioNs, such as tHosE BaseD oN mObile apPs, which can effIcIently achiEvE epIdemic CONtrol in lArge scale. Interestingly, tHIs coincIdeS with The eMpirical aNalYsis of The COntrol MeasurEs in CHiNa, bY tIan et AL. [@tiAn20].
IN tHe followinG-UP diScussIoNs, we Focus on Digital *contact tracINg* sOlutions, by cauTioUsly REFeRriNG tO TheIr MAnuAL Ancestors in the aSpect of funCtIOnAl requiremENts. foR simpliCity, we oMit thE Term “digItal" in the Rest of the PaPer.
PRELimInary on *conTact tracIng*
--------------------------------
OveralL, The puRPoSe of a *ConTact trAcIng* SolutIon is tO PreVent aN epideMiC or a paNdemiC cAused by a Contagious virus or somethIng simIlar. DEpeNding on thE stANdiNg point, it May sErve (at leasT) twO puRposeS.
- At A GlobaL levEL, iT heLPs medIcal PErsonnel tO TrAce THE pAttern of virUS SPreAding, ProDUce traNsmiSsion graphs, trace tHE origin of the viRus, aND So oN. WiTH |
In the public health doma in, *conta ct tr aci ng* r efer s to the process o f ide ntification of contact s who m a y ha v ecomeinto co n ta c t wi th a n i nf e ct ed vi cti m and s ubsequentcol le ction of fur t he r informat ion about these co ntacts .Byt racin g t he co ntacts of inf ected ind iv i duals, testing t he m fo r infection, trea t in g the infectedand tr ac i ng t hei r c ontacts in t urn,p ublic h e al t h aim s to reduce in fections in the popul at ion [ ^17].In pr ac t ice , *contacttrac ing* is w idelyp erforme d for di seases li kesexu a ll ytra ns m itt e dinf e cti ons (inc lu di ng HI V) a n d v irus in fect ions(e.g. SARS-Co V a nd S A RS- CoV-2 /COVI D-19 ). Desp ite so me pi on eering attempts inapplyingadv an ced I CT te c hnolog ies , e .g. the *FluPh o ne* [ @ Y o ne ki11], *contact tr ac i n g* has mai nly be e nim p lemented m anu ally b y med ical pe rsonnel. Regar d le ss , thisap proach h asbee n pro v en e ffecti ve in co mbati n g contiguous d i seases becaus e i t ca n atlea st (1) inte rrup t ong oing tr ans m issio n and r e du c e spread, alert con ta cts to thepossibility o f infectio n a nd offer pre v en t ive counseling or p rophylacti c care an d (2) allow t he medica l professi ona lstolea r n a bout the epid e m iolo gy in a p art icularpop ula tio n.
! [Quaranti ne Plan(t ak en f rom [@Fe r retti20] )[ ]{d at a-l abel= " quaran tine" }](f lo w)
Fe rrettie ta l . [@ Fe rr etti 20] i nvest igat e d t he keyparameter s o f epi de mi c sprea d for COVID-1 9and conclu de d t hat vi r a l spread is too fast to be cont a ined by ma nual*con tact trac ing * (see an illust ration in F ig . \ [ q uaran t i ne \]) b ut could b e con troll ed ifthis pr ocess was faster,m ore efficient an d h appe n e dats ca l e.Th i s n e c essitates digit al *contac tt ra cing* solu t ion s, such a s those base d on mob ile apps, which ca neffi c i ent ly achieve epidemi c control in la r ge scal e.Intere st ing ly, t his co i nci des w ith th eempiri cal a na lysis of the control measures i n Chin a, by Ti an et al. [@ T ian 20].
Inthefollowing- updis cussi ons , we f ocus on di g ital*con t act traci n g* so l u ti ons, by cau t i o usl y ref err i ng tothei r manual ancestor s in the aspect off u nct ion a l re qu irements. Forsim pl i c ity, weom it the term “digita l" in th e rest of th e paper .
P r elimin aryon*contacttra ci n g*
---- -- -- - ------ ---- -- ------ -----Over a l l, the purpose o f a * c o ntact tra cing* s olution is t o preventan epidemic or apand emiccausedby a con tag io us virus o r somethin g sim ilar. D ep endi ngon the sta n d ing p oint ,itmay serve ( at le as t ) t wo p urpos es .
- At a g l obal lev el, it help smed i c al per s on n e l to trace th e pat t e rn of viru s spr e ad i ng, p roduce trans mission gra ph s, trac e t h e origin o f the vir u s,an d so on. Wit h |
In the_public health_domain, *contact tracing* refers_to the_process_of identification_of_contacts who may_have come into_contact with an infected_victim and subsequent_collection_of further information about these contacts. By tracing the contacts of infected individuals, testing_them_for infection,_treating_the_infected and tracing their contacts_in turn, public health aims_to reduce_infections in the population[^17]. In practice, *contact tracing*_is_widely performed for_diseases like sexually transmitted infections (including HIV) and virus_infections (e.g. SARS-CoV and SARS-CoV-2/COVID-19). Despite_some pioneering attempts_in_applying_advanced ICT technologies, e.g._the *FluPhone* [@Yoneki11], *contact tracing* has_mainly been implemented manually by medical_personnel. Regardless, this approach has been proven_effective in combating contiguous diseases because_it can at least (1)_interrupt ongoing_transmission and reduce spread, alert_contacts to the_possibility of_infection and offer_preventive counseling or prophylactic care and_(2) allow the_medical professionals to learn about the_epidemiology_in a particular_population.
![Quarantine_Plan_(taken from_[@Ferretti20])[]{data-label="quarantine"}](flow)
Ferretti et al._[@Ferretti20]_investigated the_key_parameters of epidemic spread for COVID-19_and_concluded that viral spread is too fast_to be contained by_manual_*contact tracing* (see an_illustration in Fig. \[quarantine\]) but_could be controlled if this process_was faster,_more efficient_and happened at scale. This necessitates digital *contact tracing* solutions, such_as those based on mobile apps,_which can efficiently achieve_epidemic control_in_large scale. Interestingly,_this_coincides with_the empirical analysis of the control measures_in China,_by Tian et al. [@Tian20].
In the_following-up discussions, we focus_on_digital *contact tracing* solutions, by cautiously_referring to their manual ancestors in_the aspect of functional requirements._For_simplicity,_we omit the term “digital"_in the rest of the paper.
Preliminary_on *contact tracing*
--------------------------------
Overall,_the purpose of a *contact tracing* solution_is_to prevent an epidemic or a_pandemic_caused by a contagious virus or_something_similar._Depending on the standing point,_it may serve (at least) two_purposes.
- At a global level, it helps_medical personnel to_trace the pattern of virus_spreading,_produce_transmission graphs, trace the origin of the virus, and so_on. With |
{C}^{1,2,2}(\Lambda\times\mathbb{R})$ satisfies the equation $(4.1)$.$\Box$
[99]{}
R. Cont and D. A. Fourni, A functional extension of the Ito formula, C. R. Math. Acad. Sci. Paris 348 (2010), no. 1-2, 57-61.
R. Cont and D. A. Fourni, Change of variable formulas for non-anticipative functionals on path space, J. Funct. Anal. 259 (2010), no. 4, 1043-1072.
A. Cosso, Viscosity solutions of path-dependent PDEs and non-Markovian forward-backward stochastic systems, Preprint, 2012.
I. Ekren, C. Keller, N. Touzi and J. Zhang, On Viscosity Solutions of Path Dependent PDEs, arXiv:1109.5971, 2011.
[^1]: Institute for Financial Studies and Institute of Mathematics, Shandong University, Jinan, Shandong 250100, PR China ([email protected], Fax: +86 0531 88564100).
[^2]: School of mathematics, Shandong University, Jinan, Shandong 250100, PR China. ([email protected]).
[^3]: This work was supported by National Natural Science Foundation of China (No. 11171187, No. 10871118 and No. 10921101).
---
abstract: 'In this paper, we construct a higher rank Euler system for the multiplicative group over a totally real field by using the Iwasawa main conjecture proved by Wiles. A key ingredient of the construction is to generalize the notion of the characteristic ideal. Under certain technical assumptions, we prove that all higher Fitting ideals of a certain $p$-ramified Iwasawa module are described by analytic invariants canonically associated with Stickelberger elements.'
address: 'Graduate School of Mathematical Sciences, The University of Tokyo, 3-8-1 Komaba, Meguro-Ku, Tokyo, 153-8914, Japan'
author:
- Ryotaro Sakamoto
title: A higher rank Euler system for the multiplicative group over a totally real field
---
Introduction
============
Euler systems were introduced by Koly | { C}^{1,2,2}(\Lambda\times\mathbb{R})$ satisfies the equation $ (4.1)$.$\Box$
[ 99 ] { }
R. Cont and D. A. Fourni, A functional extension of the Ito rule, C. R. Math. Acad. Sci. Paris 348 (2010), no. 1 - 2, 57 - 61.
R. Cont and D. A. Fourni, variety of variable formulas for non - anticipant functionals on path distance, J. Funct. Anal. 259 (2010), no. 4, 1043 - 1072.
A. Cosso, Viscosity solutions of way - subject PDEs and non - Markovian forward - backward stochastic organization, Preprint, 2012.
I. Ekren, C. Keller, N. Touzi and J. Zhang, On Viscosity Solutions of Path Dependent PDEs, arXiv:1109.5971, 2011.
[ ^1 ]: Institute for Financial Studies and Institute of Mathematics, Shandong University, Jinan, Shandong 250100, PR China ([email protected], Fax: +86 0531 88564100).
[ ^2 ]: School of mathematics, Shandong University, Jinan, Shandong 250100, PR China. ([email protected]).
[ ^3 ]: This work was back by National Natural Science Foundation of China (No. 11171187, No. 10871118 and No. 10921101).
---
abstract:' In this paper, we construct a higher rank Euler system for the multiplicative group over a wholly real field by using the Iwasawa independent conjecture proved by Wiles. A cardinal ingredient of the structure is to generalize the notion of the characteristic ideal. Under certain technical assumptions, we prove that all higher Fitting ideal of a certain $ p$-ramified Iwasawa module are described by analytic invariants canonically associated with Stickelberger elements.'
address:' Graduate School of Mathematical Sciences, The University of Tokyo, 3 - 8 - 1 Komaba, Meguro - Ku, Tokyo, 153 - 8914, Japan'
author:
- Ryotaro Sakamoto
title: A higher rank Euler system for the multiplicative group over a totally real field
---
Introduction
= = = = = = = = = = = =
Euler systems were precede by Koly | {C}^{1,2,2}(\Lalbda\times\mathbb{R})$ satisfits the equation $(4.1)$.$\Box$
[99]{}
R. Conv and D. A. Fournk, A functional extension of vhe Uto firmula, C. R. Math. Acad. Szi. Paris 348 (2010), no. 1-2, 57-61.
R. Xont qnd D. A. Fondni, Chakye of yariaylx formulas for kon-anticipadive functionans ou path space, J. Funct. Anal. 259 (2010), no. 4, 1043-1072.
A. Cofso, Visvodity solutions of katr-depsndent PDEs and non-Markovian forwadd-backwerd stochastic xystems, Preprint, 2012.
I. Ekren, C. Kelper, N. Touzi and J. Xhang, On Viwcosyry Solutions of Path Dtpzndent PDEs, arXiv:1109.5971, 2011.
[^1]: Institute for Financial Studnes and Insjnrutf of Mathemavics, Srandong Univcgsity, Jhnan, Shsndong 250100, PR Chika ([email protected], Fax: +86 0531 88564100).
[^2]: School mf mathematics, Shagdong Unieexsity, Jinan, Shandong 250100, PE Chita. ([email protected]).
[^3]: This wock was suppkrted by Narional Natural Sciemcq Foundation or Chinw (Go. 11171187, No. 10871118 and No. 10921101).
---
abstract: 'In this paper, fe donstruct a higher rank Euler system for the multiplisative group over a totally real field by using tve Iwesxwa mwkb fonjecture proved by Wiles. A key ingredient os tne construction ls to generalize tne npjion of the chxractexjsfic ideal. Under cegtain tgchnicql assumpuions, we prove that all higher Futting idealf of a certain $p$-ramnfied Iwasawc moduke arr described by analytic invzriants canlnically zrsociated with Sgiclenberger tuements.'
address: 'Grwduate Sciool pf Mathdmatocal Ssiences, Thf Uniyarsity of Tokyo, 3-8-1 Klmaba, Maguro-Ku, Tonyo, 153-8914, Japan'
author:
- Ryotaro Sakamovp
title: A hignes rdnk Eulex systcm for the multyplicative groop over a totauly real fpeld
---
Introvuction
============
Euler systems were lntroduced bb Koly | {C}^{1,2,2}(\Lambda\times\mathbb{R})$ satisfies the equation $(4.1)$.$\Box$ [99]{} R. D. Fourni, A extension of the Acad. Paris 348 (2010), 1-2, 57-61. R. and D. A. Fourni, Change of formulas for non-anticipative functionals on path space, J. Funct. Anal. 259 (2010), no. 1043-1072. A. Cosso, Viscosity solutions of path-dependent PDEs and non-Markovian forward-backward stochastic systems, 2012. Ekren, Keller, Touzi and J. Zhang, On Viscosity Solutions of Path Dependent PDEs, arXiv:1109.5971, 2011. [^1]: Institute for Studies and Institute of Mathematics, Shandong University, Jinan, 250100, PR China ([email protected], +86 0531 88564100). [^2]: School mathematics, University, Jinan, 250100, China. [^3]: This work supported by National Natural Science Foundation of China (No. 11171187, No. 10871118 and No. 10921101). --- abstract: this paper, a higher Euler for multiplicative group over real field by using the Iwasawa by Wiles. A key ingredient of the construction to generalize notion of the characteristic ideal. Under technical assumptions, we prove that all higher Fitting of a certain $p$-ramified Iwasawa module are described by analytic invariants canonically associated with Stickelberger 'Graduate School of Mathematical The University of 3-8-1 Meguro-Ku, 153-8914, author: - Sakamoto title: A higher rank Euler system for the multiplicative group a totally real field --- Introduction ============ Euler systems were Koly | {C}^{1,2,2}(\Lambda\times\mathbb{R})$ satisfIes the equaTion $(4.1)$.$\BOx$
[99]{}
R. conT aNd D. A. fourNi, A functional eXTensIon of the Ito formula, C. R. MaTh. AcaD. SCI. ParIS 348 (2010), nO. 1-2, 57-61.
R. ConT and D. A. FOUrNI, chaNgE oF vaRiABlE formUlaS for non-AnticipatiVe fUnCtionals on paTH sPace, J. Funct. anaL. 259 (2010), no. 4, 1043-1072.
A. Cosso, VisCosIty solUtIonS Of patH-dePendeNt PDEs ANd non-MArkovian fOrWArd-bacKWard stoCHAsTic sYstems, Preprint, 2012.
I. EkREn, c. keller, N. Touzi anD J. ZhanG, ON viSCOsiTy SOlutions of paTh DepENdent PDeS, aRxIV:1109.5971, 2011.
[^1]: InSTitute for FinaNcial StudieS And institUtE of mAthemaTics, SHaNDonG University, jinaN, Shandong 250100, pR ChinA ([email protected], Fax: +86 0531 88564100).
[^2]: SChool oF maTheMatiCS, SHaNdoNg uNivERsIty, jInaN, ShandonG 250100, Pr CHina. ([email protected]).
[^3]: ThIs worK was supported By NAtioNAl NAturaL ScieNce FOuNdatiOn of ChIna (No. 11171187, no. 10871118 And No. 10921101).
---
abstract: 'In This Paper, we coNstRuCt a HiGher rANk EuleR sySteM for the MultiplICatIvE GROuP over a totally real fIeLD By Using the iwasawA MaIn COnjecturE pRovEd by wILes. A kEy inGReDient of tHe consTRuCtIon is to GeNeraliZe The NotIon of THe chAracteRistic idEal. UnDEr certain technICal assumptionS, We PROvE That All Higher FittiNg idEAls oF a ceRTaIn $p$-RAmifiEd IwaSaWA mODule are described by aNaLytic iNvariAnts canonicalLy associatED WIth StickElbeRGeR Elements.'
addresS: 'GradUate School OF MathemaTical sciences, the UniverSITy of TokyO, 3-8-1 KoMabA, MeGurO-kU, TOkyo, 153-8914, Japan'
authOR:
- ryotArO SakamoTo
tItle: A hiGheR raNk EUleR sYstem for tHe multipLiCaTiVe GroUp oveR A totally ReAl fIeLd
---
INtrodUCtion
============
EUler sYsteMs WeRE inTroduceD By kOLy | {C}^{1,2,2}(\Lambda\times\ mathbb{R}) $ sat isf ies t he e quat ion $(4.1)$.$\ B ox$
[99]{}
R. Cont and D . A.Fo u rni, Afunct ional e x te n s ion o fthe I t oformu la, C. R.Math. Acad . S ci . Paris 348( 20 10), no. 1 -2, 57-61.
R.Con t andD. A. Fourn i,Chang e of v a riable formulas f o r non- a nticipa t i ve fun ctionals on paths pa c e, J. Funct. A nal. 2 59 (2 0 1 0), no . 4, 1043- 10 72.
A . Cosso , V i s c osi t y solutions o f path-depe n den t PDEs a ndn on-Mar kovia nf orw ard-backwar d st ochasticsystem s , Prepr i nt, 201 2.
I. Ek ren , C. Ke ll er, N . To u zi an d J. Zhang,On V iscos ityS o l u tion s o f Pa th De pendent PDEs, ar Xiv: 1 109 .5971 , 201 1.
[^ 1]: I nstitu te fo rFinancial Studi es a nd Instit ute o f M at hemat i cs, Sh and ong Univer sity, J i nan ,S h a nd ong 250100, PR Chi na ( Js [email protected] u.cn,F ax :+ 86 053188 564 100) .
[^2] : Sc h oo l of mat hemati c s, S handong U nivers it y,Jin an, S h ando ng 250 100, PRChina . (yangsz@mail. s du.edu.cn).
[ ^3 ] : T h is w ork was suppor tedb y Na tion a lNat u ral S cienc eF ou n dation of China (No .111711 87, N o. 10871118 a nd No. 109 2 1 1 01).
-- -
ab s tr a ct: 'In this p aper, we constr u ct a hig her r ank Eule r systemf o r the mu lti pli cat ive g ro up over a tot a l ly r ea l field by usingthe Iw asa wama in conjec ture pro ve dby W ile s. Ak ey ingre di ent o f t he co n struct ion i s to g en e ral ize the no t i on o fth e ch ara ct erist ic i d eal . Under certaintec h nica las sumptio ns, we proveth at all hig he r F itting i deals of a certain $p$-ramified Iwasawa mo dulearedescribed by analy tic invari ants c anoni ca lly a ssoci a t ed wi th Stickelbe r g ereleme nt s.'address : 'Graduate School ofMathematicalSci ence s , T heU ni v ers it y of T okyo, 3-8-1 Kom aba, Megur o- K u, Tokyo, 15 3 -89 14 , Japan '
autho r:
-R yotaroSakamototitle: Ahi gher r ank Euler sys tem forthe multi p licat i ve grou p o ver ato tal ly re al fie l d
- --
I ntrodu ct ion
== ===== == ===
Eul er systems were introdu ced by Koly | {C}^{1,2,2}(\Lambda\times\mathbb{R})$ satisfies_the equation_$(4.1)$.$\Box$
[99]{}
R. Cont and D._A. Fourni,_A_functional extension_of_the Ito formula,_C. R. Math._Acad. Sci. Paris 348_(2010), no. 1-2,_57-61.
R._Cont and D. A. Fourni, Change of variable formulas for non-anticipative functionals on path_space,_J. Funct._Anal._259_(2010), no. 4, 1043-1072.
A. Cosso,_Viscosity solutions of path-dependent PDEs_and non-Markovian_forward-backward stochastic systems, Preprint, 2012.
I. Ekren, C. Keller,_N._Touzi and J._Zhang, On Viscosity Solutions of Path Dependent PDEs, arXiv:1109.5971,_2011.
[^1]: Institute for Financial Studies and_Institute of Mathematics,_Shandong_University,_Jinan, Shandong 250100, PR_China ([email protected], Fax: +86 0531 88564100).
[^2]:_School of mathematics, Shandong University, Jinan,_Shandong 250100, PR China. ([email protected]).
[^3]: This work_was supported by National Natural Science_Foundation of China (No. 11171187,_No. 10871118_and No. 10921101).
---
abstract: 'In_this paper, we_construct a_higher rank Euler_system for the multiplicative group over_a totally real_field by using the Iwasawa main_conjecture_proved by Wiles._A_key_ingredient of_the construction is_to_generalize the_notion_of the characteristic ideal. Under certain_technical_assumptions, we prove that all higher Fitting_ideals of a certain_$p$-ramified_Iwasawa module are described_by analytic invariants canonically associated_with Stickelberger elements.'
address: 'Graduate School of_Mathematical Sciences,_The University_of Tokyo, 3-8-1 Komaba, Meguro-Ku, Tokyo, 153-8914, Japan'
author:
- Ryotaro Sakamoto
title: A_higher rank Euler system for the_multiplicative group over a_totally real_field
---
Introduction
============
Euler_systems were introduced_by_Koly |
:numerical_setup}
---------------
All shown simulations are conducted on a 3D box with dimensions $\left[4741.12 M_\odot\times4741.12 M_\odot\times4741.12 M_\odot\right]$ with a grid spacing of $\Delta_{x,y,z}=74.08M_\odot$ on the coarsest grid level. For the chosen magnetar model of radius $R_*=9.26M_\odot$ ($\simeq {13.7\,\text{km}}$km) this corresponds to a $\left[512R_*\times 512R_*\times 512R_*\right]$ box with a grid spacing of $\Delta_{x,y,z}=8R_*$. For the low and high resolution tests we employ seven and eight additional levels of mesh refinement, each increasing the resolution by a factor of two and encompassing the central object, respectively. This means that the finest resolution of our models (close to the magnetar surface) are $\Delta_{x,y,z}^{\rm min}=0.0625\times R_* = 0.5787M_\odot$ and $\Delta_{x,y,z}^{\rm min}=0.03125\times R_*= 0.2894M_\odot$ for the low and high resolution models, or in other words 16 and 32 points per $R_*$, respectively. The initial data is evolved for a period of $t=1185.28M_\odot\simeq 5.84\,$ms, which is chosen to be well below the dynamical timescale of the magnetar crust, which can be considered as a fixed boundary (see section \[sec:timecales\]).
In order to ensure the conservation properties of the algorithm, it is critical to employ [*refluxing*]{} techniques correcting numerical fluxes across different levels of mesh refinement [see, e.g. @Collins2010]. Specifically, we make use of the thorn `Refluxing`[^6] in combination with a cell-centered refinement structure [cf. @Shibata2015]. We highlight the fact that employing the refluxing algorithm makes the numerical code $2-4$ times slower for the benefit of enforcing the conservation properties of the numerical method (specially of the charge). Refluxing also reduces the numerical instabilities, which tend to develop at mesh refinement boundaries.
In conservative schemes, numerical reconstruction algorithms [we employ an MP7 scheme, cf. @S | : numerical_setup }
---------------
All shown simulations are conducted on a 3D box with dimension $ \left[4741.12 M_\odot\times4741.12 M_\odot\times4741.12 M_\odot\right]$ with a power system spacing of $ \Delta_{x, y, z}=74.08M_\odot$ on the coarsest power system level. For the choose magnetar model of radius $ R_*=9.26M_\odot$ ($ \simeq { 13.7\,\text{km}}$km) this match to a $ \left[512R_*\times 512R_*\times 512R_*\right]$ box with a grid space of $ \Delta_{x, yttrium, z}=8R_*$. For the low and high solution tests we employ seven and eight additional level of mesh refinement, each increasing the resolution by a factor of two and encompassing the cardinal object, respectively. This means that the finest resoluteness of our models (close to the magnetar open) are $ \Delta_{x, y, z}^{\rm min}=0.0625\times R _ * = 0.5787M_\odot$ and $ \Delta_{x, y, z}^{\rm min}=0.03125\times R_*= 0.2894M_\odot$ for the abject and high resolution models, or in other words 16 and 32 point per $ R_*$, respectively. The initial data is evolved for a period of $ t=1185.28M_\odot\simeq 5.84\,$ms, which is chosen to be well below the dynamical timescale of the magnetar crust, which can be regard as a fixed boundary (see section \[sec: timecales\ ]).
In order to ensure the conservation properties of the algorithm, it is critical to employ [ * refluxing * ] { } techniques correcting numeric fluxes across different floor of net refinement [ see, e.g. @Collins2010 ]. Specifically, we do use of the thorn 'Refluxing`[^6 ] in combination with a cell - centered refinement structure [ californium. @Shibata2015 ]. We highlight the fact that employing the refluxing algorithm makes the numerical code $ 2 - 4 $ times slower for the benefit of enforcing the conservation place of the numerical method (specially of the charge). Refluxing also reduces the numeric instabilities, which tend to develop at mesh refinement boundaries.
In conservative schemes, numerical reconstruction algorithms [ we use an MP7 scheme, cf. @S | :numfrical_setup}
---------------
All shown simmlations are conbycted mn a 3D box witf dimensions $\left[4741.12 M_\odot\times4741.12 M_\idot\tumes4741.12 M_\odot\right]$ with a grid spabing of $\Dwlta_{e,y,z}=74.08M_\odot$ on the rkarsest grid mcvel. Yoc the chosen mannetar moden of radius $R_*=9.26M_\mdut$ ($\simeq {13.7\,\text{km}}$km) this corresponds to a $\left[512T_*\tlmes 512R_*\times 512R_*\ryght]$ fox sptm a grid spacing of $\Delta_{x,y,z}=8R_*$. Fod the lmw and high rrsolution tests we employ devej and eight additilnal levels of iwsh refinemevt, each inbxeasing the resolution by a factor of two xnd eucompassing tye fgntral objecv, respvctively. This means tvat the finest resolubion mf iur models (close to tie magnetar surface) wre $\Delta_{f,y,v}^{\rm min}=0.0625\times R_* = 0.5787M_\odot$ abd $\Denta_{x,f,z}^{\rm nin}=0.03125\gimts C_*= 0.2894M_\kdot$ flr vhe low and high resolytion models, or in ptrvt words 16 and 32 poines per $R_*$, respectively. The initial data is evklved for a period of $t=1185.28N_\odot\simeq 5.84\,$ms, which id chosen eo be well below the dynamical timescale of the mdgneter crbwt, whkxh can be considered as a fixed boundary (see sesfipn \[xec:timecales\]).
In order to emskrr the conservajion propedties of the algorlthm, it is ceitical tj emlloy [*refluxing*]{} techniques cirrecting nuiwrical fluxes acrods differenc leveks of mesh refinement [see, e.g. @Colmins2010]. Speciflcally, we oake use of the ghogn `Rafluxing`[^6] in combination wieh a cell-rentexed refivememt strtcture [cf. @Dhibabd2015]. We highlight the fact tvat employlng the refluxing algorithm makxx the numerivan cmde $2-4$ timzs sloeer for the bqnefit of enfotcing the consdrvation pgoperties of the numqrical method (dpecially of the chawge). Eeflyxing auro reduces the numerical instabiluties, which tend tp dgvslop at mesh reynntmwnt boundaries.
On zonfegvavive fwhemes, numerhcal rezpnstrjction algovitfms [ee employ an MP7 scheke, cr. @S | :numerical_setup} --------------- All shown simulations are conducted 3D with dimensions M_\odot\times4741.12 M_\odot\times4741.12 M_\odot\right]$ $\Delta_{x,y,z}=74.08M_\odot$ the coarsest grid For the chosen model of radius $R_*=9.26M_\odot$ ($\simeq {13.7\,\text{km}}$km) corresponds to a $\left[512R_*\times 512R_*\times 512R_*\right]$ box with a grid spacing of $\Delta_{x,y,z}=8R_*$. the low and high resolution tests we employ seven and eight additional levels mesh each the by a factor of two and encompassing the central object, respectively. This means that the finest of our models (close to the magnetar surface) $\Delta_{x,y,z}^{\rm min}=0.0625\times R_* = and $\Delta_{x,y,z}^{\rm min}=0.03125\times R_*= 0.2894M_\odot$ the and high models, in words 16 and points per $R_*$, respectively. The initial data is evolved for a period of $t=1185.28M_\odot\simeq 5.84\,$ms, which is to be the dynamical of magnetar which can be a fixed boundary (see section \[sec:timecales\]). ensure the conservation properties of the algorithm, it critical to [*refluxing*]{} techniques correcting numerical fluxes across levels of mesh refinement [see, e.g. @Collins2010]. Specifically, make use of the thorn `Refluxing`[^6] in combination with a cell-centered refinement structure [cf. @Shibata2015]. the fact that employing refluxing algorithm makes numerical $2-4$ slower the benefit enforcing the conservation properties of the numerical method (specially of the Refluxing also reduces the numerical instabilities, which tend to develop refinement In conservative schemes, reconstruction algorithms [we employ MP7 cf. @S | :numerical_setup}
---------------
All shown simUlations arE condUctEd oN a 3d box With Dimensions $\left[4741.12 m_\Odot\Times4741.12 M_\odot\times4741.12 M_\odot\riGht]$ wiTh A Grid SPaCing oF $\Delta_{x,Y,Z}=74.08M_\ODOt$ oN tHe CoaRsESt Grid lEveL. For the Chosen magnEtaR mOdel of radius $r_*=9.26m_\oDot$ ($\simeq {13.7\,\teXt{kM}}$km) this correSpoNds to a $\LeFt[512R_*\TImes 512R_*\TimEs 512R_*\riGht]$ box WIth a grId spacing Of $\dElta_{x,y,Z}=8r_*$. For the LOW aNd hiGh resolution tests WE eMPloy seven and eiGht addItIOnAL LevEls Of mesh refiNeMent, eACh increASiNG THe rESolution by a faCtor of two anD EncOmpassInG thE CentraL objeCt, REspEctively. ThiS meaNs that the Finest REsolutiON of our mOdels (cLosE to The mAGnEtAr sUrFAce) ARe $\delTA_{x,y,Z}^{\rm min}=0.0625\tiMeS R_* = 0.5787m_\odot$ And $\DELTA_{X,y,z}^{\rM miN}=0.03125\timEs R_*= 0.2894M_\oDot$ for the low aNd hIgh rESolUtion ModelS, or iN oTher wOrds 16 anD 32 poinTs Per $R_*$, respectivelY. The Initial daTa iS eVolVeD for a PEriod oF $t=1185.28M_\OdoT\simeq 5.84\,$mS, which iS ChoSeN TO Be Well below the dynamiCaL TImEscale of The magNEtAr CRust, whicH cAn bE conSIDered As a fIXeD boundarY (see seCTiOn \[Sec:timeCaLes\]).
In oRdEr tO enSure tHE conServatIon propeRties OF the algorithm, iT Is critical to eMPlOY [*ReFLuxiNg*]{} tEchniques coRrecTIng nUmerICaL flUXes acRoss dIfFErENt levels of mesh refinEmEnt [see, E.g. @ColLins2010]. SpecificaLly, we make uSE OF the thorN `RefLUxINg`[^6] in combinatioN with A cell-centeREd refineMent sTructure [Cf. @Shibata2015]. wE HighlighT thE faCt tHat EMPlOying the refluXINg alGoRithm maKes The numeRicAl cOde $2-4$ TimEs Slower for The benefIt Of EnFoRciNg the COnservatIoN prOpErtIes of THe numeRical MethOd (SpECiaLly of thE ChARGe). ReFlUxIng aLso ReDuces The nUMerIcal insTabilitieS, whICh teNd To Develop At mesh refinemEnT boundarieS.
IN coNservaTIVe schemeS, numerical reconstructioN AlgoritHms [We empLoy aN MP7 scheme, Cf. @S | :numerical_setup}
-------- -------
A ll sh own si mu lati onsare conductedo n a3D box with dimensions $\le ft [ 4741 . 12 M_\o dot\tim e s4 7 4 1.1 2M_ \od ot \ ti mes47 41. 12 M_\o dot\right] $ w it h a grid spa c in g of $\Del ta_ {x,y,z}=74.0 8M_ \odot$ o n t h e coa rse st gr id lev e l. For the chos en magnet a r model o fradi us $R_*=9.26M_\od o t$ ($\simeq {13.7 \,\tex t{ k m} } $ km) th is corresp on ds to a $\lef t [5 1 2 R _*\ t imes 512R_*\t imes 512R_* \ rig ht]$ b ox wi t h a gr id sp ac i ngof $\Delta_ {x,y ,z}=8R_*$ . Fort he lowa nd high resol uti ontest s w eemp lo y se v en an d ei ght addi ti on al le vels o f mesh re fine ment, each increas ing the res oluti on by a f ac tor o f twoand e nc ompassing the c entr al object , r es pec ti vely. This m ean s t hat the finest res ol u t i on of our models (cl os e to the mag netars ur fa c e) are $ \D elt a_{x , y ,z}^{ \rmm in }=0.0625 \times R_ *= 0.578 7M _\odot $and $\ Delta _ {x,y ,z}^{\ rm min}= 0.031 2 5\times R_*= 0 . 2894M_\odot$f or t he lowand high resol utio n mod els, or in other word s1 6a nd 32 points per $R _* $, res pecti vely. The ini tial datai s evolvedfora p e riod of $t=118 5.28M _\odot\sim e q 5.84\, $ms,which is chosen t o be wellbel owthe dy n a mi cal timescale o f th emagneta r c rust, w hic h c anbeco nsideredas a fix ed b ou nd ary (see section\[ sec :t ime cales \ ]).
I n ord er t oen s ure the co n se r v atio npr oper tie sof th e al g ori thm, it is criti cal to e mp lo y [*ref luxing*]{} te ch niques cor re cti ng num e r ical flu xes across different le v els ofmes h ref inem ent [see, e. g. @Co lli n s2010] . Spec ifica ll y,w e make u se of t he thorn ` R e flu xing` [^ 6] i n combi nation with a cell - cen tered refinem ent str u c tu re[ cf . @S hi b ata 2 0 15]. We highlig ht the fac tt ha t employin g th erefluxi ng algo rithm makes t he numeri cal code$2 -4$t i mes slower fo r the be nefit ofe nforc i ng thecon servat io n p roper ties o f th e num erical m ethod(spec ia lly of t he charge). Refluxing a lso re duces th e numeric ali nst abilities , wh ich tend t o d eve lop a t m e sh re fine m en t b o undar ies.
In conse r va tiv e sc hemes, nume r i c alrecon str u ctionalgo rithms [we employ an MP7 scheme, cf. @ S | :numerical_setup}
---------------
All shown_simulations are_conducted on a 3D_box with_dimensions_$\left[4741.12 M_\odot\times4741.12_M_\odot\times4741.12_M_\odot\right]$ with a_grid spacing of_$\Delta_{x,y,z}=74.08M_\odot$ on the coarsest_grid level. For_the_chosen magnetar model of radius $R_*=9.26M_\odot$ ($\simeq {13.7\,\text{km}}$km) this corresponds to a $\left[512R_*\times 512R_*\times_512R_*\right]$_box with_a_grid_spacing of $\Delta_{x,y,z}=8R_*$. For the_low and high resolution tests_we employ_seven and eight additional levels of mesh refinement,_each_increasing the resolution_by a factor of two and encompassing the central_object, respectively. This means that the_finest resolution of_our_models_(close to the magnetar_surface) are $\Delta_{x,y,z}^{\rm min}=0.0625\times R_* =_0.5787M_\odot$ and $\Delta_{x,y,z}^{\rm min}=0.03125\times R_*= 0.2894M_\odot$_for the low and high resolution models,_or in other words 16 and_32 points per $R_*$, respectively._The initial_data is evolved for a_period of $t=1185.28M_\odot\simeq_5.84\,$ms, which_is chosen to_be well below the dynamical timescale_of the magnetar_crust, which can be considered as_a_fixed boundary (see_section \[sec:timecales\]).
In_order_to ensure_the conservation properties_of_the algorithm,_it_is critical to employ [*refluxing*]{} techniques_correcting_numerical fluxes across different levels of mesh_refinement [see, e.g. @Collins2010]._Specifically,_we make use of_the thorn `Refluxing`[^6] in combination_with a cell-centered refinement structure [cf._@Shibata2015]. We_highlight the_fact that employing the refluxing algorithm makes the numerical code $2-4$_times slower for the benefit of_enforcing the conservation properties_of the_numerical_method (specially of_the_charge). Refluxing_also reduces the numerical instabilities, which tend_to develop_at mesh refinement boundaries.
In conservative schemes,_numerical reconstruction algorithms [we_employ_an MP7 scheme, cf. @S |
DCT). There are several types of DCT’s. We used a boundary including DCT, which is sometimes referred to as the *DCT of the first type* (DCT-1). The DCT-1 of a time-dependent function, $g(t)$, sampled at equidistant time points, $$t_k, \qquad k=0,1,\ldots,N_t$$ is defined as: $$\begin{aligned}
& \bar{g}(\omega_j) = \sqrt{\frac{2}{N_t}}\sum_{k=0}^{N_t}\frac{1}{h_k} g(t_k)\cos\left(\frac{jk\pi}{N_t}\right), & &j=0,1,\ldots,N_t \label{eq:DCT}\\
& h_k =
\begin{cases}
2 & \qquad k=0 \text{ or } k=N_t \\
1 & \qquad 1\leq k\leq N_t-1
\end{cases} \label{eq:hdef}\end{aligned}$$ The inverse transform is defined by $$\begin{aligned}
& g(t_k) = \sqrt{\frac{2}{N_t}}\sum_{j=0}^{N_t}\frac{1}{h_j} \bar{g}(\omega_j)\cos\left(\frac{jk\pi}{N_t}\right), & & k=0,1,\ldots,N_t \label{eq:iDCT}\end{aligned}$$ It can be observed from Eqs. , that DCT-1 is its own inverse.
The argument of the cosine function in Eqs. , is equivalent to a discretized variant of the continuous argument $\omega t$ in the continuous transform, where the variables are discretized as follows: $$\begin{aligned}
& t_k = k\frac{T}{N_t} & k=0,1,\ldots,N_t \label{eq:t_samp}\\
& \omega_j = j\frac{\pi}{T} & j=0,1,\ldots,N_t \label{eq:w_samp}\end{aligned}$$
In order to be consistent with the continuous formulation, | DCT). There are several types of DCT ’s. We used a limit include DCT, which is sometimes referred to as the * DCT of the first character * (DCT-1). The DCT-1 of a time - subject function, $ g(t)$, sampled at equidistant clock time points, $ $ t_k, \qquad k=0,1,\ldots, N_t$$ is defined as: $ $ \begin{aligned }
& \bar{g}(\omega_j) = \sqrt{\frac{2}{N_t}}\sum_{k=0}^{N_t}\frac{1}{h_k } g(t_k)\cos\left(\frac{jk\pi}{N_t}\right), & & j=0,1,\ldots, N_t \label{eq: DCT}\\
& h_k =
\begin{cases }
2 & \qquad k=0 \text { or } k = N_t \\
1 & \qquad 1\leq k\leq N_t-1
\end{cases } \label{eq: hdef}\end{aligned}$$ The inverse transform is specify by $ $ \begin{aligned }
& g(t_k) = \sqrt{\frac{2}{N_t}}\sum_{j=0}^{N_t}\frac{1}{h_j } \bar{g}(\omega_j)\cos\left(\frac{jk\pi}{N_t}\right), & & k=0,1,\ldots, N_t \label{eq: iDCT}\end{aligned}$$ It can be observed from Eqs. , that DCT-1 is its own inverse.
The argument of the cosine routine in Eqs. , is equivalent to a discretized variant of the continuous argument $ \omega t$ in the continuous transform, where the variable are discretized as follow: $ $ \begin{aligned }
& t_k = k\frac{T}{N_t } & k=0,1,\ldots, N_t \label{eq: t_samp}\\
& \omega_j = j\frac{\pi}{T } & j=0,1,\ldots, N_t \label{eq: w_samp}\end{aligned}$$
In order to be consistent with the continuous formulation, | DCT). There are several types of DCT’s. We useb a bouidary ihcluding DCT, which is sometimes refecred to aw the *DCT of the first type* (DCT-1). The DCT-1 of e time-dependent hhnction, $g(t)$, sajiled ct equidistant tlme points, $$d_k, \qquad k=0,1,\ldotv,N_g$$ ns defined as: $$\begin{aligned}
& \bar{g}(\oiega_j) = \xqgt{\frac{2}{N_t}}\sum_{k=0}^{N_t}\srac{1}{n_h} g(t_i)\bow\left(\frac{jk\pi}{N_t}\right), & &j=0,1,\ldots,N_f \label{tq:DCT}\\
& h_k =
\brgin{cases}
2 & \qquad k=0 \texh{ or } k=N_t \\
1 & \qquad 1\leq k\lez N_t-1
\end{cares} \label{eq:hdef}\end{alighed}$$ The inverse transform is dewined by $$\begin{aoitnef}
& g(t_k) = \swrt{\frwc{2}{N_t}}\sum_{j=0}^{N_t}\frac{1}{h_j} \bar{c}(\omega_j)\vos\left(\frac{jk\pl}{N_t}\rijht), & & k=0,1,\ldots,N_t \label{eq:iDCV}\end{aligned}$$ It can bg observed fxom Eqs. , that DCT-1 is irs own hnvesse.
Tfw afgujeit kf the coaine functjon in Eqs. , us equivalent to a cifbtetized variznt of tre continuous argument $\omega t$ in the cmntjnuous transform, where rhe variables are disfretized ws follows: $$\begin{aligned}
& t_k = k\frac{T}{N_t} & k=0,1,\ldots,T_t \laueu{eq:u_sqmp}\\
& \lmega_j = j\frac{\pi}{T} & j=0,1,\ldots,N_t \label{eq:w_samp}\end{alidhec}$$
Ik order to be coksistent with the vojtoguous formulajion, | DCT). There are several types of DCT’s. a including DCT, is sometimes referred the type* (DCT-1). The of a time-dependent $g(t)$, sampled at equidistant time points, \qquad k=0,1,\ldots,N_t$$ is defined as: $$\begin{aligned} & \bar{g}(\omega_j) = \sqrt{\frac{2}{N_t}}\sum_{k=0}^{N_t}\frac{1}{h_k} g(t_k)\cos\left(\frac{jk\pi}{N_t}\right), & &j=0,1,\ldots,N_t & h_k = \begin{cases} 2 & \qquad k=0 \text{ or } k=N_t \\ & 1\leq N_t-1 \label{eq:hdef}\end{aligned}$$ The inverse transform is defined by $$\begin{aligned} & g(t_k) = \sqrt{\frac{2}{N_t}}\sum_{j=0}^{N_t}\frac{1}{h_j} \bar{g}(\omega_j)\cos\left(\frac{jk\pi}{N_t}\right), & & k=0,1,\ldots,N_t It can be observed from Eqs. , that is its own inverse. argument of the cosine function Eqs. is equivalent a variant the continuous argument t$ in the continuous transform, where the variables are discretized as follows: $$\begin{aligned} & t_k = k\frac{T}{N_t} k=0,1,\ldots,N_t \label{eq:t_samp}\\ = j\frac{\pi}{T} j=0,1,\ldots,N_t In to be consistent continuous formulation, | DCT). There are several types of dCT’s. We used A bounDarY inClUdinG DCT, Which is sometimES refErred to as the *DCT of the fiRst tyPe* (dcT-1). ThE dCt-1 of a tIme-depeNDeNT FunCtIoN, $g(t)$, SaMPlEd at eQuiDistant Time points, $$T_k, \qQuAd k=0,1,\ldots,N_t$$ is DEfIned as: $$\begiN{alIgned}
& \bar{g}(\omeGa_j) = \Sqrt{\frAc{2}{n_t}}\sUM_{k=0}^{N_t}\fRac{1}{H_k} g(t_k)\Cos\lefT(\Frac{jk\Pi}{N_t}\right), & &J=0,1,\lDOts,N_t \lABel{eq:DCt}\\
& H_K =
\bEgin{Cases}
2 & \qquad k=0 \text{ or } K=n_t \\
1 & \QQuad 1\leq k\leq N_t-1
\eNd{caseS} \lABeL{EQ:hdEf}\eNd{aligned}$$ THe InverSE transfORm IS DEfiNEd by $$\begin{aligNed}
& g(t_k) = \sqrt{\fRAc{2}{N_T}}\sum_{j=0}^{N_T}\fRac{1}{H_J} \bar{g}(\oMega_j)\CoS\LefT(\frac{jk\pi}{N_t}\RighT), & & k=0,1,\ldots,N_t \Label{eQ:IDCT}\end{ALigned}$$ IT can be ObsErvEd frOM EQs. , ThaT Dct-1 is ITs Own INveRse.
The arGuMeNt of tHe coSINE FuncTioN in EQs. , is eQuivalent to a dIscRetiZEd vArianT of thE conTiNuous ArgumeNt $\omeGa T$ in the continuouS traNsform, wheRe tHe VarIaBles aRE discrEtiZed As folloWs: $$\begin{ALigNeD}
& T_K = K\fRac{T}{N_t} & k=0,1,\ldots,N_t \labeL{eQ:T_SaMp}\\
& \omega_j = J\frac{\pI}{t} & j=0,1,\LdOTs,N_t \labeL{eQ:w_sAmp}\eND{AlignEd}$$
In ORdEr to be coNsisteNT wItH the conTiNuous fOrMulAtiOn, | DCT). There are several ty pes of DCT ’s. W e u sed a bou ndar y including DC T , wh ich is sometimes refer red t oa s th e * DCT o f the f i rs t typ e* ( DCT -1 ) .The D CT- 1 of atime-depen den tfunction, $g ( t) $, sampled at equidistant ti me poi nt s,$ $t_k, \q quadk=0,1, \ ldots, N_t$$ isde f ined a s : $$\be g i n{ alig ned}
& \bar{g } (\ o mega_j) = \sqr t{\fra c{ 2 }{ N _ t}} \su m_{k=0}^{N _t }\fra c {1}{h_k } g ( t _ k)\ c os\left(\frac {jk\pi}{N_t } \ri ght),&&j= 0 ,1,\ld ots,N _t \la bel{eq:DCT} \\
& h_k=
\ begin{c a ses}
2 & \q quad k= 0\te xt { or }k=N _ t \ \
1& \qq uad1 \ l e q k\ leq N_t -1
\end{cases} \l abel { eq: hdef} \end{ alig ne d}$$The in verse t ransform is def ined by $$\be gin {a lig ne d}
& g(t _k) =\sqrt{\ frac{2} { N_t }} \ s u m_ {j=0}^{N_t}\frac{1 }{ h _ j} \bar{g} (\omeg a _j )\ c os\left( \f rac {jk\ p i }{N_t }\ri g ht ), & & k =0,1,\ l do ts ,N_t \l ab el{eq: iD CT} \en d{ali g ned} $$ Itcan be o bserv e d from Eqs. ,t hat DCT-1 isi ts o wn inve rse .
The argu ment of t he c o si nef uncti on in E q s. , is equivalent toadiscre tized variant of t he continu o u s argumen t $\ o me g a t$ in the co ntinu ous transf o rm, wher e the variabl es are di s c retizedasfol low s:$ $ \b egin{aligned} & t _k = k\ fra c{T}{N_ t}& k=0 ,1, \l dots,N_t\label{e q: t_ sa mp }\\
& \omega_ j= j \f rac {\pi} { T} & j =0,1, \ldo ts ,N _ t \ label{e q :w _ s amp} \e nd {ali gne d} $$
I n or d erto be c onsistent wi t h th eco ntinuou s formulation , | DCT). There_are several_types of DCT’s. We_used a_boundary_including DCT,_which_is sometimes referred_to as the_*DCT of the first_type* (DCT-1). The_DCT-1_of a time-dependent function, $g(t)$, sampled at equidistant time points, $$t_k, \qquad k=0,1,\ldots,N_t$$ is_defined_as: $$\begin{aligned}
___ & \bar{g}(\omega_j) = \sqrt{\frac{2}{N_t}}\sum_{k=0}^{N_t}\frac{1}{h_k}_g(t_k)\cos\left(\frac{jk\pi}{N_t}\right), & &j=0,1,\ldots,N_t \label{eq:DCT}\\
_ _& h_k =
\begin{cases}
__ _ 2 & \qquad k=0 \text{ or }_k=N_t \\
_ _1_&_\qquad 1\leq k\leq N_t-1
_ \end{cases} \label{eq:hdef}\end{aligned}$$ The_inverse transform is defined by $$\begin{aligned}
_ & g(t_k) = \sqrt{\frac{2}{N_t}}\sum_{j=0}^{N_t}\frac{1}{h_j}_\bar{g}(\omega_j)\cos\left(\frac{jk\pi}{N_t}\right), & & k=0,1,\ldots,N_t \label{eq:iDCT}\end{aligned}$$ It_can be observed from Eqs. ,_that DCT-1_is its own inverse.
The argument_of the cosine_function in_Eqs. , is equivalent_to a discretized variant of the_continuous argument $\omega_t$ in the continuous transform, where_the_variables are discretized_as_follows:_$$\begin{aligned}
_ &_t_k_= k\frac{T}{N_t}_&_ k=0,1,\ldots,N_t \label{eq:t_samp}\\
_&_\omega_j = j\frac{\pi}{T} & j=0,1,\ldots,N_t \label{eq:w_samp}\end{aligned}$$
In order_to be consistent with_the_continuous formulation, |
by Lemma \[Localy-triv\], is a locally trivial fiber bundle. Since $\mathrm{Path}(P_y, 1_y)$ is contractible, using the long exact sequence in homotopy, we find that $\pi_1(\mathcal{L})\cong \pi_2(P_y)$. Since $\mathcal{G}(A_S)$ is transitive, $P_y$ and $P_x$ are diffeomorphic.
Of course, if $\pi_2(P_x)$ were assumed to be trivial, then condition (H0) would follow automatically. Since $P_x$ is 1-connected, by the Hurewicz theorem, the hypothesis that $H^2(P_x)= 0$ is equivalent to $\pi_2(P_x)$ being finite. We show that this is enough to ensure triviality of the holonomy groups.
\[lemma-Getzler2\] The holonomy group of the foliated manifold $(\mathcal{X}, \mathcal{F})$ is trivial at any point $a\in \mathcal{Y}$.
Let $a\in \mathcal{Y}$ and let $\Gamma$ be the holonomy group at $a$. Let $\mathcal{T}$ be a transversal of $(\mathcal{X}, \mathcal{F})$ through $a$. Since $\Gamma$ is finite, $\mathcal{T}$ can be chosen small enough so that the holonomy transformations $\mathrm{hol}_{u}$ define an action of $\Gamma$ on $\mathcal{T}$. Denote by $\mathcal{T}_{\mathcal{Y}}:= \mathcal{T}\cap \mathcal{Y}$. Since the holonomy of $(\mathcal{Y}, \mathcal{F}_{\mathcal{Y}})$ is trivial, by making $\mathcal{T}$ smaller, we may assume:
1. the action of $\Gamma$ on $\mathcal{T}_{\mathcal{Y}}$ is trivial.
Note also that the submersion $\sigma: \mathcal{T}\to M$ satisfies:
1. $\sigma$ is $\Gamma$-invariant.
2. $\sigma^{-1}(\sigma(a))\subset \mathcal{T}_{\mathcal{Y}}$.
For C2, just note that, for $a'\in \mathcal{T}$, since $a'$ and $\mathrm{hol}_{u}(a')$ are in the same leaf (i.e.cot | by Lemma \[Localy - triv\ ], is a locally trivial fiber bundle. Since $ \mathrm{Path}(P_y, 1_y)$ is contractible, using the retentive accurate sequence in homotopy, we find that $ \pi_1(\mathcal{L})\cong \pi_2(P_y)$. Since $ \mathcal{G}(A_S)$ is transitive, $ P_y$ and $ P_x$ are diffeomorphic.
Of naturally, if $ \pi_2(P_x)$ were assumed to be superficial, then condition (H0) would follow mechanically. Since $ P_x$ is 1 - plug in, by the Hurewicz theorem, the hypothesis that $ H^2(P_x)= 0 $ is equivalent to $ \pi_2(P_x)$ being finite. We show that this is enough to ensure technicality of the holonomy groups.
\[lemma - Getzler2\ ] The holonomy group of the foliated manifold paper $ (\mathcal{X }, \mathcal{F})$ is trivial at any point $ a\in \mathcal{Y}$.
Let $ a\in \mathcal{Y}$ and let $ \Gamma$ be the holonomy group at $ a$. Let $ \mathcal{T}$ be a transversal of $ (\mathcal{X }, \mathcal{F})$ through $ a$. Since $ \Gamma$ is finite, $ \mathcal{T}$ can be chosen small enough so that the holonomy transformations $ \mathrm{hol}_{u}$ specify an action of $ \Gamma$ on $ \mathcal{T}$. Denote by $ \mathcal{T}_{\mathcal{Y}}:= \mathcal{T}\cap \mathcal{Y}$. Since the holonomy of $ (\mathcal{Y }, \mathcal{F}_{\mathcal{Y}})$ is trivial, by making $ \mathcal{T}$ small, we may assume:
1. the action of $ \Gamma$ on $ \mathcal{T}_{\mathcal{Y}}$ is trivial.
Note besides that the submersion $ \sigma: \mathcal{T}\to M$ satisfies:
1. $ \sigma$ is $ \Gamma$-invariant.
2. $ \sigma^{-1}(\sigma(a))\subset \mathcal{T}_{\mathcal{Y}}$.
For C2, precisely note that, for $ a'\in \mathcal{T}$, since $ a'$ and $ \mathrm{hol}_{u}(a')$ are in the same leaf (i.e.cot | by Lemma \[Localy-triv\], is a lucally trivial yuber bnndle. Sjnce $\matfrm{Path}(P_y, 1_y)$ is contractible, nsint the long exact sequence iv homotopj, we find thau $\pi_1(\mathcal{L})\cong \'j_2(P_y)$. Singz $\matggal{G}(A_V)$ is transitive, $P_y$ and $P_x$ are diffeomor[hkc.
Lf course, if $\pi_2(P_x)$ were assumed to bq trivisl, then conditiog (H0) ejuld follow automatically. Since $P_x$ is 1-donnecttd, by the Hurewicz theorem, the hypothesis thwt $H^2(O_x)= 0$ is equivalent ho $\pi_2(P_x)$ beibg fybite. We show that this is enough tk ensure triviality of the holovomy yroups.
\[lemma-Terzlft2\] The holonony grjup of the foliated kanifolc $(\mathcal{X}, \matmcal{F})$ is trivial at any point $a\in \mathcal{Y}$.
Let $a\yn \mathcan{Y}$ and let $\Gamma$ be tye hononoky gfiup at $a$. Lst $\matjcam{T}$ be a trznsversal od $(\mathcal{X}, \mathcal{F})$ trgpugh $a$. Since $\Gamma$ if finite, $\mathcal{T}$ can be chosen small etoufh so that the holonomy transformations $\mathtm{hol}_{u}$ defyne an action of $\Gamma$ on $\mathcal{T}$. Denote by $\mathwal{T}_{\metfcao{Y}}:= \oqtjcal{T}\cap \mathcal{Y}$. Since the holonomy of $(\mathcwm{Y}, \msthcal{F}_{\mathcal{N}})$ is trivial, by malijg $\iathcal{T}$ smaluer, we mag assume:
1. the actiln of $\Gwmma$ in $\mathcaj{T}_{\mayhcal{Y}}$ is trivial.
Note also rhat the subiwrsion $\sigma: \mathccl{T}\to M$ satirfiex:
1. $\sibma$ is $\Gamma$-invariant.
2. $\rigmz^{-1}(\sigma(a))\subsft \mathcam{G}_{\mathcal{Y}}$.
For C2, jjst nmte that, for $a'\in \mathcal{T}$, fince $a'$ aid $\machrm{hol}_{u}(x')$ arg in thq same leav (i.e.cot | by Lemma \[Localy-triv\], is a locally trivial Since 1_y)$ is using the long find $\pi_1(\mathcal{L})\cong \pi_2(P_y)$. Since is transitive, $P_y$ $P_x$ are diffeomorphic. Of course, if were assumed to be trivial, then condition (H0) would follow automatically. Since $P_x$ 1-connected, by the Hurewicz theorem, the hypothesis that $H^2(P_x)= 0$ is equivalent to being We that is enough to ensure triviality of the holonomy groups. \[lemma-Getzler2\] The holonomy group of the foliated $(\mathcal{X}, \mathcal{F})$ is trivial at any point $a\in Let $a\in \mathcal{Y}$ and $\Gamma$ be the holonomy group $a$. $\mathcal{T}$ be transversal $(\mathcal{X}, through $a$. Since is finite, $\mathcal{T}$ can be chosen small enough so that the holonomy transformations $\mathrm{hol}_{u}$ define an action $\Gamma$ on by $\mathcal{T}_{\mathcal{Y}}:= \mathcal{Y}$. the of $(\mathcal{Y}, \mathcal{F}_{\mathcal{Y}})$ by making $\mathcal{T}$ smaller, we may action of $\Gamma$ on $\mathcal{T}_{\mathcal{Y}}$ is trivial. Note that the $\sigma: \mathcal{T}\to M$ satisfies: 1. $\sigma$ $\Gamma$-invariant. 2. $\sigma^{-1}(\sigma(a))\subset \mathcal{T}_{\mathcal{Y}}$. For C2, just note for $a'\in \mathcal{T}$, since $a'$ and $\mathrm{hol}_{u}(a')$ are in the same leaf (i.e.cot | by Lemma \[Localy-triv\], is a localLy trivial fIber bUndLe. SInCe $\maThrm{path}(P_y, 1_y)$ is contrACtibLe, using the long exact seqUence In HOmotOPy, We finD that $\pi_1(\MAtHCAl{L})\CoNg \Pi_2(P_Y)$. SINcE $\mathCal{g}(A_S)$ is trAnsitive, $P_y$ And $p_x$ Are diffeomorPHiC.
Of course, iF $\pi_2(p_x)$ were assumeD to Be trivIaL, thEN condItiOn (H0) woUld folLOw autoMatically. siNCe $P_x$ is 1-COnnecteD, BY tHe HuRewicz theorem, the hYPoTHesis that $H^2(P_x)= 0$ is EquivaLeNT tO $\PI_2(P_x)$ BeiNg finite. We ShOw thaT This is eNOuGH TO enSUre triviality Of the holonoMY grOups.
\[leMmA-GeTZler2\] ThE holoNoMY grOup of the folIateD manifold $(\MathcaL{x}, \mathcaL{f})$ is trivIal at aNy pOinT $a\in \MAtHcAl{Y}$.
leT $A\in \MAtHcaL{y}$ anD let $\GammA$ bE tHe holOnomY GROUp at $A$. LeT $\matHcal{T}$ Be a transversaL of $(\MathCAl{X}, \MathcAl{F})$ thRougH $a$. since $\gamma$ iS finiTe, $\Mathcal{T}$ can be chOsen Small enouGh sO tHat ThE holoNOmy traNsfOrmAtions $\mAthrm{hoL}_{U}$ deFiNE AN aCtion of $\Gamma$ on $\mathCaL{t}$. deNote by $\maThcal{T}_{\MAtHcAL{Y}}:= \mathcaL{T}\Cap \MathCAL{Y}$. SinCe thE HoLonomy of $(\MathcaL{y}, \mAtHcal{F}_{\maThCal{Y}})$ is TrIviAl, bY makiNG $\matHcal{T}$ sMaller, we May asSUme:
1. the action of $\gAmma$ on $\mathcal{t}_{\MaTHCaL{y}}$ is tRivIal.
Note also That THe suBmerSIoN $\siGMa: \matHcal{T}\To m$ SaTIsfies:
1. $\sigma$ is $\Gamma$-iNvAriant.
2. $\Sigma^{-1}(\Sigma(a))\subset \mAthcal{T}_{\matHCAL{Y}}$.
For C2, juSt noTE tHAt, for $a'\in \mathcaL{T}$, sinCe $a'$ and $\mathRM{hol}_{u}(a')$ arE in thE same leaF (i.e.cot | by Lemma \[Localy-triv\], is a loca lly t riv ial f iber bun dle. Since $\m a thrm {Path}(P_y, 1_y)$ is c ontra ct i ble, us ing t he long ex a c t s eq ue nce i n h omoto py, we fin d that $\p i_1 (\ mathcal{L})\ c on g \pi_2(P_ y)$ . Since $\ma thc al{G}( A_ S)$ is tr ans itive , $P_y $ and $ P_x$ aredi f feomor p hic.
O f co urse , if $\pi_2(P_x)$ we r e assumed to b e triv ia l ,t h encon dition (H0 )would followa ut o m a tic a lly. Since $P _x$ is 1-co n nec ted, b ythe Hurewi cz th eo r em, the hypoth esis that $H^ 2(P_x) = 0$ ise quivale nt to$\p i_2 (P_x ) $be ing f i nit e .Wes how that th is i s eno ught o e nsur e t rivi ality of the holon omy gro u ps.
\[l emma- Getz le r2\]The ho lonom ygroup of the fo liat ed manifo ld$( \ma th cal{X } , \mat hca l{F })$ istrivial atan y p oi nt $a\in \mathcal{ Y} $ .
Let $a\i n \mat h ca l{ Y }$ and l et $\ Gamm a $ be t he h o lo nomy gro up at$ a$ .Let $\m at hcal{T }$ be atrans v ersa l of $ (\mathca l{X}, \mathcal{F})$t hrough $a$. S i nc e $\ G amma $ i s finite, $ \mat h cal{ T}$c an be chose n sma ll en o ugh so that the hol on omy tr ansfo rmations $\ma thrm{hol}_ { u } $ define ana ct i on of $\Gamma$ on $ \mathcal{T } $. Denot e by$\mathca l{T}_{\ma t h cal{Y}}: = \ mat hca l{T } \ ca p \mathcal{Y} $ . Sin ce the ho lon omy of$(\ mat hca l{Y }, \mathcal {F}_{\ma th ca l{ Y} })$ is t r ivial, b ymak in g $ \math c al{T}$ smal ler, w em ayassume:
1 . the a ct ionof$\ Gamma $ on $\m athcal{ T}_{\math cal { Y}}$ i strivial .
Note alsoth at the sub me rsi on $\s i g ma: \mat hcal{T}\to M$ satisfies :
1. $ \si gma$is $ \Gamma$-i nva riant.
2 . $\si gma^{- 1}(\s ig ma( a ) )\sub s e t\ma th cal{T}_{\m a t hca l{Y}} $.
Fo r C2, j ust note that, for $a' \in \mathcal{ T}$ , si n c e$a' $ a n d $ \m a thr m { hol}_{u}(a')$ a re in thesa m eleaf (i.e. c ot | by_Lemma \[Localy-triv\],_is a locally trivial_fiber bundle._Since_$\mathrm{Path}(P_y, 1_y)$_is_contractible, using the_long exact sequence_in homotopy, we find_that $\pi_1(\mathcal{L})\cong \pi_2(P_y)$._Since_$\mathcal{G}(A_S)$ is transitive, $P_y$ and $P_x$ are diffeomorphic.
Of course, if $\pi_2(P_x)$ were assumed to_be_trivial, then_condition_(H0)_would follow automatically. Since $P_x$_is 1-connected, by the Hurewicz_theorem, the_hypothesis that $H^2(P_x)= 0$ is equivalent to $\pi_2(P_x)$_being_finite. We show_that this is enough to ensure triviality of the_holonomy groups.
\[lemma-Getzler2\] The holonomy group of_the foliated manifold_$(\mathcal{X},_\mathcal{F})$_is trivial at any_point $a\in \mathcal{Y}$.
Let $a\in \mathcal{Y}$ and_let $\Gamma$ be the holonomy group_at $a$. Let $\mathcal{T}$ be a transversal_of $(\mathcal{X}, \mathcal{F})$ through $a$. Since_$\Gamma$ is finite, $\mathcal{T}$ can_be chosen_small enough so that the_holonomy transformations $\mathrm{hol}_{u}$_define an_action of $\Gamma$_on $\mathcal{T}$. Denote by $\mathcal{T}_{\mathcal{Y}}:= \mathcal{T}\cap_\mathcal{Y}$. Since the_holonomy of $(\mathcal{Y}, \mathcal{F}_{\mathcal{Y}})$ is trivial,_by_making $\mathcal{T}$ smaller,_we_may_assume:
1. _the action of_$\Gamma$_on $\mathcal{T}_{\mathcal{Y}}$_is_trivial.
Note also that the submersion $\sigma:_\mathcal{T}\to_M$ satisfies:
1. $\sigma$ is $\Gamma$-invariant.
2. _$\sigma^{-1}(\sigma(a))\subset \mathcal{T}_{\mathcal{Y}}$.
For C2, just_note_that, for $a'\in \mathcal{T}$,_since $a'$ and $\mathrm{hol}_{u}(a')$ are_in the same leaf (i.e.cot |
Subsets and Splits