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
|
|---|---|---|---|---|---|---|
1) +H(U_2)-H(U_2|S_2) -H(U_1, U_2) \notag \\
& \qquad \quad + H(U_1, U_2|Z) - H(U_2) + H(U_2|U_1) + H(U_1) - H(U_1) \\
& =H(U_1, U_2|Z)- H(U_1|S_1) -H(U_2|S_2) \\
& =H(U_1, U_2|Z)- H(U_1|S_1, S_2) -H(U_2|S_2, S_1, U_1) \label{midtrm}\\
& =H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1) \\
& =H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1, Z) \label{lasttrm} \\
& =I(U_1, U_2;S_2, S_1| Z) \label{verylasttrm}\end{aligned}$$ is from $U_1-S_1-S_2$ and $U_2-S_2-S_1U_1$, is from $U_1U_2-S_1S_2-Z$.
Combining and, we conclude that yields, i.e., $$I(U_1, U_2;S_2, S_1| Z) < I(U_1, U_2; Y|Z).$$
Proof of Theorem \[lemma2\] {#appendixA}
===========================
Achievability
-------------
The source coding part is based on lossy source coding at the two encoders conditioned on the side information $Z$ shared between the encoder and decoder [@gray1972conditional], after which the conditional rate distortion functions given in can be achieved for $S_1$ and $S_2$, respectively. The channel coding part is performed based on classical multiple access channel coding with independent channel inputs [@cover2012elements].
Converse | 1) + H(U_2)-H(U_2|S_2) -H(U_1, U_2) \notag \\
& \qquad \quad + H(U_1, U_2|Z) - H(U_2) + H(U_2|U_1) + H(U_1) - H(U_1) \\
& = H(U_1, U_2|Z)- H(U_1|S_1) -H(U_2|S_2) \\
& = H(U_1, U_2|Z)- H(U_1|S_1, S_2) -H(U_2|S_2, S_1, U_1) \label{midtrm}\\
& = H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1) \\
& = H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1, Z) \label{lasttrm } \\
& = I(U_1, U_2;S_2, S_1| Z) \label{verylasttrm}\end{aligned}$$ is from $ U_1 - S_1 - S_2 $ and $ U_2 - S_2 - S_1U_1 $, is from $ U_1U_2 - S_1S_2 - Z$.
Combining and, we conclude that yields, i.e., $ $ I(U_1, U_2;S_2, S_1| Z) < I(U_1, U_2; Y|Z).$$
Proof of Theorem \[lemma2\ ] { # appendixA }
= = = = = = = = = = = = = = = = = = = = = = = = = = =
Achievability
-------------
The source gull character is based on lossy source coding at the two encoders discipline on the side information $ Z$ share between the encoder and decoder [ @gray1972conditional ], after which the conditional pace distortion functions given in can be achieved for $ S_1 $ and $ S_2 $, respectively. The channel tease part is performed base on authoritative multiple access channel gull with independent channel input signal [ @cover2012elements ].
Converse | 1) +H(U_2)-J(U_2|S_2) -H(U_1, U_2) \notag \\
& \qquad \quxd + H(U_1, U_2|Z) - H(U_2) + H(U_2|U_1) + H(U_1) - H(U_1) \\
& =H(U_1, U_2|Z)- H(U_1|S_1) -H(U_2|S_2) \\
& =H(U_1, U_2|Z)- H(U_1|S_1, S_2) -H(U_2|S_2, S_1, U_1) \lavel{midtrm}\\
& =H(U_1, U_2|Z) -H(U_1, U_2|S_2, R_1) \\
& =H(U_1, U_2|Z) -J(U_1, U_2|S_2, S_1, Z) \lauel{lasttrm} \\
& =I(U_1, U_2;S_2, S_1| Z) \lanzl{vergpastcrn}\end{aligned}$$ is from $U_1-S_1-S_2$ dnd $U_2-S_2-S_1U_1$, is frmm $U_1B_2-S_1S_2-Z$.
Combining and, we conclude that yyelds, i.r., $$L(U_1, U_2;S_2, S_1| Z) < I(U_1, O_2; Y|Z).$$
Pgojf or Theorem \[lemma2\] {#appendixA}
===========================
Achievabiljty
-------------
The vource coding part is based on lossy sokrce coding at the two encoders cindieuoned on the side infogkation $Z$ sgared between the encoder and ddcodex [@gray1972condijninap], after whici the bonditional rate distmrtion gunctions givek in ran ve achieved for $S_1$ and $S_2$, respectively. Thg channel wobing part is performee vased on wlasrucau mhlvipme accfss channel ckding with undependent channel igiits [@cover2012elejents].
Cjnderse | 1) +H(U_2)-H(U_2|S_2) -H(U_1, U_2) \notag \\ & + U_2|Z) - + H(U_2|U_1) + =H(U_1, H(U_1|S_1) -H(U_2|S_2) \\ =H(U_1, U_2|Z)- H(U_1|S_1, -H(U_2|S_2, S_1, U_1) \label{midtrm}\\ & =H(U_1, -H(U_1, U_2|S_2, S_1) \\ & =H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1, Z) \label{lasttrm} \\ =I(U_1, U_2;S_2, S_1| Z) \label{verylasttrm}\end{aligned}$$ is from $U_1-S_1-S_2$ and $U_2-S_2-S_1U_1$, is from $U_1U_2-S_1S_2-Z$. and, conclude yields, $$I(U_1, U_2;S_2, S_1| Z) < I(U_1, U_2; Y|Z).$$ Proof of Theorem \[lemma2\] {#appendixA} =========================== Achievability ------------- source coding part is based on lossy source at the two encoders on the side information $Z$ between encoder and [@gray1972conditional], which conditional rate distortion given in can be achieved for $S_1$ and $S_2$, respectively. The channel coding part is performed based classical multiple coding with channel [@cover2012elements]. | 1) +H(U_2)-H(U_2|S_2) -H(U_1, U_2) \notag \\
& \qquad \quad + H(U_1, U_2|z) - H(U_2) + H(U_2|U_1) + H(U_1) - H(U_1) \\
& =h(U_1, U_2|Z)- H(u_1|S_1) -H(u_2|S_2) \\
& =H(u_1, U_2|z)- H(U_1|S_1, s_2) -H(U_2|S_2, s_1, U_1) \label{midtrm}\\
& =H(u_1, u_2|Z) -H(U_1, u_2|S_2, S_1) \\
& =H(U_1, U_2|Z) -H(U_1, U_2|S_2, S_1, Z) \label{lastTrm} \\
& =I(U_1, u_2;S_2, s_1| z) \labEL{vErylaSttrm}\enD{AlIGNed}$$ Is FrOm $U_1-s_1-S_2$ ANd $u_2-S_2-S_1U_1$, iS frOm $U_1U_2-S_1S_2-Z$.
combining aNd, wE cOnclude that yIElDs, i.e., $$I(U_1, U_2;S_2, S_1| Z) < i(U_1, U_2; y|Z).$$
Proof of TheOreM \[lemma2\] {#ApPenDIxA}
===========================
AcHieVabilIty
-------------
The SOurce cOding part Is BAsed on LOssy souRCE cOdinG at the two encoders COnDItioned on the siDe infoRmATiON $z$ shAreD between thE eNcodeR And decoDEr [@GRAY1972coNDitional], after Which the conDItiOnal raTe DisTOrtion FunctIoNS giVen in can be aChieVed for $S_1$ anD $S_2$, respECtively. tHe channEl codiNg pArt Is peRFoRmEd bAsED on CLaSsiCAl mUltiple aCcEsS chanNel cODING witH inDepeNdent Channel inputs [@CovEr2012elEMenTs].
ConVerse | 1) +H(U_2)-H(U_2|S_2) -H(U _1, U_2) \ notag \\
&\q quad \qu ad + H(U_1, U _ 2|Z) - H(U_2) + H(U_2|U _1) +H(U_ 1 )- H(U _1) \\& = H ( U_1 ,U_ 2|Z )- H( U_1|S _1) -H(U_2 |S_2) \\
& =H (U _1, U_2|Z)-H (U _1|S_1, S_ 2)-H(U_2|S_2,S_1 , U_1) \ lab e l{mid trm }\\
& =H(U_ 1 , U_2| Z) -H(U_1 ,U _2|S_2 , S_1) \ \ &=H(U _1, U_2|Z) -H(U_1 , U _ 2|S_2, S_1, Z) \labe l{ l as t t rm} \\
& =I(U_1, U _2;S_ 2 , S_1|Z )\ l a bel { verylasttrm}\ end{aligned } $$is fro m$U_ 1 -S_1-S _2$ a nd $U_ 2-S_2-S_1U_ 1$,is from $ U_1U_2 - S_1S_2- Z $.
Com bining an d,we c o nc lu deth a t y i el ds, i.e ., $$I(U _1 ,U_2;S _2,S _ 1 | Z)< I (U_1 , U_2 ; Y|Z).$$
Pr oof ofT heo rem \ [lemm a2\] { #appe ndixA}
==== == =============== ==== ==
Achie vab il ity
- ----- - ------
T hesourcecodingp art i s b as ed on lossy source c o d in g at the two e n co de r s condit io ned ont h e sid e in f or mation $ Z$ sha r ed b etweenth e enco de r a nddecod e r [@ gray19 72condit ional ] , after whicht he conditiona l r a t ed isto rti on function s gi v en i n ca n b e a c hieve d for $ S _1 $ and $S_2$, respect iv ely. T he ch annel codingpart is pe r f o rmed bas ed o n c l assical multip le ac cess chann e l coding with indepen dent chan n e l inputs [@ cov er2 012 e l em ents].
Conve r s e | 1) +H(U_2)-H(U_2|S_2)_-H(U_1, U_2)_\notag \\
& \qquad \quad_+ H(U_1,__U_2|Z) _-_H(U_2) +_ H(U_2|U_1) _+ H(U_1) -_H(U_1) \\
& =H(U_1,_U_2|Z)-_H(U_1|S_1) -H(U_2|S_2) \\
& =H(U_1, U_2|Z)- H(U_1|S_1, S_2) -H(U_2|S_2, S_1, U_1) \label{midtrm}\\
& =H(U_1, U_2|Z) -H(U_1,_U_2|S_2,_S_1) \\
&_=H(U_1,_U_2|Z)_-H(U_1, U_2|S_2, S_1, Z) \label{lasttrm}_\\
& =I(U_1, U_2;S_2, S_1| Z)_\label{verylasttrm}\end{aligned}$$ is_from $U_1-S_1-S_2$ and $U_2-S_2-S_1U_1$, is from $U_1U_2-S_1S_2-Z$.
Combining and,_we_conclude that yields,_i.e., $$I(U_1, U_2;S_2, S_1| Z) < I(U_1, U_2; Y|Z).$$
Proof_of Theorem \[lemma2\] {#appendixA}
===========================
Achievability
-------------
The source coding part_is based on_lossy_source_coding at the two_encoders conditioned on the side information_$Z$ shared between the encoder and_decoder [@gray1972conditional], after which the conditional rate_distortion functions given in can be_achieved for $S_1$ and $S_2$,_respectively. The_channel coding part is performed_based on classical_multiple access_channel coding with_independent channel inputs [@cover2012elements].
Converse |
(2n-2p+1)! p! }\right)^2
\label{eq.A2j+1}$$ For small oscillation amplitudes, $\theta\ll 1$, one can restrict the second finite sum to the power $p=0$. The $A_j$ coefficients are then temperature independent and one has:
$$A_{0}(\delta,-\delta)\approx
\left(\sum_{n=0}^{\infty}\frac{\left(-\tilde \delta^2\right)^{n}}{n! }
\right)^2 = e^{-2\tilde \delta^2}=e^{- \frac{4\delta^2}{w_0^2}},$$
$$A_{1}(\delta,-\delta)\approx
\tilde \delta^2 \left(\sum_{n=0}^{\infty} 2 \frac{\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2 =4\tilde\delta^2 e^{-2\tilde\delta^2},$$
$$A_{2}(\delta,-\delta)\approx
\frac{1}{2} \left(\sum_{n=0}^{\infty} 2 \frac{(2n+2 -1)\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2 =\frac{1}{2} \left( 4 \frac{d}{d(-\tilde\delta^2)} {(-\tilde\delta^2 e^{-\tilde\delta^2})} -2 {e^{-\tilde\delta^2}} \right)^2 =2(1-2\tilde\delta^2)^2 e^{-2\tilde\delta^2}.$$ From that it is clear that $A_2$ is suppressed for $\tilde\delta=\tilde\delta^\star\equiv 1/\sqrt{2}$, i.e. $\delta=w_0/2$. One then obtains: $$A_{0}(\delta^\star,-\delta^\star)\approx1/e\,\,, \, A_{1}(\delta^\star,-\delta^\star)\approx2/e$$ For higher spatial spreading $\theta_x$, the previous simplification is not possible. The dependency of the $A_j$ coefficient in $\delta$ and $\theta_x$ is shown in \[FigS1\]. It is still possible to cancel the $A_2$ term by choosing the proper detector displacement | (2n-2p+1)! p! } \right)^2
\label{eq. A2j+1}$$ For small oscillation amplitudes, $ \theta\ll 1 $, one can restrict the second finite kernel to the exponent $ p=0$. The $ A_j$ coefficients are then temperature independent and one has:
$ $ A_{0}(\delta,-\delta)\approx
\left(\sum_{n=0}^{\infty}\frac{\left(-\tilde \delta^2\right)^{n}}{n! }
\right)^2 = e^{-2\tilde \delta^2}=e^{- \frac{4\delta^2}{w_0 ^ 2}},$$
$ $ A_{1}(\delta,-\delta)\approx
\tilde \delta^2 \left(\sum_{n=0}^{\infty } 2 \frac{\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2 = 4\tilde\delta^2 e^{-2\tilde\delta^2},$$
$ $ A_{2}(\delta,-\delta)\approx
\frac{1}{2 } \left(\sum_{n=0}^{\infty } 2 \frac{(2n+2 -1)\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2 = \frac{1}{2 } \left (4 \frac{d}{d(-\tilde\delta^2) } { (-\tilde\delta^2 e^{-\tilde\delta^2 }) } -2 { e^{-\tilde\delta^2 } } \right)^2 = 2(1 - 2\tilde\delta^2)^2 e^{-2\tilde\delta^2}.$$ From that it is clear that $ A_2 $ is oppress for $ \tilde\delta=\tilde\delta^\star\equiv 1/\sqrt{2}$, i.e. $ \delta = w_0/2$. One then obtains: $ $ A_{0}(\delta^\star,-\delta^\star)\approx1 / e\,\, , \, A_{1}(\delta^\star,-\delta^\star)\approx2 / e$$ For higher spatial circulate $ \theta_x$, the previous simplification is not potential. The colony of the $ A_j$ coefficient in $ \delta$ and $ \theta_x$ is shown in \[FigS1\ ]. It is even possible to cancel the $ A_2 $ terminus by choose the proper detector displacement |
(2n-2p+1)! o! }\right)^2
\label{eq.A2j+1}$$ For smaul oscillation cnplituves, $\thefa\ll 1$, ond can restrict the second fiiite sum uj the power $p=0$. The $X_j$ coeffibients arw thtn temperature invspendenb and lne kaw:
$$A_{0}(\delta,-\delta)\apkrox
\left(\sum_{t=0}^{\infty}\frac{\left(-\diudz \delta^2\right)^{n}}{n! }
\right)^2 = e^{-2\tilde \delta^2}=e^{- \srac{4\delya^2}{a_0^2}},$$
$$A_{1}(\delta,-\delta)\apkrox
\tplqe \dslta^2 \left(\sum_{n=0}^{\infty} 2 \frac{\left(-\tilde\dslta^2\rigit)^{n}}{n! }
\right)^2 =4\tildr\delta^2 e^{-2\tilde\delta^2},$$
$$A_{2}(\delta,-\depta)\aoprox
\frac{1}{2} \left(\sum_{n=0}^{\lnfty} 2 \frac{(2b+2 -1)\lesr(-\tilde\delta^2\rkght)^{n}}{n! }
\right)^2 =\frac{1}{2} \left( 4 \frac{d}{d(-\tilde\delta^2)} {(-\tilde\delta^2 e^{-\gilde\belta^2})} -2 {e^{-\tilee\eelhd^2}} \right)^2 =2(1-2\tilve\deltw^2)^2 e^{-2\tilde\delta^2}.$$ From tvat it os clear that $S_2$ iv syppressed for $\tilde\denta=\tilde\delta^\star\ezuiv 1/\sqrt{2}$, i.z. $\delta=w_0/2$. One then obtqibs: $$A_{0}(\dglta^\sdar,-\ddota^\rtad)\a'prkx1/e\,\,, \, W_{1}(\demta^\star,-\delfa^\star)\approz2/e$$ For higher spatisl wpreading $\thefa_x$, thq [revious simplification is not possible. Ths dependency of the $A_j$ xoefficient in $\delta$ wnd $\theta_v$ is shown in \[FigS1\]. It is still possible to cancel the $E_2$ gern ny cfioding the proper detector displacement | (2n-2p+1)! p! }\right)^2 \label{eq.A2j+1}$$ For small oscillation 1$, can restrict second finite sum $A_j$ are then temperature and one has: \left(\sum_{n=0}^{\infty}\frac{\left(-\tilde \delta^2\right)^{n}}{n! } \right)^2 = e^{-2\tilde \frac{4\delta^2}{w_0^2}},$$ $$A_{1}(\delta,-\delta)\approx \tilde \delta^2 \left(\sum_{n=0}^{\infty} 2 \frac{\left(-\tilde\delta^2\right)^{n}}{n! } \right)^2 =4\tilde\delta^2 e^{-2\tilde\delta^2},$$ $$A_{2}(\delta,-\delta)\approx \frac{1}{2} 2 \frac{(2n+2 -1)\left(-\tilde\delta^2\right)^{n}}{n! } \right)^2 =\frac{1}{2} \left( 4 \frac{d}{d(-\tilde\delta^2)} {(-\tilde\delta^2 e^{-\tilde\delta^2})} -2 {e^{-\tilde\delta^2}} =2(1-2\tilde\delta^2)^2 From it clear that $A_2$ is suppressed for $\tilde\delta=\tilde\delta^\star\equiv 1/\sqrt{2}$, i.e. $\delta=w_0/2$. One then obtains: $$A_{0}(\delta^\star,-\delta^\star)\approx1/e\,\,, \, A_{1}(\delta^\star,-\delta^\star)\approx2/e$$ higher spatial spreading $\theta_x$, the previous simplification is possible. The dependency of $A_j$ coefficient in $\delta$ and is in \[FigS1\]. is possible cancel the $A_2$ by choosing the proper detector displacement |
(2n-2p+1)! p! }\right)^2
\label{eq.A2j+1}$$ For small OscillatioN amplItuDes, $\ThEta\lL 1$, one Can restrict the SEconD finite sum to the power $p=0$. THe $A_j$ cOeFFiciENtS are tHen tempERaTURe iNdEpEndEnT AnD one hAs:
$$A_{0}(\Delta,-\deLta)\approx
\lEft(\SuM_{n=0}^{\infty}\frac{\lEFt(-\Tilde \delta^2\RigHt)^{n}}{n! }
\right)^2 = e^{-2\tiLde \Delta^2}=e^{- \FrAc{4\dELta^2}{w_0^2}},$$
$$A_{1}(\DelTa,-\delTa)\apprOX
\tilde \Delta^2 \left(\SuM_{N=0}^{\infty} 2 \FRac{\left(-\TILdE\delTa^2\right)^{n}}{n! }
\right)^2 =4\tilDE\dELta^2 e^{-2\tilde\delta^2},$$
$$a_{2}(\delta,-\DeLTa)\APProX
\frAc{1}{2} \left(\sum_{n=0}^{\InFty} 2 \frAC{(2n+2 -1)\left(-\tILdE\DELta^2\RIght)^{n}}{n! }
\right)^2 =\frAc{1}{2} \left( 4 \frac{d}{D(-\TilDe\deltA^2)} {(-\tIldE\Delta^2 e^{-\Tilde\DeLTa^2})} -2 {e^{-\Tilde\delta^2}} \rIght)^2 =2(1-2\Tilde\deltA^2)^2 e^{-2\tildE\Delta^2}.$$ FrOM that it Is cleaR thAt $A_2$ Is suPPrEsSed FoR $\TilDE\dEltA=\TilDe\delta^\sTaR\eQuiv 1/\sQrt{2}$, i.E. $\DELTa=w_0/2$. ONe tHen oBtainS: $$A_{0}(\delta^\star,-\deLta^\Star)\APprOx1/e\,\,, \, A_{1}(\dElta^\sTar,-\dElTa^\staR)\approX2/e$$ For HiGher spatial spreAdinG $\theta_x$, thE prEvIouS sImpliFIcatioN is Not PossiblE. The depENdeNcY OF ThE $A_j$ coefficient in $\deLtA$ ANd $\Theta_x$ is Shown iN \[fiGS1\]. iT is still PoSsiBle tO CAncel The $A_2$ TErM by choosIng the PRoPeR detectOr DisplaCeMenT |
(2n-2p+1)! p! }\right)^2\label{eq. A2j+1 }$$ Fo rsmal l os cillation ampl i tude s, $\theta\ll 1$, onecan r es t rict th e sec ond fin i te s umto t hepo w er $p=0 $.The $A_ j$ coeffic ien ts are then te m pe rature ind epe ndent and on e h as:
$ $A _{0 } (\del ta, -\del ta)\ap p rox
\ left(\sum _{ n =0}^{\ i nfty}\f r a c{ \lef t(-\tilde \delta^ 2 \r i ght)^{n}}{n! }
\righ t) ^ 2= e^{ -2\ tilde \del ta ^2}=e ^ {- \fra c {4 \ d e lta ^ 2}{w_0^2}},$$
$$A_{1}(\ d elt a,-\de lt a)\ a pprox\tild e\ del ta^2 \left( \sum _{n=0}^{\ infty} 2 \frac { \left(- \tilde \de lta ^2\r i gh t) ^{n }} { n!}
\ rig h t)^ 2 =4\til de \d elta^ 2 e^ { - 2 \ tild e\d elta ^2},$ $
$$A_{2}(\d elt a,-\ d elt a)\ap prox\fra c{ 1}{2} \left (\sum _{ n=0}^{\infty} 2 \fr ac{(2n+2-1) \l eft (- \tild e \delta ^2\ rig ht)^{n} }{n! }\ rig ht ) ^ 2 = \frac{1}{2} \left( 4 \ fr ac{d}{d( -\tild e \d el t a^2)} {( -\ til de\d e l ta^2e^{- \ ti lde\delt a^2})} -2 { e^{-\ti ld e\delt a^ 2}} \r ight) ^ 2 =2 (1-2\t ilde\del ta^2) ^ 2 e^{-2\tilde\ d elta^2}.$$ Fr o mt h at it i s c lear that $ A_2$ is s uppr e ss edf or $\ tilde \d e lt a =\tilde\delta^\star \e quiv 1 /\sqr t{2}$, i.e. $ \delta=w_0 / 2 $ . One th en o b ta i ns: $$A_{0}(\d elta^ \star,-\de l ta^\star )\app rox1/e\, \,, \, A _ { 1}(\delt a^\ sta r,- \de l t a^ \star)\approx 2 / e$$Fo r highe r s patialspr ead ing $\ th eta_x$, t he previ ou ssi mp lif icati o n is not p oss ib le. Thed epende ncy o f th e$A _ j$coeffic i en t in $ \d el ta$and $ \thet a_x$ isshown i n \[FigS1 \]. It i sst ill pos sible to canc el the $A_2$ t erm by ch o o sing the proper detector displa c ement |
(2n-2p+1)! p!_}\right)^2
\label{eq.A2j+1}$$ For_small oscillation amplitudes, $\theta\ll_1$, one_can_restrict the_second_finite sum to_the power $p=0$._The $A_j$ coefficients are_then temperature independent_and_one has:
$$A_{0}(\delta,-\delta)\approx
\left(\sum_{n=0}^{\infty}\frac{\left(-\tilde \delta^2\right)^{n}}{n! }
\right)^2 = e^{-2\tilde \delta^2}=e^{- \frac{4\delta^2}{w_0^2}},$$
$$A_{1}(\delta,-\delta)\approx
\tilde \delta^2 \left(\sum_{n=0}^{\infty} 2 \frac{\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2_=4\tilde\delta^2_e^{-2\tilde\delta^2},$$
$$A_{2}(\delta,-\delta)\approx
\frac{1}{2} \left(\sum_{n=0}^{\infty}_2_\frac{(2n+2_-1)\left(-\tilde\delta^2\right)^{n}}{n! }
\right)^2 =\frac{1}{2} \left( 4_\frac{d}{d(-\tilde\delta^2)} {(-\tilde\delta^2 e^{-\tilde\delta^2})} -2 {e^{-\tilde\delta^2}}_\right)^2 =2(1-2\tilde\delta^2)^2_e^{-2\tilde\delta^2}.$$ From that it is clear that $A_2$_is_suppressed for $\tilde\delta=\tilde\delta^\star\equiv_1/\sqrt{2}$, i.e. $\delta=w_0/2$. One then obtains: $$A_{0}(\delta^\star,-\delta^\star)\approx1/e\,\,, \, _A_{1}(\delta^\star,-\delta^\star)\approx2/e$$ For higher spatial spreading $\theta_x$,_the previous simplification_is_not_possible. The dependency of_the $A_j$ coefficient in $\delta$ and_$\theta_x$ is shown in \[FigS1\]. It_is still possible to cancel the $A_2$_term by choosing the proper detector_displacement |
2+1} \{C_{E_2} C_{E_1} - \prod_{k=1}^{2} C_{e_{k,2}} C_{e_{k,1}} \}\,, \mbox{ when $r=2$.} \nonumber \end{aligned}$$
\[Conway3braugmented\]
Eq. \[Alexander3brApprox\] is essentially given by Prop. 4.2, p. 13-14 [@69].
Induction, plus (\[Conwayrecur\]) and (\[Conwaysumrel\]) for the first and last equations. Eq. \[Conwaydiffofprodlast\] is also helpful to prove the order of magnitude result in (\[Conway3brApprox\]).
When $\gamma \in B_3$ and $a >0$ we have: $$\begin{aligned}
P_{\widehat {\alpha_2^3 \gamma} } &=& v^6 P_{\widehat {\gamma} } + P_{T_{w(\gamma)+5}} - v^6 P_{T_{w(\gamma)+1}} \,, \\
%
P_{\widehat {\alpha_2^{3a} \gamma} } &=&
v^{6a} P_{\widehat {\gamma} }+ \sum_{j=1}^a v^{6a-6j} P_{T_{w(\gamma)+6j-1}} - \sum_{j=1}^a v^{6+6a-6j} P_{T_{w(\gamma)+6j-5}}\,, \\
%
P_{\widehat {\alpha_2^{-3a} \gamma} } &=&
v^{-6a} P_{\widehat {\gamma} }+ \sum_{j=1}^a v^{6j-6a} P_{T_{w(\gamma)-6j+1}} - \sum_{j=1}^a v^{6j-6a-6} P_{T_{w(\gamma)-6j+5}}\,.\end{aligned}$$ \[Homflypt3braugmented\]
Induction, plus Prop. \[Conway3braugmented\] and Prop. \[H3braidstdformprop\] for the first equation.
Proof of Prop. \[HConway3br\] (When do 3 braid links have $P_{\widehat \beta} = P_{\ | 2 + 1 } \{C_{E_2 } C_{E_1 } - \prod_{k=1}^{2 } C_{e_{k,2 } } C_{e_{k,1 } } \}\, , \mbox { when $ r=2$. } \nonumber \end{aligned}$$
\[Conway3braugmented\ ]
Eq. \[Alexander3brApprox\ ] is essentially given by Prop. 4.2, p. 13 - 14 [ @69 ].
Induction, plus (\[Conwayrecur\ ]) and (\[Conwaysumrel\ ]) for the first and last equation. Eq. \[Conwaydiffofprodlast\ ] is besides helpful to prove the order of order of magnitude result in (\[Conway3brApprox\ ]).
When $ \gamma \in B_3 $ and $ a > 0 $ we receive: $ $ \begin{aligned }
P_{\widehat { \alpha_2 ^ 3 \gamma } } & = & v^6 P_{\widehat { \gamma } } + P_{T_{w(\gamma)+5 } } - v^6 P_{T_{w(\gamma)+1 } } \, , \\
%
P_{\widehat { \alpha_2^{3a } \gamma } } & = &
v^{6a } P_{\widehat { \gamma } } + \sum_{j=1}^a v^{6a-6j } P_{T_{w(\gamma)+6j-1 } } - \sum_{j=1}^a v^{6 + 6a-6j } P_{T_{w(\gamma)+6j-5}}\, , \\
%
P_{\widehat { \alpha_2^{-3a } \gamma } } & = &
v^{-6a } P_{\widehat { \gamma } } + \sum_{j=1}^a v^{6j-6a } P_{T_{w(\gamma)-6j+1 } } - \sum_{j=1}^a v^{6j-6a-6 } P_{T_{w(\gamma)-6j+5}}\,.\end{aligned}$$ \[Homflypt3braugmented\ ]
Induction, plus Prop. \[Conway3braugmented\ ] and Prop. \[H3braidstdformprop\ ] for the first equation.
Proof of Prop. \[HConway3br\ ] (When do 3 braid link have $ P_{\widehat \beta } = P_{\ | 2+1} \{C_{E_2} C_{E_1} - \prod_{k=1}^{2} C_{e_{k,2}} C_{e_{k,1}} \}\,, \mbox{ when $r=2$.} \nonumbet \wnd{alijned}$$
\[Consay3braugoented\]
Eq. \[Alexander3brApprox\] is ewsentually given by Prop. 4.2, p. 13-14 [@69].
Knduction, plus (\[Cobwaycecur\]) and (\[Conwaysumrel\]) for the nirst end last equatipns. Eq. \[Conwdydiffofprodlavt\] id also helpful to prove the order os magniyufe result in (\[Cjnwaj3bwAppdox\]).
When $\gamma \in B_3$ and $a >0$ we have: $$\gegin{alpgned}
P_{\widehat {\alpna_2^3 \gamma} } &=& v^6 P_{\widehat {\gammw} } + O_{T_{w(\gamma)+5}} - v^6 P_{T_{w(\gamla)+1}} \,, \\
%
P_{\widehaj {\al[ya_2^{3a} \gamma} } &=&
x^{6a} P_{\widehau {\yamma} }+ \sum_{j=1}^z v^{6a-6j} P_{T_{w(\gamma)+6j-1}} - \sum_{j=1}^a v^{6+6a-6j} P_{T_{w(\gxmma)+6j-5}}\,, \\
%
P_{\widehat {\qlpha_2^{-3w} \gamma} } &=&
v^{-6a} P_{\widvhat {\gamma} }+ \smk_{j=1}^a v^{6j-6d} P_{T_{w(\gakma)-6j+1}} - \sum_{j=1}^a v^{6j-6s-6} P_{V_{w(\ganma)-6j+5}}\,.\end{aligned}$$ \[Homfly't3braugmented\]
Inductijn, plus Pso'. \[Conway3braugmented\] ane Prop. \[H3traigstdwirmorok\] fpr the flrsv equation.
Pdoof of Prop. \[HConway3br\] (When do 3 bwqid links havs $P_{\widqhwt \beta} = P_{\ | 2+1} \{C_{E_2} C_{E_1} - \prod_{k=1}^{2} C_{e_{k,2}} C_{e_{k,1}} when \nonumber \end{aligned}$$ Eq. \[Alexander3brApprox\] is p. [@69]. Induction, plus and (\[Conwaysumrel\]) for first and last equations. Eq. \[Conwaydiffofprodlast\] also helpful to prove the order of magnitude result in (\[Conway3brApprox\]). When $\gamma B_3$ and $a >0$ we have: $$\begin{aligned} P_{\widehat {\alpha_2^3 \gamma} } &=& v^6 {\gamma} + - P_{T_{w(\gamma)+1}} \,, \\ % P_{\widehat {\alpha_2^{3a} \gamma} } &=& v^{6a} P_{\widehat {\gamma} }+ \sum_{j=1}^a v^{6a-6j} P_{T_{w(\gamma)+6j-1}} \sum_{j=1}^a v^{6+6a-6j} P_{T_{w(\gamma)+6j-5}}\,, \\ % P_{\widehat {\alpha_2^{-3a} \gamma} &=& v^{-6a} P_{\widehat {\gamma} \sum_{j=1}^a v^{6j-6a} P_{T_{w(\gamma)-6j+1}} - \sum_{j=1}^a P_{T_{w(\gamma)-6j+5}}\,.\end{aligned}$$ Induction, plus \[Conway3braugmented\] Prop. for the first Proof of Prop. \[HConway3br\] (When do 3 braid links have $P_{\widehat \beta} = P_{\ | 2+1} \{C_{E_2} C_{E_1} - \prod_{k=1}^{2} C_{e_{k,2}} C_{e_{k,1}} \}\,, \mbox{ when $r=2$.} \nOnumber \end{AlignEd}$$
\[COnwAy3BrauGmenTed\]
Eq. \[Alexander3BRAppRox\] is essentially given bY Prop. 4.2, P. 13-14 [@69].
INDuctIOn, Plus (\[COnwayreCUr\]) AND (\[CoNwAySumReL\]) FoR the fIrsT and lasT equations. eq. \[COnWaydiffofproDLaSt\] is also heLpfUl to prove the OrdEr of maGnItuDE resuLt iN (\[ConwAy3brApPRox\]).
WheN $\gamma \in B_3$ AnD $A >0$ we havE: $$\Begin{alIGNeD}
P_{\wiDehat {\alpha_2^3 \gamma} } &=& v^6 P_{\WIdEHat {\gamma} } + P_{T_{w(\gamMa)+5}} - v^6 P_{T_{w(\GaMMa)+1}} \,, \\
%
p_{\WIdeHat {\Alpha_2^{3a} \gammA} } &=&
v^{6A} P_{\widEHat {\gammA} }+ \SuM_{J=1}^A V^{6a-6j} p_{t_{w(\gamma)+6j-1}} - \sum_{j=1}^a V^{6+6a-6j} P_{T_{w(\gamma)+6J-5}}\,, \\
%
p_{\wiDehat {\aLpHa_2^{-3a} \GAmma} } &=&
v^{-6a} p_{\wideHaT {\GamMa} }+ \sum_{j=1}^a v^{6j-6a} P_{t_{w(\gaMma)-6j+1}} - \sum_{j=1}^a V^{6j-6a-6} P_{T_{w(\GAmma)-6j+5}}\,.\enD{Aligned}$$ \[homflyPt3bRauGmenTEd\]
inDucTiON, plUS PRop. \[cOnwAy3braugmEnTeD\] and PRop. \[H3BRAIDstdForMproP\] for tHe first equatiOn.
PRoof OF PrOp. \[HCoNway3bR\] (WheN dO 3 braiD links Have $P_{\WiDehat \beta} = P_{\ | 2+1} \{C_{E_2} C_{E_1} - \ prod_{k=1} ^{2}C_{ e_{ k, 2}}C_{e _{k,1}} \}\,,\ mbox { when $r=2$.} \nonumb er \e nd { alig n ed }$$
\[Conwa y 3b r a ugm en te d\]
E q. \[Al exa nder3br Approx\] i s e ss entially giv e nby Prop. 4 .2, p. 13-14 [@ 69] .
Ind uc tio n , plu s ( \[Con wayrec u r\]) a nd (\[Con wa y sumrel \ ]) fort h efirs t and last equati o ns . Eq. \[Conwayd iffofp ro d la s t \]isalso helpf ul to p r ove the or d e r of magnitude res ult in (\[C o nwa y3brAp pr ox\ ] ).
Wh en $\ ga m ma\in B_3$ an d $a >0$ we h ave: $ $ \begin{ a ligned}
P_{\w ide hat {\a l ph a_ 2^3 \ g amm a }} & = & v ^6 P_{\w id eh at {\ gamm a } } + P _{T _{w( \gamm a)+5}} - v^6P_{ T_{w ( \ga mma)+ 1}} \ ,, \ \%
P_{ \wideh at {\ al pha_2^{3a} \gam ma}} &=&
v^ {6a }P_{ \w ideha t {\gam ma} }+ \sum_{ j=1}^av ^{6 a- 6 j } P _{T_{w(\gamma)+6j- 1} } -\sum_{j= 1}^a v ^ {6 +6 a -6j} P_{ T_ {w( \gam m a )+6j- 5}}\ , ,\\
%
P_{ \wideh a t{\ alpha_2 ^{ -3a} \ ga mma } } &=&v^{- 6a} P_ {\wideha t {\g a mma} }+ \sum_{ j =1}^a v^{6j-6 a }P _ {T _ {w(\ gam ma)-6j+1}}- \s u m_{j =1}^ a v ^{6 j -6a-6 } P_{ T_ { w( \ gamma)-6j+5}}\,.\en d{ aligne d}$$\[Homflypt3br augmented\ ] Inductio n, p l us Prop. \[Conway 3brau gmented\]a nd Prop. \[H3 braidstd formprop\ ] for thefir stequ ati o n .
Proof of Pro p . \[H Co nway3br \](When d o 3 br aid li nk s have $P _{\wideh at \ be ta } = P_{\ | 2+1} \{C_{E_2}_C_{E_1} -_\prod_{k=1}^{2} C_{e_{k,2}} C_{e_{k,1}} \}\,,_\mbox{ when_$r=2$.}_\nonumber \end{aligned}$$
\[Conway3braugmented\]
Eq. \[Alexander3brApprox\]_is_essentially given by_Prop. 4.2, p. 13-14 [@69].
Induction,_plus (\[Conwayrecur\]) and (\[Conwaysumrel\])_for the first_and_last equations. Eq. \[Conwaydiffofprodlast\] is also helpful to prove the order of magnitude result in_(\[Conway3brApprox\]).
When_$\gamma \in_B_3$_and_$a >0$ we have: $$\begin{aligned}
P_{\widehat_{\alpha_2^3 \gamma} } &=& v^6_P_{\widehat {\gamma}_} + P_{T_{w(\gamma)+5}} - v^6 P_{T_{w(\gamma)+1}} \,, \\
%
P_{\widehat_{\alpha_2^{3a}_\gamma} } &=&_
v^{6a} P_{\widehat {\gamma} }+ \sum_{j=1}^a v^{6a-6j} P_{T_{w(\gamma)+6j-1}} - \sum_{j=1}^a_v^{6+6a-6j} P_{T_{w(\gamma)+6j-5}}\,, \\
%
P_{\widehat {\alpha_2^{-3a} \gamma} }_&=&
v^{-6a} P_{\widehat_{\gamma}_}+_\sum_{j=1}^a v^{6j-6a} P_{T_{w(\gamma)-6j+1}} -_\sum_{j=1}^a v^{6j-6a-6} P_{T_{w(\gamma)-6j+5}}\,.\end{aligned}$$ \[Homflypt3braugmented\]
Induction, plus Prop. \[Conway3braugmented\]_and Prop. \[H3braidstdformprop\] for the first equation.
Proof_of Prop. \[HConway3br\] (When do 3 braid links_have $P_{\widehat \beta} = P_{\ |
{G}}$ in $X\left({}^{\vee}{G}^{\Gamma}\right)$. Then we define the micro-packet of geometric parameters attached to $\varphi$ as $$\begin{aligned}
\Xi^{z}({G}/\mathbb{R})_{\varphi}=\Xi\left(X\left({}^{\vee}{G}^{\Gamma}\right),{}^{\vee}{G}\right)^{mic}_{S_{\varphi}}.\end{aligned}$$ For any complete parameter $\xi\in \Xi^{z}\left({G}/\mathbb{R}\right)$, let $\pi(\xi)$ be the representation in $\Pi^{z}({G}/\mathbb{R})$ associated to $\xi$ by Theorem \[theo:4.1\]. Then the micro-packet of $\varphi$ is defined as $$\begin{aligned}
\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}=\{\pi(\xi'):\xi'\in \Xi^{z}({G}/\mathbb{R})^{mic}_{\varphi}\}.\end{aligned}$$ Finally, let $\delta$ be a strong real form of $G^{\Gamma}$, then we define the restriction of $\Pi({G}/\mathbb{R})^{mic}_{\varphi}$ to $\delta$ as $$\begin{aligned}
\Pi^{z}({G}(\mathbb{R},\delta))^{mic}_{\varphi}=\{\pi\in \Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}: \pi \text{ is a representation of }G(\mathbb{R},\delta) \}.\end{aligned}$$
We notice that the Langlands packet attached to a Langlands parameter $\varphi$ is always contained in the corresponding micro-packet $$\begin{aligned}
\label{eq:contentionoflpackets}
\Pi^{z}({G}/\mathbb{R})_{\varphi}\subset\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}.\end{aligned}$$ This is a consequence of $(ii)$ of the following lemma. Point $(i)$ of the lemma will show to be quite useful later in this section. For a proof, see Lemma 19.14 [@ABV].
\[lem:lemorbit\] Let $\eta=(S,\mathcal{V})$ and $\eta'=\left(S',\mathcal{V}'\right)$ be two geometric parameters for ${}^{\vee}{G}^{\Gamma}$.
i. If $\eta'\in \Xi\left(X\left({}^{\vee}{G | { G}}$ in $ X\left({}^{\vee}{G}^{\Gamma}\right)$. Then we define the micro - packet of geometric parameters attached to $ \varphi$ as $ $ \begin{aligned }
\Xi^{z}({G}/\mathbb{R})_{\varphi}=\Xi\left(X\left({}^{\vee}{G}^{\Gamma}\right),{}^{\vee}{G}\right)^{mic}_{S_{\varphi}}.\end{aligned}$$ For any accomplished argument $ \xi\in \Xi^{z}\left({G}/\mathbb{R}\right)$, let $ \pi(\xi)$ be the representation in $ \Pi^{z}({G}/\mathbb{R})$ consociate to $ \xi$ by Theorem \[theo:4.1\ ]. Then the micro - mailboat of $ \varphi$ is defined as $ $ \begin{aligned }
\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}=\{\pi(\xi'):\xi'\in \Xi^{z}({G}/\mathbb{R})^{mic}_{\varphi}\}.\end{aligned}$$ Finally, permit $ \delta$ be a strong real kind of $ G^{\Gamma}$, then we define the restriction of $ \Pi({G}/\mathbb{R})^{mic}_{\varphi}$ to $ \delta$ as $ $ \begin{aligned }
\Pi^{z}({G}(\mathbb{R},\delta))^{mic}_{\varphi}=\{\pi\in \Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi }: \pi \text { is a theatrical performance of } G(\mathbb{R},\delta) \}.\end{aligned}$$
We notice that the Langlands packet attached to a Langlands parameter $ \varphi$ is constantly contained in the corresponding micro - packet $ $ \begin{aligned }
\label{eq: contentionoflpackets }
\Pi^{z}({G}/\mathbb{R})_{\varphi}\subset\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}.\end{aligned}$$ This is a consequence of $ (ii)$ of the following lemma. Point $ (i)$ of the lemma will show to be quite useful later in this part. For a proof, see Lemma 19.14 [ @ABV ].
\[lem: lemorbit\ ] Let $ \eta=(S,\mathcal{V})$ and $ \eta'=\left(S',\mathcal{V}'\right)$ be two geometric parameter for $ { } ^{\vee}{G}^{\Gamma}$.
i. If $ \eta'\in \Xi\left(X\left({}^{\vee}{G | {G}}$ ij $X\left({}^{\vee}{G}^{\Gamma}\right)$. Thtn we define the micro-parket of geometrkc parameters attached to $\vacphi$ as $$\btyin{aligned}
\Xi^{z}({G}/\mathbb{F})_{\varphi}=\Xi\peft(X\lefr({}^{\vee}{J}^{\Gamma}\right),{}^{\vee}{G}\rmfht)^{mic}_{S_{\ycrphi}}.\skd{aliynxd}$$ For any compkete paramater $\xi\in \Xi^{z}\lafg({G}/\lathbb{R}\right)$, let $\pi(\xi)$ be the represqntatiom ln $\Pi^{z}({G}/\mathbb{R})$ asspsiatsd to $\xi$ by Theorem \[theo:4.1\]. Then the jicro-pabket of $\varphi$ is defined as $$\begin{aligned}
\Pi^{x}({G}/\mahhbb{R})^{mic}_{\varphi}=\{\pi(\xi'):\di'\in \Xi^{z}({G}/\majgbb{W})^{nic}_{\varphi}\}.\end{xligned}$$ Finally, let $\demta$ be a strong real form of $G^{\Gxmma}$, chen we defunw tjg restrictioi of $\Pp({G}/\mathbb{R})^{mic}_{\varphi}$ to $\delta$ ss $$\begin{alignec}
\Pi^{v}({G}(\marhbb{R},\delta))^{mic}_{\varphi}=\{\pi\mn \Pi^{z}({G}/\mathbb{R})^{mic}_{\varkhi}: \pi \texd{ ns a representation od }T(\mathtb{R},\dalta) \}.\wnd{xlifnxd}$$
Ws notife vhat the Lahglands pacjet attached to a Lsndoands parametsr $\var[hy$ is always contained in the correspondpng jicro-packet $$\begin{alignee}
\label{eq:contentionoflkackets}
\Pi^{z}({D}/\mathbb{R})_{\varphi}\subset\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}.\end{aligneg}$$ Thia is a zindequence of $(ii)$ of the following lemma. Point $(i)$ kf tme lemma will shjw to be quotf ifeful later iv this sedtion. For a proof, dee Lemia 19.14 [@AVV].
\[lem:lemowbit\] Let $\eta=(S,\mathcal{V})$ and $\eta'=\ledt(S',\mathcal{V}'\rpght)$ be two geometric 'arameters fur ${}^{\vge}{G}^{\Gamka}$.
i. If $\eta'\in \Xi\left(X\leyt({}^{\vee}{G | {G}}$ in $X\left({}^{\vee}{G}^{\Gamma}\right)$. Then we define the geometric attached to as $$\begin{aligned} \Xi^{z}({G}/\mathbb{R})_{\varphi}=\Xi\left(X\left({}^{\vee}{G}^{\Gamma}\right),{}^{\vee}{G}\right)^{mic}_{S_{\varphi}}.\end{aligned}$$ \Xi^{z}\left({G}/\mathbb{R}\right)$, $\pi(\xi)$ be the in $\Pi^{z}({G}/\mathbb{R})$ associated $\xi$ by Theorem \[theo:4.1\]. Then the of $\varphi$ is defined as $$\begin{aligned} \Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}=\{\pi(\xi'):\xi'\in \Xi^{z}({G}/\mathbb{R})^{mic}_{\varphi}\}.\end{aligned}$$ Finally, let $\delta$ be a real form of $G^{\Gamma}$, then we define the restriction of $\Pi({G}/\mathbb{R})^{mic}_{\varphi}$ to $\delta$ $$\begin{aligned} \Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}: \text{ a representation of }G(\mathbb{R},\delta) \}.\end{aligned}$$ We notice that the Langlands packet attached to a Langlands parameter is always contained in the corresponding micro-packet $$\begin{aligned} \Pi^{z}({G}/\mathbb{R})_{\varphi}\subset\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}.\end{aligned}$$ This is a of $(ii)$ of the following Point of the will to quite useful later this section. For a proof, see Lemma 19.14 [@ABV]. \[lem:lemorbit\] Let $\eta=(S,\mathcal{V})$ and $\eta'=\left(S',\mathcal{V}'\right)$ be two geometric for ${}^{\vee}{G}^{\Gamma}$. $\eta'\in \Xi\left(X\left({}^{\vee}{G | {G}}$ in $X\left({}^{\vee}{G}^{\Gamma}\right)$. TheN we define tHe micRo-pAckEt Of geOmetRic parameters aTTachEd to $\varphi$ as $$\begin{alignEd}
\Xi^{z}({g}/\mAThbb{r})_{\VaRphi}=\XI\left(X\lEFt({}^{\VEE}{G}^{\GAmMa}\RigHt),{}^{\VEe}{g}\righT)^{miC}_{S_{\varphI}}.\end{aligneD}$$ FoR aNy complete paRAmEter $\xi\in \Xi^{Z}\leFt({G}/\mathbb{R}\riGht)$, Let $\pi(\xI)$ bE thE RepreSenTatioN in $\Pi^{z}({g}/\Mathbb{r})$ associatEd TO $\xi$ by THEorem \[thEO:4.1\]. thEn thE micro-packet of $\varPHi$ IS defined as $$\begiN{alignEd}
\pI^{z}({g}/\MAthBb{R})^{Mic}_{\varphi}=\{\pI(\xI'):\xi'\in \xI^{z}({G}/\mathBB{R})^{MIC}_{\VarPHi}\}.\end{aligned}$$ FInally, let $\deLTa$ bE a stroNg ReaL Form of $g^{\GammA}$, tHEn wE define the rEstrIction of $\PI({G}/\mathBB{R})^{mic}_{\vaRPhi}$ to $\deLta$ as $$\bEgiN{alIgneD}
\pi^{Z}({G}(\MatHbB{r},\deLTa))^{Mic}_{\VArpHi}=\{\pi\in \Pi^{Z}({G}/\MaThbb{R})^{Mic}_{\vARPHI}: \pi \tExt{ Is a rEpresEntation of }G(\maThbB{R},\deLTa) \}.\eNd{aliGned}$$
WE notIcE that The LanGlandS pAcket attached to A LanGlands parAmeTeR $\vaRpHi$ is aLWays coNtaIneD in the cOrrespoNDinG mICRO-pAcket $$\begin{aligned}
\lAbEL{Eq:ContentiOnoflpACkEtS}
\pi^{z}({G}/\mathBb{r})_{\vaRphi}\SUBset\PI^{z}({G}/\mAThBb{R})^{mic}_{\vaRphi}.\enD{AlIgNed}$$ This Is A conseQuEncE of $(Ii)$ of tHE folLowing Lemma. PoiNt $(i)$ of THe lemma will shoW To be quite usefUL lATEr IN thiS seCtion. For a prOof, sEE LemMa 19.14 [@ABv].
\[LeM:leMOrbit\] let $\etA=(S,\MAtHCal{V})$ and $\eta'=\left(S',\mathCaL{V}'\righT)$ be twO geometric parAmeters for ${}^{\VEE}{g}^{\Gamma}$.
i. IF $\eta'\IN \XI\Left(X\left({}^{\vee}{G | {G}}$ in $X\left({}^{\vee} {G}^{\Gamm a}\ri ght )$. T henwe d efine the micr o -pac ket of geometric param eters a t tach e dto $\ varphi$ as $ $\b eg in {al ig n ed }
\Xi ^{z }({G}/\ mathbb{R}) _{\ va rphi}=\Xi\le f t( X\left({}^ {\v ee}{G}^{\Gam ma} \right ), {}^ { \vee} {G} \righ t)^{mi c }_{S_{ \varphi}} .\ e nd{ali g ned}$$F o ranycomplete paramete r $ \ xi\in \Xi^{z}\ left({ G} / \m a t hbb {R} \right)$,le t $\p i (\xi)$b et h e re p resentation i n $\Pi^{z}( { G}/ \mathb b{ R}) $ assoc iated t o $\ xi$ by Theo rem\[theo:4. 1\]. T h en them icro-pa cket o f $ \va rphi $ i sdef in e d a s $ $\b e gin {aligned }\P i^{z} ({G} / \ m a thbb {R} )^{m ic}_{ \varphi}=\{\p i(\ xi') : \xi '\in\Xi^{ z}({ G} /\mat hbb{R} )^{mi c} _{\varphi}\}.\e nd{a ligned}$$ Fi na lly ,let $ \ delta$ be astrongreal fo r m o f$ G ^ {\ Gamma}$, then we d ef i n ethe rest rictio n o f$ \Pi({G}/ \m ath bb{R } ) ^{mic }_{\ v ar phi}$ to $\del t a$ a s $$\be gi n{alig ne d}\Pi ^{z}( { G}(\ mathbb {R},\del ta))^ { mic}_{\varphi} = \{\pi\in \Pi^ { z} ( { G} / \mat hbb {R})^{mic}_ {\va r phi} : \p i \ tex t { isa rep re s en t ation of }G(\mathbb {R },\del ta) \ }.\end{aligne d}$$
We n o t i ce thattheL an g lands packet a ttach ed to a La n glands p arame ter $\va rphi$ isa l ways con tai ned in th e co rresponding m i c ro-p ac ket $$\ beg in{alig ned }
\ lab el{ eq :contenti onoflpac ke ts }\P i^{ z}({G } /\mathbb {R })_ {\ var phi}\ s ubset\ Pi^{z }({G }/ \m a thb b{R})^{ m ic } _ {\va rp hi }.\e nd{ al igned }$$T his is a c onsequenc e o f $(i i) $of thefollowing lem ma . Point $( i) $ o f thel e mma will show to be quite usefu l laterinthissect ion. Fora p roof,see Lemma19.14[@ABV ].
\ [ l em:le m o rb it\ ]Let $\eta= ( S ,\m athca l{ V})$ and $\ eta'=\left(S',\mat h cal {V}'\right)$betwog e om etr i cp ara me t ers f or ${}^{\vee}{G }^{\Gamma} $.
i . If $\et a '\i n\Xi\lef t(X\lef t({}^ { \vee}{G | {G}}$ in_$X\left({}^{\vee}{G}^{\Gamma}\right)$. Then_we define the micro-packet_of geometric_parameters_attached to_$\varphi$_as $$\begin{aligned}
\Xi^{z}({G}/\mathbb{R})_{\varphi}=\Xi\left(X\left({}^{\vee}{G}^{\Gamma}\right),{}^{\vee}{G}\right)^{mic}_{S_{\varphi}}.\end{aligned}$$ For_any complete parameter_$\xi\in \Xi^{z}\left({G}/\mathbb{R}\right)$, let $\pi(\xi)$_be the representation_in_$\Pi^{z}({G}/\mathbb{R})$ associated to $\xi$ by Theorem \[theo:4.1\]. Then the micro-packet of $\varphi$ is defined_as_$$\begin{aligned}
\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}=\{\pi(\xi'):\xi'\in \Xi^{z}({G}/\mathbb{R})^{mic}_{\varphi}\}.\end{aligned}$$_Finally,_let_$\delta$ be a strong real_form of $G^{\Gamma}$, then we_define the_restriction of $\Pi({G}/\mathbb{R})^{mic}_{\varphi}$ to $\delta$ as $$\begin{aligned}
\Pi^{z}({G}(\mathbb{R},\delta))^{mic}_{\varphi}=\{\pi\in \Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}:_\pi_\text{ is a_representation of }G(\mathbb{R},\delta) \}.\end{aligned}$$
We notice that the Langlands packet_attached to a Langlands parameter $\varphi$_is always contained_in_the_corresponding micro-packet $$\begin{aligned}
\label{eq:contentionoflpackets}
\Pi^{z}({G}/\mathbb{R})_{\varphi}\subset\Pi^{z}({G}/\mathbb{R})^{mic}_{\varphi}.\end{aligned}$$ This_is a consequence of $(ii)$ of_the following lemma. Point $(i)$ of_the lemma will show to be quite_useful later in this section. For_a proof, see Lemma 19.14_[@ABV].
\[lem:lemorbit\] Let_$\eta=(S,\mathcal{V})$ and $\eta'=\left(S',\mathcal{V}'\right)$ be two_geometric parameters for_${}^{\vee}{G}^{\Gamma}$.
i. _If $\eta'\in \Xi\left(X\left({}^{\vee}{G |
left( \phi\right) }{\partial\phi}=0,
\label{motequ}$$ has an implicit solution$$\beta_{\mu}x^{\mu}+\int\frac{\sqrt{-\beta^{2}}}{\sqrt{2\left[ \frac{1}{2}m^{2}\phi^{2}+V\left( \phi\right) -G\right] }}d\phi\equiv F_{1}\left(
x^{\mu},\phi\right) =0. \label{gensol}$$
From the solution (\[gensol\]) we obtain $$\partial_{\mu}\phi=-\frac{\frac{\partial F_{1}\left( x^{\mu},\phi\right)
}{\partial x^{\mu}}}{\frac{\partial F_{1}\left( x^{\mu},\phi\right)
}{\partial\phi}}=-\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\phi
^{2}+V\left( \phi\right) -G\right] }}{\sqrt{-\beta^{2}}}. \label{rdphi1}$$
Furthermore, by Eq. (\[rdphi1\]), letting$$\partial_{\mu}\phi+\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\phi
^{2}+V\left( \phi\right) -G\right] }}{\sqrt{-\beta^{2}}}\equiv F_{2}\left(
x^{\mu},\phi,\partial_{\mu}\phi\right) =0,$$ we arrive at $$\begin{aligned}
\square\phi & =\partial^{\mu}\partial_{\mu}\phi=\frac{\partial}{\partial
x_{\mu}}\partial_{\mu}\phi\nonumber\\
& =-\frac{\frac{\partial F_{2}\left( x^{\mu},\phi,\partial_{\mu}\phi\right)
}{\partial x_{\mu}}}{\frac{\partial F_{2}\left( x^{\mu},\phi,\partial_{\mu
}\phi\right) }{\partial_{\mu}\phi}}-\frac{\frac{\partial F_{2}\left( x^{\mu
},\phi,\partial_{\mu}\phi\right) }{\partial\phi}}{\frac{\partial F_{2}\left(
x^{\mu},\phi | left ( \phi\right) } { \partial\phi}=0,
\label{motequ}$$ has an implicit solution$$\beta_{\mu}x^{\mu}+\int\frac{\sqrt{-\beta^{2}}}{\sqrt{2\left [ \frac{1}{2}m^{2}\phi^{2}+V\left ( \phi\right) -G\right ] } } d\phi\equiv F_{1}\left (
x^{\mu},\phi\right) = 0. \label{gensol}$$
From the solution (\[gensol\ ]) we obtain $ $ \partial_{\mu}\phi=-\frac{\frac{\partial F_{1}\left ( x^{\mu},\phi\right)
} { \partial x^{\mu}}}{\frac{\partial F_{1}\left ( x^{\mu},\phi\right)
} { \partial\phi}}=-\frac{\beta_{\mu}\sqrt{2\left [ \frac{1}{2}m^{2}\phi
^{2}+V\left ( \phi\right) -G\right ] } } { \sqrt{-\beta^{2 } } }. \label{rdphi1}$$
Furthermore, by Eq. (\[rdphi1\ ]), letting$$\partial_{\mu}\phi+\frac{\beta_{\mu}\sqrt{2\left [ \frac{1}{2}m^{2}\phi
^{2}+V\left ( \phi\right) -G\right ] } } { \sqrt{-\beta^{2}}}\equiv F_{2}\left (
x^{\mu},\phi,\partial_{\mu}\phi\right) = 0,$$ we arrive at $ $ \begin{aligned }
\square\phi & = \partial^{\mu}\partial_{\mu}\phi=\frac{\partial}{\partial
x_{\mu}}\partial_{\mu}\phi\nonumber\\
& = -\frac{\frac{\partial F_{2}\left ( x^{\mu},\phi,\partial_{\mu}\phi\right)
} { \partial x_{\mu}}}{\frac{\partial F_{2}\left ( x^{\mu},\phi,\partial_{\mu
} \phi\right) } { \partial_{\mu}\phi}}-\frac{\frac{\partial F_{2}\left ( x^{\mu
}, \phi,\partial_{\mu}\phi\right) } { \partial\phi}}{\frac{\partial F_{2}\left (
x^{\mu},\phi | lefh( \phi\right) }{\partial\phi}=0,
\lxbel{motequ}$$ has cb implmcit somution$$\bega_{\mu}x^{\mu}+\int\frac{\sqrt{-\beta^{2}}}{\sqrt{2\leht[ \drac{1}{2}m^{2}\kki^{2}+V\left( \phi\right) -G\fight] }}d\pji\equiv D_{1}\lefu(
x^{\mu},\phi\right) =0. \lausl{gensol}$$
From fme sonntion (\[gensol\]) we obtain $$\pastial_{\mu}\phi=-\frac{\xrxc{\'artial F_{1}\left( x^{\mu},\phi\right)
}{\partial x^{\mt}}}{\frac{\pattlal F_{1}\left( x^{\mu},\khi\ribrt)
}{\padnicl\phi}}=-\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\pgi
^{2}+V\left( \phi\right) -G\tight] }}{\sqrt{-\beta^{2}}}. \label{rdphi1}$$
Vurtjermore, by Eq. (\[rdphl1\]), letting$$\patfiaj_{\nu}\phi+\frac{\betx_{\mu}\sqrt{2\lefu[ \frac{1}{2}m^{2}\phi
^{2}+V\meft( \phi\right) -G\right] }}{\sqrt{-\betx^{2}}}\equir F_{2}\left(
x^{\mu},\pyi,\parhhal_{\mu}\phi\rigit) =0,$$ wv arrive at $$\bcbin{alicned}
\squsre\phi & =\partisl^{\mn}\parrial_{\mu}\phi=\frac{\partial}{\pertial
x_{\mu}}\partial_{\mu}\phy\nonumber\\
& =-\yrac{\frac{\partial F_{2}\left( z^{\mu},\phh,\pardial_{\oy}\phk\rifhv)
}{\padtial d_{\mu}}}{\hrac{\partial F_{2}\left( x^{\mu},\phi,\partial_{\mu
}\phi\righu) }{\[qrtial_{\mu}\phi}}-\frzc{\frac{\[awtial F_{2}\left( x^{\mu
},\phi,\partial_{\mu}\phi\right) }{\pdrtjal\phi}}{\frac{\partial F_{2}\left(
z^{\mu},\phi | left( \phi\right) }{\partial\phi}=0, \label{motequ}$$ has an implicit \phi\right) }}d\phi\equiv F_{1}\left( =0. \label{gensol}$$ From $$\partial_{\mu}\phi=-\frac{\frac{\partial x^{\mu},\phi\right) }{\partial x^{\mu}}}{\frac{\partial x^{\mu},\phi\right) }{\partial\phi}}=-\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\phi \phi\right) -G\right] }}{\sqrt{-\beta^{2}}}. \label{rdphi1}$$ Furthermore, by (\[rdphi1\]), letting$$\partial_{\mu}\phi+\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\phi ^{2}+V\left( \phi\right) -G\right] }}{\sqrt{-\beta^{2}}}\equiv F_{2}\left( x^{\mu},\phi,\partial_{\mu}\phi\right) =0,$$ we arrive at \square\phi & =\partial^{\mu}\partial_{\mu}\phi=\frac{\partial}{\partial x_{\mu}}\partial_{\mu}\phi\nonumber\\ & =-\frac{\frac{\partial F_{2}\left( x^{\mu},\phi,\partial_{\mu}\phi\right) }{\partial x_{\mu}}}{\frac{\partial F_{2}\left( x^{\mu},\phi,\partial_{\mu }\phi\right) F_{2}\left( },\phi,\partial_{\mu}\phi\right) F_{2}\left( | left( \phi\right) }{\partial\phi}=0,
\labEl{motequ}$$ haS an imPliCit SoLutiOn$$\beTa_{\mu}x^{\mu}+\int\frac{\SQrt{-\bEta^{2}}}{\sqrt{2\left[ \frac{1}{2}m^{2}\phi^{2}+V\leFt( \phi\RiGHt) -G\rIGhT] }}d\phi\Equiv F_{1}\lEFt(
X^{\MU},\phI\rIgHt) =0. \lAbEL{gEnsol}$$
froM the solUtion (\[gensoL\]) we ObTain $$\partial_{\mU}\PhI=-\frac{\frac{\pArtIal F_{1}\left( x^{\mu},\pHi\rIght)
}{\paRtIal X^{\Mu}}}{\fraC{\paRtial f_{1}\left( x^{\MU},\phi\riGht)
}{\partiaL\pHI}}=-\frac{\bETa_{\mu}\sqrT{2\LEfT[ \fraC{1}{2}m^{2}\phi
^{2}+V\left( \phi\righT) -g\rIGht] }}{\sqrt{-\beta^{2}}}. \labEl{rdphI1}$$
FURtHERmoRe, bY Eq. (\[rdphi1\]), leTtIng$$\paRTial_{\mu}\pHI+\fRAC{\BetA_{\Mu}\sqrt{2\left[ \fraC{1}{2}m^{2}\phi
^{2}+V\left( \pHI\riGht) -G\riGhT] }}{\sqRT{-\beta^{2}}}\eQuiv F_{2}\LeFT(
x^{\mU},\phi,\partial_{\Mu}\phI\right) =0,$$ we aRrive aT $$\Begin{alIGned}
\squAre\phi & =\ParTiaL^{\mu}\pARtIaL_{\mu}\PhI=\FraC{\PaRtiAL}{\paRtial
x_{\mu}}\PaRtIal_{\mu}\Phi\nONUMBer\\
& =-\fRac{\Frac{\PartiAl F_{2}\left( x^{\mu},\phi,\ParTial_{\MU}\phI\righT)
}{\partIal x_{\Mu}}}{\Frac{\pArtial f_{2}\left( X^{\mU},\phi,\partial_{\mu
}\phI\rigHt) }{\partial_{\Mu}\pHi}}-\FraC{\fRac{\paRTial F_{2}\lEft( X^{\mu
},\Phi,\partIal_{\mu}\phI\RigHt) }{\PARTiAl\phi}}{\frac{\partial F_{2}\lEfT(
X^{\Mu},\Phi | left( \phi\right) }{\par tial\phi}= 0,
\l abe l{m ot equ} $$ h as an implicit solu tion$$\beta_{\mu}x^{\m u}+\i nt \ frac { \s qrt{- \beta^{ 2 }} } { \sq rt {2 \le ft [ \frac {1} {2}m^{2 }\phi^{2}+ V\l ef t( \phi\rig h t) -G\right ] }}d\phi\equi v F _{1}\l ef t(x ^{\mu },\ phi\r ight)=0. \l abel{gens ol } $$
Fr o m the s o l ut ion(\[gensol\]) we o b ta i n $$\partial_{ \mu}\p hi = -\ f r ac{ \fr ac{\partia lF_{1} \ left( x ^{ \ m u },\ p hi\right)
}{\ partial x^{ \ mu} }}{\fr ac {\p a rtialF_{1} \l e ft( x^{\mu},\ phi\ right)
}{ \parti a l\phi}} = -\frac{ \beta_ {\m u}\ sqrt { 2\ le ft[ \ fra c {1 }{2 } m^{ 2}\phi
^ {2 }+ V\lef t( \ p h i \rig ht) -G \righ t] }}{\sqrt{ -\b eta^ { 2}} }. \l abel{ rdph i1 }$$
Furthe rmore ,by Eq. (\[rdphi 1\]) , letting $$\ pa rti al _{\mu } \phi+\ fra c{\ beta_{\ mu}\sqr t {2\ le f t [ \frac{1}{2}m^{2}\p hi ^ {2 }+V\left ( \ph i \r ig h t) -G\r ig ht] }} { \ sqrt{ -\be t a^ {2}}}\eq uiv F_ { 2} \l eft(
x^ {\ mu},\p hi ,\p art ial_{ \ mu}\ phi\ri ght) =0 ,$$ w e arrive at $$\ b egin{aligned} \s q u ar e \phi & =\partial^ {\mu } \par tial _ {\ mu} \ phi=\ frac{ \p a rt i al}{\partial
x_{\mu }} \parti al_{\ mu}\phi\nonum ber\\
& = - \ f rac{\fra c{\p a rt i al F_{2}\left( x^{ \mu},\phi, \ partial_ {\mu} \phi\rig ht)
}{\pa r t ial x_{\ mu} }}{ \fr ac{ \ p ar tial F_{2}\le f t ( x ^{ \mu},\p hi, \partia l_{ \mu
}\ phi \r ight) }{ \partial _{ \m u} \p hi} }-\fr a c{\frac{ \p art ia l F _{2}\ l eft( x^{\m u
}, \p hi , \pa rtial_{ \ mu } \ phi\ ri gh t) }{\ pa rtial \phi } }{\ frac{\p artial F_ {2} \ left (x^ {\mu},\ phi | left( _\phi\right) _}{\partial\phi}=0,
\label{motequ}$$ has an implicit_solution$$\beta_{\mu}x^{\mu}+\int\frac{\sqrt{-\beta^{2}}}{\sqrt{2\left[ _\frac{1}{2}m^{2}\phi^{2}+V\left(_ \phi\right)__-G\right] }}d\phi\equiv_F_{1}\left(
x^{\mu},\phi\right) =0._\label{gensol}$$
From the solution (\[gensol\])_we obtain $$\partial_{\mu}\phi=-\frac{\frac{\partial_F_{1}\left(_ x^{\mu},\phi\right)
}{\partial x^{\mu}}}{\frac{\partial F_{1}\left( x^{\mu},\phi\right)
}{\partial\phi}}=-\frac{\beta_{\mu}\sqrt{2\left[ \frac{1}{2}m^{2}\phi
^{2}+V\left( \phi\right) -G\right] }}{\sqrt{-\beta^{2}}}._\label{rdphi1}$$
Furthermore,_by Eq._(\[rdphi1\]),_letting$$\partial_{\mu}\phi+\frac{\beta_{\mu}\sqrt{2\left[_ \frac{1}{2}m^{2}\phi
^{2}+V\left( \phi\right) _-G\right] }}{\sqrt{-\beta^{2}}}\equiv F_{2}\left(
x^{\mu},\phi,\partial_{\mu}\phi\right) _=0,$$ we_arrive at $$\begin{aligned}
\square\phi & =\partial^{\mu}\partial_{\mu}\phi=\frac{\partial}{\partial
x_{\mu}}\partial_{\mu}\phi\nonumber\\
& =-\frac{\frac{\partial_F_{2}\left(_ x^{\mu},\phi,\partial_{\mu}\phi\right)
}{\partial x_{\mu}}}{\frac{\partial_F_{2}\left( x^{\mu},\phi,\partial_{\mu
}\phi\right) }{\partial_{\mu}\phi}}-\frac{\frac{\partial F_{2}\left( x^{\mu
},\phi,\partial_{\mu}\phi\right) _}{\partial\phi}}{\frac{\partial F_{2}\left(
x^{\mu},\phi |
FT at the edge of the SFib state, this theory for the parent medium leads to the theory $\mathrm{Fib}\times\overline{\mathbb{Z}_2\times\mathbb{Z}_2}|_R$, or equivalently $SU(2)_3\times\overline{SU(2)}_1|_R$ for the SFib to vacuum edge.
There is one additional subtlety, however, which is that the SFib phase is superconducting. Hence, Bogoliubov-de Gennes (BdG) quasiparticles, rather than electrons, are the quasiparticles of a superconductor with (possibly mobile) vortices. Thus, the “electron” of the previous paragraph should actually be interpreted as a BdG quasiparticle. Our previous statements are unchanged but must be supplemented by the fact that a $-1$ results when a BdG quasiparticle encircles an $hc/2e$ vortex.
CS theory of topological superconductors with $SU(2)\times SU(2)$ gauge symmetry {#secsu2su2}
================================================================================
Combining the preceding elements, we postulate a bulk low-energy effective field theory for the electronic SFib state that takes the form of an $SU(2)_3\times\overline{SU(2)}_1|_R$ Chern-Simons gauge theory: $$\label{eqn:Fib-electron-effective-action}
{\cal L} = {\cal L}_\text{top} + {\cal L}_\text{SC} + {\cal L}_\text{qp}$$ In this expression, $$\begin{gathered}
\label{eqn:Fib-electron-top-action}
{\cal L}_\text{top} =\frac{3}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(a_\mu \partial_\nu a_\lambda + \frac{2}{3} a_\mu a_\nu a_\lambda \right)\\
- \frac{1}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(b_\mu \partial_\nu b_\lambda + \frac{2}{3} b_\mu b_\nu b_\lambda \right),\end{gathered}$$ Here, $a_\mu$ and $b_\mu$ are SU(2) gauge fields. Each is a triplet of gauge fields, corresponding to the three generators of | FT at the edge of the SFib state, this theory for the parent medium lead to the hypothesis $ \mathrm{Fib}\times\overline{\mathbb{Z}_2\times\mathbb{Z}_2}|_R$, or equivalently $ SU(2)_3\times\overline{SU(2)}_1|_R$ for the SFib to vacuum edge.
There is one extra subtlety, however, which is that the SFib phase is superconducting. therefore, Bogoliubov - de Gennes (BdG) quasiparticles, rather than electrons, are the quasiparticle of a superconductor with (possibly mobile) vortices. Thus, the “ electron ” of the former paragraph should actually be interpret as a BdG quasiparticle. Our previous statement are unchanged but must be supplemented by the fact that a $ -1 $ solution when a BdG quasiparticle encircle an $ hc/2e$ vortex.
CS theory of topological superconductors with $ SU(2)\times SU(2)$ gauge symmetry { # secsu2su2 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Combining the predate elements, we postulate a bulk low - department of energy effective field theory for the electronic SFib state of matter that takes the kind of an $ SU(2)_3\times\overline{SU(2)}_1|_R$ Chern - Simons gauge theory: $ $ \label{eqn: Fib - electron - effective - action }
{ \cal L } = { \cal L}_\text{top } + { \cal L}_\text{SC } + { \cal L}_\text{qp}$$ In this expression, $ $ \begin{gathered }
\label{eqn: Fib - electron - top - action }
{ \cal L}_\text{top } = \frac{3}{4\pi}\epsilon^{\mu\nu\lambda } \text{Tr}\left(a_\mu \partial_\nu a_\lambda + \frac{2}{3 } a_\mu a_\nu a_\lambda \right)\\
- \frac{1}{4\pi}\epsilon^{\mu\nu\lambda } \text{Tr}\left(b_\mu \partial_\nu b_\lambda + \frac{2}{3 } b_\mu b_\nu b_\lambda \right),\end{gathered}$$ Here, $ a_\mu$ and $ b_\mu$ are SU(2) bore fields. Each is a triplet of gauge fields, corresponding to the three generators of | FT wt the edge of the SFib rtate, this theoty for tie pareht mediuo leads to the theory $\mathrm{Hib}\tumes\ocerline{\mathbb{Z}_2\times\matfbb{Z}_2}|_R$, or vquivalenrly $WY(2)_3\times\overline{SU(2)}_1|_R$ njr tgc SFiy vo vacuum edge.
Tmere is one additional suttuecy, however, which is that the SFib phwse is xuoerconducting. Renct, Bjgoljlbiv-de Gennes (BdG) quasiparticlea, ratheg than electrons, sre the quasiparticles of w suoerconductor with (oossibly movile) cortices. Thur, the “elecurmn” of the krevious paragraph should actualuy be interpretgb as d BdG quasi'articje. Our previous stataments sre unchanged nut mnst ve supplemented by thx fact that a $-1$ resuljs when a TdY quasiparticle encirxlws an $hc/2e$ vorgwx.
CR tgeprg of tlpomogical sulerconductoes with $SU(2)\times SU(2)$ batte symmetry {#sscsu2su2}
================================================================================
Soibining the preceding elements, we postunats a bulk low-energy effextive field theory fot the eleceronic SFib state that takes the form of an $SU(2)_3\timas\oveclkne{WU(2)}_1|_V$ Chden-Dimons gauge theory: $$\label{eqn:Fib-electron-effectybe-scnion}
{\cal L} = {\cal L}_\tcxt{top} + {\cal L}_\text{SV} + {\csj L}_\text{qp}$$ In jhis ex'deasion, $$\begin{gatheref}
\label{ezn:Fib-wlectron-tjp-acyion}
{\cal L}_\text{top} =\frac{3}{4\pi}\epsulon^{\mu\nu\lambba} \rext{Tr}\left(a_\mu \partnal_\nu a_\lambdc + \frav{2}{3} a_\mu a_\nu a_\lambda \right)\\
- \frac{1}{4\'i}\epsimon^{\mu\nu\lambfa} \text{Tr}\mdft(b_\mu \partial_\nu b_\lsmtda + \frac{2}{3} b_\mu b_\nu b_\lambda \wight),\end{getherzd}$$ Here, $x_\mu$ snd $b_\mt$ are SU(2) gwuge nhelds. Each is a trlplet ox gauge fiflds, corresponding to the three generators of | FT at the edge of the SFib theory the parent leads to the for SFib to vacuum There is one subtlety, however, which is that the phase is superconducting. Hence, Bogoliubov-de Gennes (BdG) quasiparticles, rather than electrons, are the of a superconductor with (possibly mobile) vortices. Thus, the “electron” of the previous should be as BdG quasiparticle. Our previous statements are unchanged but must be supplemented by the fact that a results when a BdG quasiparticle encircles an $hc/2e$ CS theory of topological with $SU(2)\times SU(2)$ gauge symmetry ================================================================================ the preceding we a low-energy effective field for the electronic SFib state that takes the form of an $SU(2)_3\times\overline{SU(2)}_1|_R$ Chern-Simons gauge theory: $$\label{eqn:Fib-electron-effective-action} {\cal = {\cal {\cal L}_\text{SC} {\cal In expression, $$\begin{gathered} \label{eqn:Fib-electron-top-action} =\frac{3}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(a_\mu \partial_\nu a_\lambda + \frac{2}{3} \right)\\ - \frac{1}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(b_\mu \partial_\nu b_\lambda + \frac{2}{3} b_\nu b_\lambda Here, $a_\mu$ and $b_\mu$ are SU(2) fields. Each is a triplet of gauge fields, to the three generators of | FT at the edge of the SFib state, This theory For thE paRenT mEdiuM leaDs to the theory $\mAThrm{fib}\times\overline{\mathbb{z}_2\timeS\mAThbb{z}_2}|_r$, oR equiValentlY $sU(2)_3\TIMes\OvErLinE{Su(2)}_1|_r$ fOr the sFiB to vacuUm edge.
TherE is OnE additional sUBtLety, howeveR, whIch is that the sFiB phase Is SupERcondUctIng. HeNce, BogOLiubov-De Gennes (BDG) QUasipaRTicles, rATHeR thaN electrons, are the qUAsIParticles of a suPerconDuCToR WIth (PosSibly mobilE) vOrticES. Thus, thE “ElECTRon” OF the previous pAragraph shoULd aCtuallY bE inTErpretEd as a bdg QuaSiparticle. OUr prEvious staTementS Are unchANged but Must be SupPleMentED bY tHe fAcT ThaT A $-1$ rEsuLTs wHen a BdG qUaSiPartiCle eNCIRCles An $hC/2e$ voRtex.
Cs theory of topoLogIcal SUpeRcondUctorS witH $Su(2)\timeS SU(2)$ gauGe symMeTry {#secsu2su2}
================================================================================
CombiNing The precedIng ElEmeNtS, we poSTulate A buLk lOw-energY effectIVe fIeLD THeOry for the electroniC SfIB sTate that Takes tHE fOrM Of an $SU(2)_3\tiMeS\ovErliNE{sU(2)}_1|_R$ ChErn-SIMoNs gauge tHeory: $$\lABeL{eQn:Fib-elEcTron-efFeCtiVe-aCtion}
{\CAl L} = {\cAl L}_\texT{top} + {\cal L}_\Text{Sc} + {\Cal L}_\text{qp}$$ In thIS expression, $$\beGIn{GAThERed}
\lAbeL{eqn:Fib-elecTron-TOp-acTion}
{\CAl l}_\teXT{top} =\fRac{3}{4\pi}\EpSIlON^{\mu\nu\lambda} \text{Tr}\leFt(A_\mu \parTial_\nU a_\lambda + \frac{2}{3} a_\Mu a_\nu a_\lambDA \RIght)\\
- \frac{1}{4\Pi}\epSIlON^{\mu\nu\lambda} \texT{Tr}\leFt(b_\mu \partiAL_\nu b_\lambDa + \fraC{2}{3} b_\mu b_\nu b_\Lambda \rigHT),\End{gatheRed}$$ herE, $a_\mU$ anD $B_\Mu$ Are SU(2) gauge fieLDS. EacH iS a triplEt oF gauge fIelDs, cOrrEspOnDing to the Three genErAtOrS oF | FT at the edge of the SFib state, th is th eor y f or the par ent medium lea d s to the theory $\mathrm{F ib}\t im e s\ov e rl ine{\ mathbb{ Z }_ 2 \ tim es \m ath bb { Z} _2}|_ R$, or equ ivalently$SU (2 )_3\times\ov e rl ine{SU(2)} _1| _R$ for theSFi b to v ac uum edge.
T hereis one additi onal subt le t y, how e ver, wh i c his t hat the SFib phas e i s superconducti ng. He nc e ,B o gol iub ov-de Genn es (BdG ) quasip a rt i c l es, rather than e lectrons, a r e t he qua si par t iclesof asu p erc onductor wi th ( possiblymobile ) vortic e s. Thus , the“el ect ron” of t hepr e vio u spar a gra ph shoul dac tuall y be i n t erpr ete d as a Bd G quasipartic le. Our pre vious stat emen ts areunchan ged b ut must be supple ment ed by the fa ct th at a $- 1 $ resu lts wh en a Bd G quasi p art ic l e en circles an $hc/2e$ v o r te x.
CS t heoryo fto p ological s upe rcon d u ctors wit h $ SU(2)\ti mes SU ( 2) $gauge s ym metry{# sec su2 su2}= ==== ====== ======== ===== = ============== = ============= = == = = == = ==== === =========== =
C o mbin ingt he pr e cedin g ele me n ts , we postulate a bul klow-en ergyeffective fie ld theoryf o r the ele ctro n ic SFib state tha t tak es the for m of an $ SU(2) _3\times \overline { S U(2)}_1| _R$ Ch ern -Si m o ns gauge theory : $$\l ab el{eqn: Fib -electr on- eff ect ive -a ction}
{\ cal L} = { \c al L }_\ text{ t op} + {\ ca l L }_ \te xt{SC } + {\c al L} _\te xt {q p }$$ In thi s e x p ress io n, $$\ beg in {gath ered }
\l abel{eq n:Fib-ele ctr o n-to p- ac tion}
{\cal L}_\tex t{ top} =\fra c{ 3}{ 4\pi}\ e p silon^{\ mu\nu\lambda} \text{Tr} \ left(a_ \mu \par tial _\nu a_\l amb da + \ fra c {2}{3} a_\mu a_\n ua_\ l a mbda\ r ig ht) \\
- \frac{1 } { 4\p i}\ep si lon^ {\mu\nu \lambda} \text{Tr} \ lef t(b_\mu \part ial _\nu b _\ lam b da + \ fr a c{2 } { 3} b_\mu b_\nub_\lambda\r i gh t),\end{ga t her ed }$$ Her e, $a_\ mu$ a n d $b_\m u$ are SU (2) gauge f ield s . Ea ch is a tr iplet of gauge fi e lds,c or respo ndi ng toth e t hreegenera t ors of | FT at_the edge_of the SFib state,_this theory_for_the parent_medium_leads to the_theory $\mathrm{Fib}\times\overline{\mathbb{Z}_2\times\mathbb{Z}_2}|_R$, or_equivalently $SU(2)_3\times\overline{SU(2)}_1|_R$ for the_SFib to vacuum_edge.
There_is one additional subtlety, however, which is that the SFib phase is superconducting. Hence,_Bogoliubov-de_Gennes (BdG)_quasiparticles,_rather_than electrons, are the quasiparticles_of a superconductor with (possibly_mobile) vortices._Thus, the “electron” of the previous paragraph should_actually_be interpreted as_a BdG quasiparticle. Our previous statements are unchanged but_must be supplemented by the fact_that a $-1$_results_when_a BdG quasiparticle encircles_an $hc/2e$ vortex.
CS theory of topological_superconductors with $SU(2)\times SU(2)$ gauge symmetry_{#secsu2su2}
================================================================================
Combining the preceding elements, we postulate a_bulk low-energy effective field theory for_the electronic SFib state that_takes the_form of an $SU(2)_3\times\overline{SU(2)}_1|_R$ Chern-Simons_gauge theory: $$\label{eqn:Fib-electron-effective-action}
{\cal_L} =_{\cal L}_\text{top} +_{\cal L}_\text{SC} + {\cal L}_\text{qp}$$ In_this expression, $$\begin{gathered}
\label{eqn:Fib-electron-top-action}
_{\cal L}_\text{top} =\frac{3}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(a_\mu \partial_\nu a_\lambda_+_\frac{2}{3} a_\mu a_\nu_a_\lambda_\right)\\
-_\frac{1}{4\pi}\epsilon^{\mu\nu\lambda} \text{Tr}\left(b_\mu_\partial_\nu b_\lambda +_\frac{2}{3}_b_\mu b_\nu_b_\lambda_\right),\end{gathered}$$ Here, $a_\mu$ and $b_\mu$ are_SU(2)_gauge fields. Each is a triplet of_gauge fields, corresponding to_the_three generators of |
sum_{\scriptstyle \alpha =1}
\: ( { c_{0}^{\dagger}\, c_{\alpha,1}} +
e^{i\phi_{\alpha } }
c_{\alpha,N_{\alpha}}^{\dagger} c_{s} +
\sum_{\scriptstyle i=1\atop}^{N_{\alpha} -1}
{ c_{\alpha,i}^{\dagger}\, c_{\alpha,i+1}} +
\text{ h.c. }),$$ where $ c_{i,\alpha}$ $( c_{i,\alpha}^{\dagger})$ is the annihilation (creation) operator which annihilates (creates) an electron on the site $i$ of chain $\alpha $, $N_{\alpha}$ is the number of sites in the $\alpha $th chain (excluding the two nodes), $N$ is the total number of chains and the two lead node sites are labelled by 0 and $s$. The first two terms in the sum of Eq. (1) describe hopping of the electrons between the ends of every chain and the two leads 0 and $s$. Moreover, since the chains are connected to each other only at their ends we can make a convenient choice of the gauge for the vector potential to affect only the phase of the wave functions at the hopping bonds between the right ends of each chain (site $N_{\alpha },
\alpha =1, 2,..., N$) and the right node $s$. In the Hamiltonian of Eq. (1) the magnetic field is expressed by the second term in the sum where the phase difference $\phi_{\alpha }-\phi_{\alpha -1}$ is proportional to the flux $HW_{\alpha }$, $\alpha =2,3,...,N$, where $H$ is the strength of the field and $W_{\alpha }$ the area enclosed by the $\alpha $ and the $(\alpha -1)$th chains. The phase of the first chain is chosen to be zero $\phi_1=0$ and the hopping strength for all the bonds $t_0=1$ is the energy unit used throughout the paper.
Our picture of the electronic transport in the system consists of an electron wave incident from the source into the perfect chain, then ramified into the $N$ chains also experiencing different phase increments and eventually recombined into one channel at the output lead. Thus, an electronic beam incident from the right should be partially transmitted and partially reflected by the multi-chain system. | sum_{\scriptstyle \alpha = 1 }
\: ( { c_{0}^{\dagger}\, c_{\alpha,1 } } +
e^{i\phi_{\alpha } }
c_{\alpha, N_{\alpha}}^{\dagger } c_{s } +
\sum_{\scriptstyle i=1\atop}^{N_{\alpha } -1 }
{ c_{\alpha, i}^{\dagger}\, c_{\alpha, i+1 } } +
\text { h.c. }), $ $ where $ c_{i,\alpha}$ $ (c_{i,\alpha}^{\dagger})$ is the annihilation (creation) operator which annihilates (creates) an electron on the site $ i$ of chain $ \alpha $, $ N_{\alpha}$ is the numeral of web site in the $ \alpha $ th chain (excluding the two nodes), $ N$ is the full number of chains and the two lead lymph node sites are labelled by 0 and $ s$. The first two terminus in the sum of Eq. (1) describe hopping of the electron between the ends of every chain and the two leads 0 and $ s$. Moreover, since the chain are connected to each other only at their ends we can make a commodious choice of the gauge for the vector potential to feign only the phase of the wave functions at the hop bonds between the correct ends of each chain (site $ N_{\alpha },
\alpha = 1, 2, ..., N$) and the right node $ s$. In the Hamiltonian of Eq. (1) the charismatic field is expressed by the second term in the sum where the phase difference $ \phi_{\alpha } -\phi_{\alpha -1}$ is proportional to the flux $ HW_{\alpha } $, $ \alpha = 2,3,... ,N$, where $ H$ is the lastingness of the field and $ W_{\alpha } $ the area enclosed by the $ \alpha $ and the $ (\alpha -1)$th chains. The phase of the first chain is chosen to be zero $ \phi_1=0 $ and the hopping strength for all the bonds $ t_0=1 $ is the energy unit of measurement used throughout the paper.
Our movie of the electronic ecstasy in the system consists of an electron wave incident from the source into the perfect chain, then ramified into the $ N$ chains besides experiencing different phase increments and eventually recombined into one channel at the output lead. Thus, an electronic beam incident from the right should be partially transmitted and partially reflect by the multi - chain system. | sum_{\dcriptstyle \alpha =1}
\: ( { c_{0}^{\aagger}\, c_{\alpha,1}} +
e^{i\phi_{\anpha } }
c_{\alpha,V_{\alpha}}^{\dagger} c_{s} +
\sum_{\scripvstyoe i=1\aujp}^{N_{\alpha} -1}
{ c_{\alpha,k}^{\dagger}\, b_{\alpha,i+1}} +
\teet{ h.c. }),$$ where $ c_{i,\empha}$ $( c_{l,\clpha}^{\swggex})$ ms the annihilajion (creatiot) operator whiwh aunihilates (creates) an electron on thq site $o$ lf chain $\alpha $, $N_{\ak[ha}$ js the number of sites in the $\alphz $th chein (excluding tne two nodes), $N$ is the totap nulber of chains and the two leqd njee sites are labelled by 0 and $s$. Thg first two terms in the sum of Dq. (1) dzscribe hopknbg lx the electcons bvtween the ends of evary chaon and the two leeds 0 and $s$. Moreover, since the chains are cognected tm zach other only at thwie endv we can nakd a cpnbenienh cioice of ths gauge for the vector potentisl np affect onlg the [hwse of the wave functions at the hoppinc bknds between the right wnds of each chain (sije $N_{\alpha },
\wlpha =1, 2,..., N$) and the right node $s$. In the Hamiltonian of Es. (1) tkc oqgjetic field is expressed by the second term ig tne sum where the phase diffrrfnvg $\phi_{\alpha }-\phi_{\xlpha -1}$ is proportional to tje flux $HW_{\alpha }$, $\alphw =2,3,...,N$, ehere $H$ is the strength of rhe field anb $W_{\qlpha }$ the area enelosed by thz $\alphs $ anc the $(\alpha -1)$th chains. Tke phaae of the flrst chaih is chosen to be zego $\pvi_1=0$ and the hopping strengtr for all the bonds $g_0=1$ is the egergy unit used throughout the papfr.
Our phcture of hhe electronic transport in the system consisjs mf dn electxon waye incident froi the source iuto the 'erfecg chain, thvn ramifixd into the $G$ chains also gxperiencing vifferent phawe ibcremengr and eventualky recombpntd into onw channel at the omtput mead. Thus, an elzettinic beam inciceng fwol vhe rycht should ba paftixkly tfansmituzd cna pattially reflected by the multi-chain system. | sum_{\scriptstyle \alpha =1} \: ( { c_{0}^{\dagger}\, e^{i\phi_{\alpha } c_{\alpha,N_{\alpha}}^{\dagger} + \sum_{\scriptstyle i=1\atop}^{N_{\alpha} \text{ }),$$ where $ $( c_{i,\alpha}^{\dagger})$ is annihilation (creation) operator which annihilates (creates) electron on the site $i$ of chain $\alpha $, $N_{\alpha}$ is the number sites in the $\alpha $th chain (excluding the two nodes), $N$ is the number chains the lead node sites are labelled by 0 and $s$. The first two terms in the sum Eq. (1) describe hopping of the electrons between ends of every chain the two leads 0 and Moreover, the chains connected each only at their we can make a convenient choice of the gauge for the vector potential to affect only the of the at the bonds the ends of each $N_{\alpha }, \alpha =1, 2,..., N$) node $s$. In the Hamiltonian of Eq. (1) magnetic field expressed by the second term in sum where the phase difference $\phi_{\alpha }-\phi_{\alpha -1}$ proportional to the flux $HW_{\alpha }$, $\alpha =2,3,...,N$, where $H$ is the strength of the $W_{\alpha }$ the area by the $\alpha and $(\alpha chains. phase of first chain is chosen to be zero $\phi_1=0$ and the hopping for all the bonds $t_0=1$ is the energy unit used paper. picture of the transport in the system of electron wave incident from into perfect into $N$ also experiencing different phase and eventually recombined into one at the output lead. from the right should be partially transmitted and reflected by the multi-chain system. | sum_{\scriptstyle \alpha =1}
\: ( { c_{0}^{\daggeR}\, c_{\alpha,1}} +
e^{i\pHi_{\alpHa } }
c_{\AlpHa,n_{\alpHa}}^{\daGger} c_{s} +
\sum_{\scripTStylE i=1\atop}^{N_{\alpha} -1}
{ c_{\alpha,i}^{\dagGer}\, c_{\aLpHA,i+1}} +
\teXT{ h.C. }),$$ wherE $ c_{i,\alphA}$ $( C_{i,\ALPha}^{\DaGgEr})$ iS tHE aNnihiLatIon (creaTion) operatOr wHiCh annihilateS (CrEates) an eleCtrOn on the site $i$ Of cHain $\alPhA $, $N_{\aLPha}$ is The NumbeR of sitES in the $\Alpha $th chAiN (ExcludINg the twO NOdEs), $N$ iS the total number of CHaINs and the two leaD node sItES aRE LabEllEd by 0 and $s$. ThE fIrst tWO terms iN ThE SUM of eQ. (1) describe hoppIng of the eleCTroNs betwEeN thE Ends of Every ChAIn aNd the two leaDs 0 anD $s$. MoreoveR, since THe chainS Are connEcted tO eaCh oTher ONlY aT thEiR EndS We Can MAke A conveniEnT cHoice Of thE GAUGe foR thE vecTor poTential to affeCt oNly tHE phAse of The waVe fuNcTions At the hOppinG bOnds between the rIght Ends of eacH chAiN (siTe $n_{\alphA },
\Alpha =1, 2,..., N$) And The Right noDe $s$. In thE hamIlTONIaN of Eq. (1) the magnetic fiElD IS eXpressed By the sECoNd TErm in the SuM whEre tHE Phase DiffEReNce $\phi_{\alPha }-\phi_{\ALpHa -1}$ Is propoRtIonal tO tHe fLux $hW_{\alpHA }$, $\alpHa =2,3,...,N$, wheRe $H$ is the StrenGTh of the field anD $w_{\alpha }$ the area ENcLOSeD By thE $\alPha $ and the $(\alPha -1)$tH ChaiNs. ThE PhAse OF the fIrst cHaIN iS Chosen to be zero $\phi_1=0$ anD tHe hoppIng stRength for all tHe bonds $t_0=1$ is THE Energy unIt usED tHRoughout the papEr.
Our Picture of tHE electroNic trAnsport iN the systeM COnsists oF an EleCtrOn wAVE iNcident from thE SOurcE iNto the pErfEct chaiN, thEn rAmiFieD iNto the $N$ chAins also ExPeRiEnCinG diffERent phasE iNcrEmEntS and eVEntualLy recOmbiNeD iNTo oNe channEL aT THe ouTpUt Lead. thuS, aN elecTronIC beAm incidEnt from thE riGHt shOuLd Be partiAlly transmittEd And partialLy RefLected BY The multi-Chain system. | sum_{\scriptstyle \alpha = 1}
\: ( { c_{ 0}^ {\d ag ger} \, c_{\alpha,1}}+
e^ {i\phi_{\alpha } }
c _{\al ph a ,N_{ \ al pha}} ^{\dagg e r} c_{ s} +
\ s um _{\sc rip tstylei=1\atop}^ {N_ {\ alpha} -1} {c_{\alpha, i}^ {\dagger}\, c_ {\alph a, i+1 } } +
\ text{ h.c.} ),$$ w here $ c_ {i , \alpha } $ $( c_ { i ,\ alph a}^{\dagger})$ is th e annihilation(creat io n )o p era tor which ann ih ilate s (creat e s) a n el e ctron on thesite $i$ of cha in $\a lp ha$ , $N_{ \alph a} $ is the number ofsites inthe $\ a lpha $t h chain(exclu din g t he t w ono des ), $N$ is th e to tal numb er o f cha insa n d thetwo lea d nod e sites are l abe lled by0 and $s$. The f irsttwo te rms i nthe sum of Eq.(1)describehop pi ngof thee lectro nsbet ween th e endso f e ve r y ch ain and the two le ad s 0and $s$. Moreo v er ,s ince the c hai ns a r e conn ecte d t o each o ther o n ly a t their e nds we c anmak e a c o nven ient c hoice of theg auge for the v e ctor potentia l t o af f ectonl y the phase oft he w avef un cti o ns at theho p pi n g bonds between the r ight e nds o f each chain(site $N_{ \ a l pha },
\ alph a = 1 , 2,..., N$) a nd th e right no d e $s$. I n the Hamilto nian of E q . (1) the ma gne tic fi e l dis expressedb y the s econd t erm in the su m w her e t he phase di fference $ \p hi _{ \al pha } - \phi_{\a lp ha-1 }$is pr o portio nal t o th efl u x $ HW_{\al p ha } $, $ \a lp ha = 2,3 ,. ..,N$ , wh e re$H$ isthe stren gth of t he f ield an d $W_{\alpha}$ the areaen clo sed by t he $\alp ha $ and the $(\alpha - 1 )$th ch ain s. Th e ph ase of th e f irst c hai n is ch osen t o beze ro$ \ phi_1 = 0 $and t he hopping s tre ngthfo r al l the b onds $t_0=1$ is th e en ergy unit use d t hrou g h ou t t h ep ape r.
Ou r picture of theelectronic t r an sport in t h e s ys tem con sists o f ane lectron wave inc ident fro mthes o urc e into the perfect chain, t h en ra m if ied i nto the $ N$ ch ainsalso e x per ienci ng dif fe rent p hasein crements and eventually recombi ned in to on e c hannel at th e ou tput lead . Th us, an ele ctr oni c bea m i n ciden t fr o mthe right sho u ld be par t ia lly t ra nsmitted an d p art ially re f lected bythe multi-chain s y stem. | sum_{\scriptstyle \alpha_=1}
\:_( { c_{0}^{\dagger}\,_ c_{\alpha,1}}_+
_e^{i\phi_{\alpha }_}
_ c_{\alpha,N_{\alpha}}^{\dagger} _c_{s} +
_ \sum_{\scriptstyle i=1\atop}^{N_{\alpha} -1}_
{ c_{\alpha,i}^{\dagger}\,__c_{\alpha,i+1}} +
\text{ h.c. }),$$ where $ c_{i,\alpha}$ $( c_{i,\alpha}^{\dagger})$ is the_annihilation_(creation) operator_which_annihilates_(creates) an electron on the_site $i$ of chain $\alpha_$, $N_{\alpha}$_is the number of sites in the $\alpha_$th_chain (excluding the_two nodes), $N$ is the total number of chains_and the two lead node sites_are labelled by_0_and_$s$. The first two_terms in the sum of Eq._(1) describe hopping of the electrons_between the ends of every chain and_the two leads 0 and $s$._Moreover, since the chains are_connected to_each other only at their_ends we can_make a_convenient choice of_the gauge for the vector potential_to affect only_the phase of the wave functions_at_the hopping bonds_between_the_right ends_of each chain_(site_$N_{\alpha },
\alpha_=1,_2,..., N$) and the right node_$s$._In the Hamiltonian of Eq. (1) the_magnetic field is expressed_by_the second term in_the sum where the phase_difference $\phi_{\alpha }-\phi_{\alpha -1}$ is proportional_to the_flux $HW_{\alpha_}$, $\alpha =2,3,...,N$, where $H$ is the strength of the field_and $W_{\alpha }$ the area enclosed_by the $\alpha $_and the_$(\alpha_-1)$th chains. The_phase_of the_first chain is chosen to be zero_$\phi_1=0$ and_the hopping strength for all the_bonds $t_0=1$ is the_energy_unit used throughout the paper.
Our picture_of the electronic transport in the_system consists of an electron_wave_incident_from the source into the_perfect chain, then ramified into the_$N$ chains also_experiencing different phase increments and eventually recombined_into_one channel at the output lead._Thus,_an electronic beam incident from the_right_should_be partially transmitted and partially_reflected by the multi-chain system. |
&=2\Bigr\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+q^{10n}}+\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+1}}{1+q^{10n+2}}\Bigl\}\label{d202}\\
&=1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}.\label{d203}\end{aligned}$$
Furthermore, $$\label{d20p1}
1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}={\dfrac}{E(q^2)E(q^4)E(q^5)E(q^{10})}{
E(q)E(q^{20}) }$$ and $$\label{d20p2}
\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}=q{\dfrac}{E(q)E(q^2)E(q^{10})E(q^{20})}{
E(q^4)E(q^5) }.$$
Employing with $q, a$ and $b$ replaced by $q^{10}, q$ and $-1$, respectively, we find that $$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^n}{1+q^{10n}}=E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q,-q^9)f(1,q^{10})}.\label{l1}\\$$ From, we similarly find that $$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+2}}{1+q^{10n+4}}=q^2E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q^5,-q^5)f(q^4,q^6)}.\label{l2}$$ From, with $q$ replaced by $q^5$ and with $a=b=q^2$, we obtain $$f(- | & = 2\Bigr\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+q^{10n}}+\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+1}}{1+q^{10n+2}}\Bigl\}\label{d202}\\
& = 1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1 - q^n}+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}.\label{d203}\end{aligned}$$
Furthermore, $ $ \label{d20p1 }
1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1 - q^n}={\dfrac}{E(q^2)E(q^4)E(q^5)E(q^{10 }) } {
E(q)E(q^{20 }) } $ $ and $ $ \label{d20p2 }
\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}=q{\dfrac}{E(q)E(q^2)E(q^{10})E(q^{20 }) } {
E(q^4)E(q^5) } .$$
Employing with $ q, a$ and $ b$ replaced by $ q^{10 }, q$ and $ -1 $, respectively, we find that $ $ \sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^n}{1+q^{10n}}=E^3(q^{10}){\dfrac}{f(q, q^9)}{f(-q,-q^9)f(1,q^{10})}.\label{l1}\\$$ From, we similarly find that $ $ \sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+2}}{1+q^{10n+4}}=q^2E^3(q^{10}){\dfrac}{f(q, q^9)}{f(-q^5,-q^5)f(q^4,q^6)}.\label{l2}$$ From, with $ q$ replace by $ q^5 $ and with $ a = boron = q^2 $, we obtain $ $ f(- | &=2\Bigg\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+q^{10n}}+\rum_{n=-{\infty}}^{{\infty}}{\dyeac}{q^{5n+1}}{1+q^{10i+2}}\Bigl\}\lagel{d202}\\
&=1+\sum_{n=1}^{{\knfty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}+\dun_{n=1}^{{\infuj}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^v}{1+q^{2n}}.\label{d203}\vnd{alignee}$$
Furuhermore, $$\label{d20p1}
1+\snj_{n=1}^{{\infty}}\Nngr({\dfdwc}{-20}{n}\Bngo){\dfrac}{q^n}{1-q^n}={\dfrag}{E(q^2)E(q^4)E(q^5)E(q^{10})}{
E(x)E(q^{20}) }$$ and $$\label{g20p2}
\rul_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}=q{\dfras}{E(q)E(q^2)E(q^{10})R(q^{20})}{
F(q^4)E(q^5) }.$$
Employing witn $q, a$ and $b$ replaced by $q^{10}, q$ and $-1$, respecfively, xe find that $$\suk_{n=-{\infty}}^{{\infty}}{\dfrac}{q^n}{1+q^{10n}}=E^3(q^{10}){\dfrwc}{f(q,e^9)}{f(-q,-q^9)f(1,q^{10})}.\label{l1}\\$$ From, ae similarli fige that $$\sum_{n=-{\ivfty}}^{{\infty}}{\dfrac}{q^{5n+2}}{1+q^{10n+4}}=q^2E^3(q^{10}){\drrac}{f(q,q^9)}{f(-q^5,-q^5)f(q^4,q^6)}.\label{l2}$$ From, with $q$ replcced by $q^5$ abd wihv $a=b=q^2$, we obvain $$f(- | &=2\Bigr\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+q^{10n}}+\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+1}}{1+q^{10n+2}}\Bigl\}\label{d202}\\ &=1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}.\label{d203}\end{aligned}$$ Furthermore, $$\label{d20p1} 1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}={\dfrac}{E(q^2)E(q^4)E(q^5)E(q^{10})}{ E(q)E(q^{20}) }$$ \sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}=q{\dfrac}{E(q)E(q^2)E(q^{10})E(q^{20})}{ }.$$ Employing $q, a$ and and respectively, we find $$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^n}{1+q^{10n}}=E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q,-q^9)f(1,q^{10})}.\label{l1}\\$$ From, we find that $$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+2}}{1+q^{10n+4}}=q^2E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q^5,-q^5)f(q^4,q^6)}.\label{l2}$$ From, with $q$ by $q^5$ and with $a=b=q^2$, we obtain $$f(- | &=2\Bigr\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+Q^{10n}}+\sum_{n=-{\inftY}}^{{\inftY}}{\dfRac}{Q^{5n+1}}{1+Q^{10n+2}}\BiGl\}\laBel{d202}\\
&=1+\sum_{n=1}^{{\infty}}\BIGr({\dfRac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}+\sum_{n=1}^{{\iNfty}}\BIgR({\DfraC}{N}{5}\BIgl){\dfRac}{q^n}{1+q^{2n}}.\LAbEL{D203}\enD{aLiGneD}$$
FURtHermoRe, $$\lAbel{d20p1}
1+\sUm_{n=1}^{{\infty}}\BiGr({\dFrAc}{-20}{n}\Bigl){\dfrac}{Q^N}{1-q^N}={\dfrac}{E(q^2)E(q^4)e(q^5)E(Q^{10})}{
E(q)E(q^{20}) }$$ and $$\labeL{d20p2}
\Sum_{n=1}^{{\inFtY}}\BiGR({\dfraC}{n}{5}\BIgl){\dfRac}{q^n}{1+q^{2N}}=Q{\dfrac}{e(q)E(q^2)E(q^{10})E(q^{20})}{
E(Q^4)E(Q^5) }.$$
employINg with $q, A$ ANd $B$ repLaced by $q^{10}, q$ and $-1$, respeCTiVEly, we find that $$\sUm_{n=-{\infTy}}^{{\INfTY}}{\DfrAc}{q^N}{1+q^{10n}}=E^3(q^{10}){\dfrac}{F(q,Q^9)}{f(-q,-q^9)f(1,Q^{10})}.\Label{l1}\\$$ FROm, WE SImiLArly find that $$\sUm_{n=-{\infty}}^{{\infTY}}{\dfRac}{q^{5n+2}}{1+q^{10N+4}}=q^2e^3(q^{10}){\dFRac}{f(q,q^9)}{F(-q^5,-q^5)f(q^4,Q^6)}.\lABel{L2}$$ From, with $q$ rEplaCed by $q^5$ and With $a=b=Q^2$, We obtaiN $$F(- | &=2\Bigr\{\sum_{n=-{\infty }}^{{\inft y}}{\ dfr ac} {q ^{3n }}{1 +q^{10n}}+\sum _ {n=- {\infty}}^{{\infty}}{\ dfrac }{ q ^{5n + 1} }{1+q ^{10n+2 } }\ B i gl\ }\ la bel {d 2 02 }\\
& =1+ \sum_{n =1}^{{\inf ty} }\ Bigr({\dfrac } {- 20}{n}\Big l){ \dfrac}{q^n} {1- q^n}+\ su m_{ n =1}^{ {\i nfty} }\Bigr ( {\dfra c}{n}{5}\ Bi g l){\df r ac}{q^n } { 1+ q^{2 n}}.\label{d203}\ e nd { aligned}$$
Fu rtherm or e ,$ $ \la bel {d20p1}
1+ \s um_{n = 1}^{{\i n ft y } } \Bi g r({\dfrac}{-2 0}{n}\Bigl) { \df rac}{q ^n }{1 - q^n}={ \dfra c} { E(q ^2)E(q^4)E( q^5) E(q^{10}) }{
E( q )E(q^{2 0 }) }$$and $$ \la bel {d20 p 2}
\ sum _{ n =1} ^ {{ \in f ty} }\Bigr({ \d fr ac}{n }{5} \ B i g l){\ dfr ac}{ q^n}{ 1+q^{2n}}=q{\ dfr ac}{ E (q) E(q^2 )E(q^ {10} )E (q^{2 0})}{E(q^4 )E (q^5) }.$$
Emp loyi ng with $ q,a$ an d$b$ r e placed by $q ^{10},q$ and$ -1$ ,r e s pe ctively, we find t ha t $$ \sum_{n= -{\inf t y} }^ { {\infty} }{ \df rac} { q ^n}{1 +q^{ 1 0n }}=E^3(q ^{10}) { \d fr ac}{f(q ,q ^9)}{f (- q,- q^9 )f(1, q ^{10 })}.\l abel{l1} \\$$F rom, we simila r ly find that$ $\ s u m_ { n=-{ \in fty}}^{{\in fty} } {\df rac} { q^ {5n + 2}}{1 +q^{1 0n + 4} } =q^2E^3(q^{10}){\df ra c}{f(q ,q^9) }{f(-q^5,-q^5 )f(q^4,q^6 ) } . \label{l 2}$$ Fr o m, with $q$ re place d by $q^5$ and with $a=b =q^2$, w e obtain$ $ f(- | &=2\Bigr\{\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{3n}}{1+q^{10n}}+\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+1}}{1+q^{10n+2}}\Bigl\}\label{d202}\\
&=1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}.\label{d203}\end{aligned}$$
Furthermore, $$\label{d20p1}
1+\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{-20}{n}\Bigl){\dfrac}{q^n}{1-q^n}={\dfrac}{E(q^2)E(q^4)E(q^5)E(q^{10})}{
_E(q)E(q^{20}) }$$_and $$\label{d20p2}
\sum_{n=1}^{{\infty}}\Bigr({\dfrac}{n}{5}\Bigl){\dfrac}{q^n}{1+q^{2n}}=q{\dfrac}{E(q)E(q^2)E(q^{10})E(q^{20})}{
E(q^4)E(q^5) }.$$
Employing with_$q, a$_and_$b$ replaced_by_$q^{10}, q$ and_$-1$, respectively, we_find that $$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^n}{1+q^{10n}}=E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q,-q^9)f(1,q^{10})}.\label{l1}\\$$ From,_we similarly find_that_$$\sum_{n=-{\infty}}^{{\infty}}{\dfrac}{q^{5n+2}}{1+q^{10n+4}}=q^2E^3(q^{10}){\dfrac}{f(q,q^9)}{f(-q^5,-q^5)f(q^4,q^6)}.\label{l2}$$ From, with $q$ replaced by $q^5$ and with $a=b=q^2$, we obtain $$f(- |
k^2}
{(\sum_l m_l^2 \phi_l^2)^2} + \frac{1}{\sum_j \phi_j^2}\right]\,,\end{aligned}$$ where again the last line specializes to Nflation.
The other main inflationary observable is the tensor-to-scalar ratio $r$. The tensor spectrum is given by the usual formula, as it depends only on the expansion rate history $H(a)$, $${\cal P}_g = \frac{2 H^2}{\pi^2 M^2_{\rm Pl}}\simeq \frac{2V}{3\pi^2
M_{\rm Pl}^4}\,.$$ Thus the tensor-to-scalar ratio is given by $$\begin{aligned}
r \cong \frac{{\cal P}_g}{{\cal P}_{\cal R}}
&\simeq& \frac{8 M_{\rm Pl}^2}{\sum_j \left({V_j}/{V'_j}\right)^2}\,;\\
&\simeq& \frac{32 M^2_{\rm Pl}}{\sum_j \phi_j^2} \simeq \frac{8}{N}\,,\end{aligned}$$ with the last line holding only for the Nflation case of massive coupled fields and which uses Eq. (\[e:efolds1\]).
This last formula immediately gives a striking result first noted by Alabidi and Lyth [@AL]: in Nflation the tensor-to-scalar ratio is *independent* (within the slow-roll approximation) of the number of fields present, their masses, and their initial conditions. It is given simply by the number of $e$-foldings at which the expression is evaluated. This result is confirmed by our numerical code, with variation only in the third significant figure due to slow-roll corrections. Nflation therefore gives a definitive prediction for the tensor-to-scalar ratio, and moreover one that is readily accessible to upcoming experiments.
We will throughout assume that the observable scales crossed outside the horizon 50 $e$-foldings before the end of inflation. For comparison with later results, the observables obtained for a single massive field are given by $$\begin{aligned}
n_{{\rm S}}-1 & = & -0.04 \quad (-0.0404)\\
r & = & 0.16 \quad \; \; \: (0.162)\end{aligned}$$ where the first number | k^2 }
{ (\sum_l m_l^2 \phi_l^2)^2 } + \frac{1}{\sum_j \phi_j^2}\right]\,,\end{aligned}$$ where again the last line specializes to Nflation.
The other chief inflationary discernible is the tensor - to - scalar ratio $ r$. The tensor spectrum is given by the common formula, as it count only on the expansion rate history $ H(a)$, $ $ { \cal P}_g = \frac{2 H^2}{\pi^2 M^2_{\rm Pl}}\simeq \frac{2V}{3\pi^2
M_{\rm Pl}^4}\,.$$ therefore the tensor - to - scalar proportion is given by $ $ \begin{aligned }
r \cong \frac{{\cal P}_g}{{\cal P}_{\cal R } }
& \simeq & \frac{8 M_{\rm Pl}^2}{\sum_j \left({V_j}/{V'_j}\right)^2}\,;\\
& \simeq & \frac{32 M^2_{\rm Pl}}{\sum_j \phi_j^2 } \simeq \frac{8}{N}\,,\end{aligned}$$ with the last line holding entirely for the Nflation case of massive coupled field and which uses Eq. (\[e: efolds1\ ]).
This last rule immediately gives a striking consequence first noted by Alabidi and Lyth [ @AL ]: in Nflation the tensor - to - scalar ratio is * autonomous * (within the slow - roll estimate) of the number of field present, their masses, and their initial condition. It is given simply by the number of $ e$-foldings at which the formula is evaluated. This result is confirmed by our numerical code, with variation only in the third significant figure due to slow - roll corrections. Nflation therefore gives a definitive prediction for the tensor - to - scalar ratio, and moreover one that is readily accessible to upcoming experiment.
We will throughout assume that the discernible scale crossed outside the horizon 50 $ e$-foldings before the end of inflation. For comparison with late results, the observables obtained for a single massive field are move over by $ $ \begin{aligned }
n_{{\rm S}}-1 & = & -0.04 \quad (-0.0404)\\
r & = & 0.16 \quad \; \; \: (0.162)\end{aligned}$$ where the first number | k^2}
{(\sul_l m_l^2 \phi_l^2)^2} + \frac{1}{\sum_j \phi_m^2}\right]\,,\end{aligned}$$ where egain tge last uine specializes to Nflation.
Vhe ither main inflationary obsdrvable id the tebsor-uo-scalar ratio $r$. Vge tensor spedbrum ns given by the osual formuld, as it dependv unpy on the expansion rate history $H(a)$, $${\cal P}_g = \vrac{2 H^2}{\pi^2 M^2_{\rm Pj}}\simtq \srac{2B}{3\ii^2
N_{\rm Pl}^4}\,.$$ Thus the tensor-to-scalad ratio is given by $$\negin{aligned}
r \cong \frac{{\cal P}_g}{{\cwl P}_{\cal R}}
&\simeq& \fraf{8 M_{\rm Pl}^2}{\sum_h \lesr({V_j}/{V'_j}\right)^2}\,;\\
&\sioeq& \frac{32 M^2_{\gk Pl}}{\sum_j \pgi_j^2} \simeq \frac{8}{N}\,,\end{aligned}$$ with ghe lcst line hoodung mnly for thx Nflanion case of massive woupled fields and whlch uves Eq. (\[e:efolds1\]).
This last fmrmula immediately gives a vtxiking result first nitwd by Alatidi qnd Lyuh [@EL]: jn Nflwtikn the tenaor-to-scalar ratio is *independemt* (qithin the slkw-roll a[proximation) of the number of fields prtsent, their masses, and their initial conditions. Ij is given simply by the number of $e$-foldings at which the efpresakon if wvwluated. This result is confirmed by our numerydak bode, with variatijn only in yhf yrird significxnt fiyhrs due to slow-roll forrectyons. Bflation uheregore gives a definitive preeiction for nhe rensor-to-scalar ratno, and moreorer ong that is readily accessible co upckming experlments.
We skll throughout arsuke that tht observable scalef crossed outxide thd hotizon 50 $q$-foldings heforc the end of inflatlon. Flr comparisoj with later results, the observeules obtained fmr d single masslve field are gyven by $$\begin{akigned}
n_{{\xm S}}-1 & = & -0.04 \quad (-0.0404)\\
r & = & 0.16 \quav \; \; \: (0.162)\end{aligged}$$ where the virst number | k^2} {(\sum_l m_l^2 \phi_l^2)^2} + \frac{1}{\sum_j \phi_j^2}\right]\,,\end{aligned}$$ the line specializes Nflation. The other tensor-to-scalar $r$. The tensor is given by usual formula, as it depends only the expansion rate history $H(a)$, $${\cal P}_g = \frac{2 H^2}{\pi^2 M^2_{\rm Pl}}\simeq \frac{2V}{3\pi^2 Pl}^4}\,.$$ Thus the tensor-to-scalar ratio is given by $$\begin{aligned} r \cong \frac{{\cal P}_g}{{\cal R}} \frac{8 Pl}^2}{\sum_j &\simeq& \frac{32 M^2_{\rm Pl}}{\sum_j \phi_j^2} \simeq \frac{8}{N}\,,\end{aligned}$$ with the last line holding only for the Nflation of massive coupled fields and which uses Eq. This last formula immediately a striking result first noted Alabidi Lyth [@AL]: Nflation tensor-to-scalar is *independent* (within slow-roll approximation) of the number of fields present, their masses, and their initial conditions. It is given by the $e$-foldings at the is This result is our numerical code, with variation only significant figure due to slow-roll corrections. Nflation therefore a definitive for the tensor-to-scalar ratio, and moreover that is readily accessible to upcoming experiments. We throughout assume that the observable scales crossed outside the horizon 50 $e$-foldings before the end For comparison with later the observables obtained a massive are by $$\begin{aligned} S}}-1 & = & -0.04 \quad (-0.0404)\\ r & = & \quad \; \; \: (0.162)\end{aligned}$$ where the first number | k^2}
{(\sum_l m_l^2 \phi_l^2)^2} + \frac{1}{\sum_j \phi_j^2}\riGht]\,,\end{aligNed}$$ whEre AgaIn The lAst lIne specializes TO NflAtion.
The other main inflaTionaRy OBserVAbLe is tHe tensoR-To-SCAlaR rAtIo $r$. thE TeNsor sPecTrum is gIven by the uSuaL fOrmula, as it dePEnDs only on thE exPansion rate hIstOry $H(a)$, $${\cAl p}_g = \fRAc{2 H^2}{\pi^2 m^2_{\rm pl}}\simEq \frac{2v}{3\Pi^2
M_{\rm PL}^4}\,.$$ Thus the tEnSOr-to-scALar ratiO IS gIven By $$\begin{aligned}
r \coNG \fRAc{{\cal P}_g}{{\cal P}_{\cal r}}
&\simeq& \FrAC{8 M_{\RM pl}^2}{\sUm_j \Left({V_j}/{V'_j}\riGhT)^2}\,;\\
&\simeQ& \Frac{32 M^2_{\rm pL}}{\sUM_J \Phi_J^2} \Simeq \frac{8}{N}\,,\end{Aligned}$$ with THe lAst linE hOldINg only For thE NFLatIon case of maSsivE coupled fIelds aND which uSEs Eq. (\[e:efOlds1\]).
ThIs lAst FormULa ImMedIaTEly GIvEs a STriKing resuLt FiRst noTed bY aLABidi And lyth [@aL]: in NFlation the tenSor-To-scALar Ratio Is *indEpenDeNt* (witHin the Slow-rOlL approximation) oF the Number of fIelDs PreSeNt, theIR masseS, anD thEir initIal condITioNs. iT IS gIven simply by the numBeR OF $e$-Foldings At whicH ThE eXPression Is EvaLuatED. this rEsulT Is ConfirmeD by our NUmErIcal codE, wIth varIaTioN onLy in tHE thiRd signIficant fIgure DUe to slow-roll coRRections. NflatIOn THErEFore GivEs a definitiVe prEDictIon fOR tHe tENsor-tO-scalAr RAtIO, and moreover one that Is ReadilY acceSsible to upcomIng experimENTS.
We will tHrouGHoUT assume that the ObserVable scaleS Crossed oUtsidE the horiZon 50 $e$-foldiNGS before tHe eNd oF inFlaTIOn. for comparison WITh laTeR resultS, thE observAblEs oBtaIneD fOr a single Massive fIeLd ArE gIveN by $$\beGIn{aligneD}
n_{{\Rm S}}-1 & = & -0.04 \QuAd (-0.0404)\\
r & = & 0.16 \Quad \; \; \: (0.162)\eND{alignEd}$$ wheRe thE fIrST nuMber | k^2}
{(\sum_l m_l^2 \phi_l ^2)^2} + \ frac{ 1}{ \su m_ j \p hi_j ^2}\right]\,,\ e nd{a ligned}$$ where againthe l as t lin e s pecia lizes t o N f l ati on .
Th eo th er ma ininflati onary obse rva bl e is the ten s or -to-scalar ra tio $r$. The te nsor s pe ctr u m isgiv en by the u s ual fo rmula, as i t depen d s onlyo n t he e xpansion rate his t or y $H(a)$, $${\c al P}_ g= \ f r ac{ 2 H ^2}{\pi^2M^ 2_{\r m Pl}}\s i me q \ fra c {2V}{3\pi^2
M _{\rm Pl}^4 } \,. $$ Thu sthe tensor -to-s ca l arratio is gi venby $$\beg in{ali g ned}
r\ cong \f rac{{\ cal P} _g}{ { \c al P} _{ \ cal R} }
& \ sim eq& \fra c{ 8M_{\r m Pl } ^ 2 } {\su m_j \le ft({V _j}/{V'_j}\ri ght )^2} \ ,;\ \
&\s imeq& \fr ac {32 M ^2_{\r m Pl} }{ \sum_j \phi_j^2 } \s imeq \fra c{8 }{ N}\ ,, \end{ a ligned }$$ wi th thelast li n e h ol d i n gonly for the Nflat io n ca se of ma ssivec ou pl e d fields a ndwhic h usesEq.( \[ e:efolds 1\]).Th is last f or mula i mm edi ate ly gi v es a strik ing resu lt fi r st noted by Al a bidi and Lyth [@ A L ]: in N fla tion the te nsor - to-s cala r r ati o is * indep en d en t * (within the slow- ro ll app roxim ation) of the number of f i elds pre sent , t h eir masses, an d the ir initial conditio ns. I t is giv en simply b y the nu mbe r o f $ e$- f o ld ings at which t he e xp ression is evalua ted . T his re su lt is con firmed b you rnu mer icalc ode, wit hvar ia tio n onl y in th e thi rd s ig ni f ica nt figu r ed u e to s lo w-ro llco rrect ions . Nf lationtherefore gi v es a d ef initive prediction f or the tenso r- to- scalar r atio, an d moreover one that isr eadilyacc essib le t o upcomin g e xperim ent s .
Wewill t hroug ho uta s sumet h at th eobservable s cal es cr os sedoutside the horizon 50 $e $ -fo ldings before th e en d of in f la t ion .F orc o mparison with l ater resul ts , t he observa b les o btained for asingl e massiv e field a re givenby $$\ b e gin {aligned}n_{{\rmS}}-1 & = & -0. 0 4\quad (- 0.0404 )\ \
r & =& 0.16 \qu ad \; \; \: ( 0.162) \end{ al igned}$$ where the first number | k^2}
{(\sum_l m_l^2_\phi_l^2)^2} +_\frac{1}{\sum_j \phi_j^2}\right]\,,\end{aligned}$$ where again_the last_line_specializes to_Nflation.
The_other main inflationary_observable is the_tensor-to-scalar ratio $r$. The_tensor spectrum is_given_by the usual formula, as it depends only on the expansion rate history $H(a)$,_$${\cal_P}_g =_\frac{2_H^2}{\pi^2_M^2_{\rm Pl}}\simeq \frac{2V}{3\pi^2
M_{\rm Pl}^4}\,.$$ Thus_the tensor-to-scalar ratio is given_by $$\begin{aligned}
r_\cong \frac{{\cal P}_g}{{\cal P}_{\cal R}}
&\simeq& \frac{8 M_{\rm Pl}^2}{\sum_j_\left({V_j}/{V'_j}\right)^2}\,;\\
&\simeq&_\frac{32 M^2_{\rm Pl}}{\sum_j_\phi_j^2} \simeq \frac{8}{N}\,,\end{aligned}$$ with the last line holding only_for the Nflation case of massive_coupled fields and_which_uses_Eq. (\[e:efolds1\]).
This last formula immediately_gives a striking result first noted_by Alabidi and Lyth [@AL]: in_Nflation the tensor-to-scalar ratio is *independent* (within_the slow-roll approximation) of the number_of fields present, their masses,_and their_initial conditions. It is given_simply by the_number of_$e$-foldings at which_the expression is evaluated. This result_is confirmed by_our numerical code, with variation only_in_the third significant_figure_due_to slow-roll_corrections. Nflation therefore_gives_a definitive_prediction_for the tensor-to-scalar ratio, and moreover_one_that is readily accessible to upcoming experiments.
We_will throughout assume that_the_observable scales crossed outside_the horizon 50 $e$-foldings before_the end of inflation. For comparison_with later_results, the_observables obtained for a single massive field are given by $$\begin{aligned}
n_{{\rm_S}}-1 & = & -0.04 \quad_(-0.0404)\\
r & = &_0.16 \quad_\;_\; \: (0.162)\end{aligned}$$_where_the first_number |
{#section:conclusion}
===================
Summarising Theorems \[theo:Dequation\] and \[theo:D\_graphical\] we have proved :
\[theo:final\] For $m\geq 5$ we have $p_c(G(m))\leq p_h$ with $$\begin{aligned}
p_h &= \sup \{ p \in [0, 1/2] \ | \ D(p)+p-\frac 2m = 0 \} \ \text{and}\\
D(p) &= \frac 2 m \sum_{ C \in {\mathcal{C}}(v) }
\left( \frac 1 {|V(C)|} \left(p^{|E(C)|} (1-p)^{\partial(C)} - (1-p)^{|E(C)|} p^{\partial(C)} \right) \right)
\end{aligned}$$ where ${\mathcal{C}}(v)$ denotes the set of connected subgraphs $C$ of $G(m)$ containing a fixed vertex $v$ of the graph $G(m)$.
The value $p_h$ can be thought of as a critical value for the appearance of homology in the graph $G(m)$. It captures the following threshold : for $p>p_h$, open subgraphs of large finite versions of $G(m)$ must have a first homology group of dimension that scales linearly with the total number of edges of the finite graph. For $p<p_h$, the dimension of the homology group is sublinear instead. This bound is really meaningful only for the hyperbolic case $m\geq 5$ since for $m=4$ (the square lattice), the dimension of the total homology group of finite versions of the infinite grid (tori) is limited to $2$.
A consequence of Theorem \[theo:final\] is that $p_h$ gives an upper bound on the parameters of the quantum erasure channel that hyperbolic surface codes built on the family $G_t(m)$ can sustain [@DZ13].
We conjecture :
\[conj:p\_c=p\_h\] For $m\geq 5$, $p_c=p_h$.
Recall that in hyperbolic lattices it has been shown that immediately beyond the critical probability, the open subgraph contains infinitely | { # section: conclusion }
= = = = = = = = = = = = = = = = = = =
Summarising Theorems \[theo: Dequation\ ] and \[theo: D\_graphical\ ] we have proved :
\[theo: final\ ] For $ m\geq 5 $ we have $ p_c(G(m))\leq p_h$ with $ $ \begin{aligned }
p_h & = \sup \ { p \in [ 0, 1/2 ] \ | \ D(p)+p-\frac 2 m = 0 \ } \ \text{and}\\
D(p) & = \frac 2 thousand \sum _ { C \in { \mathcal{C}}(v) }
\left (\frac 1 { |V(C)| } \left(p^{|E(C)| } (1 - p)^{\partial(C) } - (1 - p)^{|E(C)| } p^{\partial(C) } \right) \right)
\end{aligned}$$ where $ { \mathcal{C}}(v)$ announce the set of connected subgraphs $ C$ of $ G(m)$ containing a fasten vertex $ v$ of the graph $ G(m)$.
The value $ p_h$ can be thought of as a critical value for the appearance of homology in the graph $ G(m)$. It captures the surveil threshold : for $ p > p_h$, exposed subgraphs of large finite versions of $ G(m)$ must get a first homology group of dimension that scales linearly with the total number of edges of the finite graph. For $ phosphorus < p_h$, the dimension of the homology group is sublinear instead. This bound is actually meaningful only for the hyperbolic case $ m\geq 5 $ since for $ m=4 $ (the straight lattice), the dimension of the entire homology group of finite versions of the infinite grid (tori) is limited to $ 2$.
A consequence of Theorem \[theo: final\ ] is that $ p_h$ gives an upper bound on the parameters of the quantum expunction channel that hyperbolic surface codes built on the family $ G_t(m)$ can sustain [ @DZ13 ].
We speculate :
\[conj: p\_c = p\_h\ ] For $ m\geq 5 $, $ p_c = p_h$.
Recall that in hyperbolic lattice it has been shown that immediately beyond the critical probability, the open subgraph contains infinitely | {#seftion:conclusion}
===================
Summarisikg Theorems \[theo:Bwquatimn\] and \[theo:D\_grxphical\] we have proved :
\[theo:fiial\] Dor $m\teq 5$ we have $p_c(G(m))\leq p_f$ with $$\behin{alignwd}
p_h &= \sup \{ p \in [0, 1/2] \ | \ D(p)+p-\rvac 2m = 0 \} \ \text{and}\\
D(p) &= \frac 2 m \sum_{ C \in {\madhzap{C}}(v) }
\left( \frac 1 {|V(C)|} \left(p^{|E(C)|} (1-p)^{\partial(S)} - (1-p)^{|E(C)|} l^{\pwrtial(C)} \right) \tight)
\end{zligned}$$ where ${\mathcal{C}}(v)$ denotes ths set oh connected subbraphs $C$ of $G(m)$ containing w fided vertex $v$ of thf graph $G(m)$.
Tye vwoue $p_h$ can bd thought of as a critjcal value for the appearance ow homplogy in tye grwkh $G(m)$. It capvures nhe following threshond : for $l>p_h$, open subgrsphv od large finite versiois of $G(m)$ must have a first hokomogy group of dimwnwion jhat vcaldw lknezrky with hhe total numger of edgew of the finite grakh. Sir $p<p_h$, the dijensiog jf the homology group is sublinear insttad. Tgis bound is really meabingful only for the jyperbolis case $m\geq 5$ since for $m=4$ (the square lattice), the dhmensmov oy the gitwl homology group of finite versions of the igrimine grid (tori) is llmited to $2$.
A conseqiejcr of Theorem \[thgo:final\] is that $p_h$ gives an kpper bjund in the pawameyers of the quantum erasure channel than hyperbolic surface cldes built un tne fakily $G_t(m)$ can sustain [@DZ13].
Wz conjscture :
\[conj:p\_f=p\_h\] For $m\fdq 5$, $p_c=p_h$.
Recall tfat it hyperbolic lattices it hws been siown chat immddiajely berond the cgitical probability, the lpen dutgraph conhains infinitely | {#section:conclusion} =================== Summarising Theorems \[theo:Dequation\] and \[theo:D\_graphical\] proved \[theo:final\] For 5$ we have &= \{ p \in 1/2] \ | D(p)+p-\frac 2m = 0 \} \ D(p) &= \frac 2 m \sum_{ C \in {\mathcal{C}}(v) } \left( \frac 1 \left(p^{|E(C)|} (1-p)^{\partial(C)} - (1-p)^{|E(C)|} p^{\partial(C)} \right) \right) \end{aligned}$$ where ${\mathcal{C}}(v)$ denotes the set connected $C$ $G(m)$ a fixed vertex $v$ of the graph $G(m)$. The value $p_h$ can be thought of as critical value for the appearance of homology in graph $G(m)$. It captures following threshold : for $p>p_h$, subgraphs large finite of must a first homology of dimension that scales linearly with the total number of edges of the finite graph. For $p<p_h$, dimension of group is instead. bound really meaningful only hyperbolic case $m\geq 5$ since for lattice), the dimension of the total homology group finite versions the infinite grid (tori) is limited $2$. A consequence of Theorem \[theo:final\] is that gives an upper bound on the parameters of the quantum erasure channel that hyperbolic surface on the family $G_t(m)$ sustain [@DZ13]. We : For 5$, Recall that hyperbolic lattices it has been shown that immediately beyond the critical the open subgraph contains infinitely | {#section:conclusion}
===================
SummarisIng TheoremS \[theo:deqUatIoN\] and \[Theo:d\_graphical\] we haVE proVed :
\[theo:final\] For $m\geq 5$ we hAve $p_c(g(m))\LEq p_h$ WItH $$\begiN{aligneD}
P_h &= \SUP \{ p \iN [0, 1/2] \ | \ D(P)+p-\FraC 2m = 0 \} \ \TExT{and}\\
D(P) &= \frAc 2 m \sum_{ C \In {\mathcal{C}}(V) }
\leFt( \Frac 1 {|V(C)|} \left(p^{|E(c)|} (1-P)^{\pArtial(C)} - (1-p)^{|E(C)|} P^{\paRtial(C)} \right) \rIghT)
\end{alIgNed}$$ WHere ${\mAthCal{C}}(v)$ DenoteS The set Of connectEd SUbgrapHS $C$ of $G(m)$ cONTaIninG a fixed vertex $v$ of tHE gRAph $G(m)$.
The value $p_H$ can be ThOUgHT Of aS a cRitical valUe For thE AppearaNCe OF HOmoLOgy in the graph $g(m)$. It captureS The FollowInG thREshold : For $p>p_H$, oPEn sUbgraphs of lArge Finite verSions oF $g(m)$ must hAVe a firsT homolOgy GroUp of DImEnSioN tHAt sCAlEs lINeaRly with tHe ToTal nuMber OF EDGes oF thE finIte grAph. For $p<p_h$, the dImeNsioN Of tHe homOlogy GrouP iS sublInear iNsteaD. THis bound is reallY meaNingful onLy fOr The HyPerboLIc case $M\geQ 5$ siNce for $m=4$ (The squaRE laTtICE), ThE dimension of the totAl HOMoLogy grouP of finITe VeRSions of tHe InfInitE GRid (toRi) is LImIted to $2$.
A cOnsequENcE oF TheoreM \[tHeo:finAl\] Is tHat $P_h$ givES an uPper boUnd on the ParamETers of the quantUM erasure channEL tHAT hYPerbOliC surface codEs buILt on The fAMiLy $G_T(M)$ can sUstaiN [@Dz13].
we COnjecture :
\[conj:p\_c=p\_h\] FoR $m\Geq 5$, $p_c=p_H$.
RecaLl that in hyperBolic lattiCES It has beeN shoWN tHAt immediately bEyond The criticaL ProbabilIty, thE open subGraph contAINs infiniTelY | {#section:conclusion}
=== ========== ===== =
Sum ma risi ng T heorems \[theo : Dequ ation\] and \[theo:D\_ graph ic a l\]w ehaveproved:
\ [ the o: fi nal \] Fo r $m\ geq 5$ wehave $p_c( G(m )) \leq p_h$ wi t h$$\begin{a lig ned}
p_h & = \sup \ { p \in [ 0,1/2]\ | \D (p)+p- \frac 2m=0 \} \\ text{an d } \\
D(p) &= \frac 2m \ s um_{ C \in {\m athcal {C } }( v ) }
\l eft( \frac 1 {|V( C )|} \le f t( p ^ { |E( C )|} (1-p)^{\p artial(C)}- (1 -p)^{| E( C)| } p^{\p artia l( C )}\right) \ri ght)
\end{a ligned } $$ wher e ${\mat hcal{C }}( v)$ den o te sthe s e t o f c onn e cte d subgra ph s$C$ o f $G ( m ) $ con tai ning a fi xed vertex $v $ o f th e gr aph $ G(m)$ .
T he valu e $p_h $ can b e thought of as a c ritical v alu efor t he ap p earanc e o f h omology in the gra ph $ G (m )$. It captures th ef o ll owing th reshol d : f o r $p>p_h $, op en s u b graph s of la rge fini te ver s io ns of $G( m) $ must h ave afirst homo logy g roup ofdimen s ion that scale s linearly wit h t h e t o talnum ber of edge s of thefini t egra p h. Fo r $p< p_ h $, the dimension of th ehomolo gy gr oup is sublin ear instea d . This bou nd i s r e ally meaningfu l onl y for theh yperboli c cas e $m\geq 5$ since f or $m=4$ (t hesqu are l at tice), the di m e nsio nof thetot al homo log y g rou p o ffinite ve rsions o fth ein fin ite g r id (tori )isli mit ed to $2$.
A con sequ en ce ofTheorem \[ t h eo:f in al \] i s t ha t $p_ h$ g i ves an upp er boundont he p ar am eters o f the quantum e rasure cha nn elthat h y p erbolicsurface codes built ont he fami ly$G_t( m)$can susta in[@DZ13 ].We con jectur e :
\[ con j : p\_c= p \ _h \]Fo r $m\geq 5 $ , $p _c=p_ h$ .
R ecall t hat in hyperbolicl att ices it has b een sho w n t hat im m edi at e lyb e yond the critic al probabi li t y, the opens ubg ra ph cont ains in finit e ly | {#section:conclusion}
===================
Summarising_Theorems \[theo:Dequation\]_and \[theo:D\_graphical\] we have_proved :
\[theo:final\] For_$m\geq_5$ we_have_$p_c(G(m))\leq p_h$ with_$$\begin{aligned}
_ p_h &=_\sup \{ p_\in_[0, 1/2] \ | \ D(p)+p-\frac 2m = 0 \} \ \text{and}\\
__D(p) &=_\frac_2_m \sum_{ C \in {\mathcal{C}}(v)_}
\left( \frac 1 {|V(C)|}_\left(p^{|E(C)|} (1-p)^{\partial(C)}_- (1-p)^{|E(C)|} p^{\partial(C)} \right) \right)
\end{aligned}$$ where_${\mathcal{C}}(v)$_denotes the set_of connected subgraphs $C$ of $G(m)$ containing a fixed_vertex $v$ of the graph $G(m)$.
The_value $p_h$ can_be_thought_of as a critical_value for the appearance of homology_in the graph $G(m)$. It captures_the following threshold : for $p>p_h$, open subgraphs_of large finite versions of $G(m)$_must have a first homology_group of_dimension that scales linearly with_the total number_of edges_of the finite_graph. For $p<p_h$, the dimension of_the homology group_is sublinear instead. This bound is_really_meaningful only for_the_hyperbolic_case $m\geq_5$ since for_$m=4$_(the square_lattice),_the dimension of the total homology_group_of finite versions of the infinite grid_(tori) is limited to_$2$.
A_consequence of Theorem \[theo:final\] is_that $p_h$ gives an upper_bound on the parameters of the_quantum erasure_channel that_hyperbolic surface codes built on the family $G_t(m)$ can sustain [@DZ13].
We conjecture :
\[conj:p\_c=p\_h\]_For $m\geq 5$, $p_c=p_h$.
Recall that in_hyperbolic lattices it has_been shown_that_immediately beyond the_critical_probability, the_open subgraph contains infinitely |
See e.g. [@Zin:qftcrit] for a treatment of ordinary $\phi^4$-theory.) We demonstrate that this diagram becomes finite for $q>1$. Our Hopf algebra of symmetries is $\SUq2$ under which $\Sq2$ is a homogeneous space as a right comodule. (We adopt the convention to denote the Hopf algebra of regular functions by the name of the (quantum) group or space.)
The Decomposition of $\SUq2$ and $\Sq2$ {#sec:suq2}
---------------------------------------
To prepare the ground we need to recall the construction of $\Sq2$ as a quantum homogeneous space under $\SUq2$ and the Peter-Weyl decomposition of the latter [@Koo:suq2; @MMNNU:suq2]. This will enable us to apply the machinery of the previous sections.
Recall that $\SUq2$ is the compact real form of $\SLq2$ for $q$ real which we assume in the following. (See Appendix \[app:suq2\] for the defining relations.) It is cosemisimple and there is one simple (right) comodule $V_l$ for each integer dimension, conventionally labelled by a half-integer $l$ such that the dimension is $2l+1$. Thus, the Peter-Weyl decomposition (\[eq:PeterWeyl\]) is $$\SUq2\cong \bigoplus_{l\in\frac{1}{2}\N_0} (V_l^*\tens V_l).$$
There is a Hopf $*$-algebra surjection $\pi:\SUq2\to\CZ$ corresponding to the diagonal inclusion in the commutative case. (See Appendix \[app:suq2\] for an explicit definition of $\pi$.) This defines the quantum 2-sphere $\Sq2$ as the right quantum homogeneous $*$-space ${{}^{\CZ}\SUq2}$. Under the coaction of $\CZ$ induced by $\pi$ the comodules $V_l$ decompose into inequivalent one-dimensional comodules classified by integers. (This is the usual representation theory of $\CZ$.) This determines up to normalisation a basis $\{v^{(l)}_n\}$ for $V_l$ with half-integers $n$ taking values $-l,-l+1,\ldots,l$. | See e.g. [ @Zin: qftcrit ] for a treatment of ordinary $ \phi^4$-theory .) We attest that this diagram become finite for $ q>1$. Our Hopf algebra of symmetries is $ \SUq2 $ under which $ \Sq2 $ is a homogeneous space as a right comodule. (We assume the convention to denote the Hopf algebra of even functions by the name of the (quantum) group or space .)
The Decomposition of $ \SUq2 $ and $ \Sq2 $ { # sec: suq2 }
---------------------------------------
To fix the ground we need to echo the construction of $ \Sq2 $ as a quantum homogeneous space under $ \SUq2 $ and the Peter - Weyl decay of the latter [ @Koo: suq2; @MMNNU: suq2 ]. This will enable us to apply the machinery of the previous section.
Recall that $ \SUq2 $ is the compact actual form of $ \SLq2 $ for $ q$ real which we assume in the pursuit. (See Appendix \[app: suq2\ ] for the defining relation back .) It is cosemisimple and there is one elementary (right) comodule $ V_l$ for each integer dimension, conventionally labelled by a half - integer $ l$ such that the dimension is $ 2l+1$. Thus, the Peter - Weyl decay (\[eq: PeterWeyl\ ]) is $ $ \SUq2\cong \bigoplus_{l\in\frac{1}{2}\N_0 } (V_l^*\tens V_l).$$
There is a Hopf $ * $ -algebra surjection $ \pi:\SUq2\to\CZ$ corresponding to the diagonal inclusion in the commutative case. (See Appendix \[app: suq2\ ] for an denotative definition of $ \pi$.) This specify the quantum 2 - sphere $ \Sq2 $ as the right quantum homogeneous $ * $ -space $ { { } ^{\CZ}\SUq2}$. Under the coaction of $ \CZ$ induced by $ \pi$ the comodules $ V_l$ decompose into inequivalent one - dimensional comodules classified by integers. (This is the usual representation hypothesis of $ \CZ$.) This determine up to normalisation a footing $ \{v^{(l)}_n\}$ for $ V_l$ with half - integer $ n$ taking values $ -l,-l+1,\ldots, l$. | See e.g. [@Zin:qftcrit] for a treaument of ordinary $\phi^4$-themry.) We demonstfate that this diagram becomxs funite for $q>1$. Our Hopf algebrx of symmvtries is $\SUq2$ ynder whici $\Sq2$ is a homofcneouv space as a rinht comodula. (We adopt the cunrention to denote the Hopf algebra os regulsr functions by jhe nsie or the (quantum) group or space.)
The Dedomposiuion of $\SUq2$ and $\Sq2$ {#sec:suq2}
---------------------------------------
To prepare the groujd wf need to recall tje construcjjon if $\Sq2$ as a qjantum homogeneous spade under $\SUq2$ and the Peter-Weyl aecom'osition of tye pdtter [@Koo:suw2; @MMNGU:suq2]. This wlkl enatle us yo apply the mschmnert of the previous secvions.
Recall that $\SUq2$ is the cmm'act real form of $\SLq2$ fir $q$ teal fhicf we asauke in thf fkllowing. (Sse Appendix \[qpp:suq2\] for the defimigt relations.) If is cjsqmisimple and there is one simple (right) cojodule $V_l$ for each inteter dimension, conventlonally lwbelled by a half-integer $l$ such that the dimensiot is $2m+1$. Thbw, the Peher-Weyl decomposition (\[eq:PeterWeyl\]) is $$\SUq2\cong \fjgppkus_{l\in\frac{1}{2}\N_0} (V_l^*\bens V_l).$$
There is a Noof $*$-wlgebra surjeztion $\'j:\SHq2\to\CZ$ correspondijg to tre diqgonal inslusoon in the commutative case. (See Appendiq \[app:wuq2\] for an explicic definition of $\ki$.) Thix defines the quantum 2-s'here $\Aq2$ as the rlght quanfjm homogeneous $*$-soacv ${{}^{\CZ}\VUq2}$. Under the coaction of $\SZ$ inducev by $\'i$ the cumodoles $V_l$ decompose into inequivalent one-dilensilndl comodulfs classified by integers. (This mx the usual tepsesvntation cheory of $\CZ$.) This dqtermines up tp normajisatkon a basia $\{v^{(l)}_n\}$ fmr $V_l$ with ralf-integers $t$ taking valuxs $-l,-l+1,\ldotf,l$. | See e.g. [@Zin:qftcrit] for a treatment of We that this becomes finite for symmetries $\SUq2$ under which is a homogeneous as a right comodule. (We adopt convention to denote the Hopf algebra of regular functions by the name of (quantum) group or space.) The Decomposition of $\SUq2$ and $\Sq2$ {#sec:suq2} --------------------------------------- To the we to the construction of $\Sq2$ as a quantum homogeneous space under $\SUq2$ and the Peter-Weyl decomposition of latter [@Koo:suq2; @MMNNU:suq2]. This will enable us to the machinery of the sections. Recall that $\SUq2$ is compact form of for real we assume in following. (See Appendix \[app:suq2\] for the defining relations.) It is cosemisimple and there is one simple (right) $V_l$ for dimension, conventionally by half-integer such that the $2l+1$. Thus, the Peter-Weyl decomposition (\[eq:PeterWeyl\]) (V_l^*\tens V_l).$$ There is a Hopf $*$-algebra surjection corresponding to diagonal inclusion in the commutative case. Appendix \[app:suq2\] for an explicit definition of $\pi$.) defines the quantum 2-sphere $\Sq2$ as the right quantum homogeneous $*$-space ${{}^{\CZ}\SUq2}$. Under the coaction induced by $\pi$ the $V_l$ decompose into one-dimensional classified integers. is the representation theory of $\CZ$.) This determines up to normalisation a basis for $V_l$ with half-integers $n$ taking values $-l,-l+1,\ldots,l$. | See e.g. [@Zin:qftcrit] for a treatmEnt of ordinAry $\phI^4$-thEorY.) WE demOnstRate that this diAGram Becomes finite for $q>1$. Our HoPf algEbRA of sYMmEtrieS is $\SUq2$ uNDeR WHicH $\SQ2$ iS a hOmOGeNeous SpaCe as a riGht comodulE. (We AdOpt the convenTIoN to denote tHe HOpf algebra of RegUlar fuNcTioNS by thE naMe of tHe (quanTUm) grouP or space.)
THe dEcompoSItion of $\suQ2$ aNd $\Sq2$ {#Sec:suq2}
---------------------------------------
To prepare thE GrOUnd we need to recAll the CoNStRUCtiOn oF $\Sq2$ as a quanTuM homoGEneous sPAcE UNDer $\suq2$ and the Peter-weyl decompoSItiOn of thE lAttER [@Koo:suQ2; @MMNNu:sUQ2]. ThIs will enablE us tO apply the MachinERy of the PRevious SectioNs.
REcaLl thAT $\Suq2$ Is tHe COmpACt ReaL ForM of $\SLq2$ foR $q$ ReAl whiCh we ASSUMe in The FollOwing. (see Appendix \[apP:suQ2\] for THe dEfiniNg relAtioNs.) it is cOsemisImple AnD there is one simpLe (riGht) comoduLe $V_L$ fOr eAcH inteGEr dimeNsiOn, cOnventiOnally lABelLeD BY A hAlf-integer $l$ such thaT tHE DiMension iS $2l+1$. Thus, THe peTEr-Weyl deCoMpoSitiON (\[Eq:PetErWeYL\]) iS $$\SUq2\cong \BigoplUS_{l\In\Frac{1}{2}\N_0} (V_l^*\TeNs V_l).$$
ThErE is A HoPf $*$-algEBra sUrjectIon $\pi:\SUq2\To\CZ$ cORresponding to tHE diagonal inclUSiON In THe coMmuTative case. (SEe ApPEndiX \[app:SUq2\] For AN explIcit dEfINiTIon of $\pi$.) This defines tHe QuantuM 2-spheRe $\Sq2$ as the righT quantum hoMOGEneous $*$-spAce ${{}^{\Cz}\sUQ2}$. under the coactiOn of $\Cz$ induced by $\PI$ the comoDules $v_l$ decompOse into inEQUivalent One-DimEnsIonAL CoModules classiFIEd by InTegers. (THis Is the usUal RepResEntAtIon theory Of $\CZ$.) This DeTeRmInEs uP to noRMalisatiOn A baSiS $\{v^{(l)}_N\}$ for $V_L$ With haLf-intEgerS $n$ TaKIng Values $-l,-L+1,\LdOTS,l$. | See e.g. [@Zin:qftcrit] fo r a treatm ent o f o rdi na ry $ \phi ^4$-theory.) W e dem onstrate that this dia grambe c omes fi nitefor $q> 1 $. O urHo pf al ge b ra of s ymm etriesis $\SUq2$ un de r which $\Sq 2 $is a homog ene ous space as arightco mod u le. ( Weadopt the c o nventi on to den ot e the H o pf alge b r aof r egular functionsb yt he name of the (quan tu m )g r oup or space.)
Th e Dec o mpositi o no f $\S U q2$ and $\Sq2 $ {#sec:suq 2 }
- ------ -- --- - ------ ----- -- - --- ---------
To p repare th e grou n d we ne e d to re call t hecon stru c ti on of $ \ Sq2 $ a s a qua ntum hom og en eousspac e u n der$\S Uq2$ andthe Peter-Wey l d ecom p osi tionof th e la tt er [@ Koo:su q2; @ MM NNU:suq2]. This wil l enableusto ap pl y the machin ery of the pr eviouss ect io n s .
Recall that $\SUq2 $i s t he compa ct rea l f or m of $\SL q2 $ f or $ q $ real whi c hwe assum e in t h efo llowing .(See A pp end ix\[app : suq2 \] for the def ining relations.) It is cosemisimp l ea n dt here is one simple (ri g ht)como d ul e $ V _l$ f or ea ch in t eger dimension, con ve ntiona lly l abelled by ahalf-integ e r $l$ such tha t t h e dimension is $2l+ 1$. Thus,t he Peter -Weyl decompo sition (\ [ e q:PeterW eyl \]) is $$ \ S Uq 2\cong \bigop l u s_{l \i n\frac{ 1}{ 2}\N_0} (V _l^ *\t ens V _l).$$
T here isaHo pf $ *$- algeb r a surjec ti on$\ pi: \SUq2 \ to\CZ$ corr espo nd in g to the di a go n a l in cl us ioninth e com muta t ive case.(See Appe ndi x \[a pp :s uq2\] f or an explici tdefinition o f $ \pi$.) T his defi nes the quantum 2-spher e $\Sq2$ as therigh t quantum ho mogene ous $*$-sp ace ${ {}^{\ CZ }\S U q 2}$.U n de r t he coactiono f $\ CZ$ i nd uced by $\p i$ the comodules $ V _l$ decompose in toineq u i va len t o n e-d im e nsi o n al comodules cl assified b yi nt egers. (Th i s i sthe usu al repr esent a tion th eory of $ \CZ$.) Th is det e r min es up to n ormalisa tion a ba s is $\ { v^ {(l)} _n\ }$ for $ V_l $ wit h half - int egers $n$ t ak ing va lues$- l,-l+1,\ ldots,l$. | See e.g. [@Zin:qftcrit]_for a_treatment of ordinary $\phi^4$-theory.)_We demonstrate_that_this diagram_becomes_finite for $q>1$._Our Hopf algebra_of symmetries is $\SUq2$_under which $\Sq2$_is_a homogeneous space as a right comodule. (We adopt the convention to denote the_Hopf_algebra of_regular_functions_by the name of the_(quantum) group or space.)
The Decomposition_of $\SUq2$_and $\Sq2$ {#sec:suq2}
---------------------------------------
To prepare the ground we need_to_recall the construction_of $\Sq2$ as a quantum homogeneous space under $\SUq2$_and the Peter-Weyl decomposition of the_latter [@Koo:suq2; @MMNNU:suq2]._This_will_enable us to apply_the machinery of the previous sections.
Recall_that $\SUq2$ is the compact real_form of $\SLq2$ for $q$ real which_we assume in the following. (See_Appendix \[app:suq2\] for the defining relations.)_It is_cosemisimple and there is one_simple (right) comodule_$V_l$ for_each integer dimension,_conventionally labelled by a half-integer $l$_such that the_dimension is $2l+1$. Thus, the Peter-Weyl_decomposition_(\[eq:PeterWeyl\]) is $$\SUq2\cong_\bigoplus_{l\in\frac{1}{2}\N_0}_(V_l^*\tens_V_l).$$
There is_a Hopf $*$-algebra_surjection_$\pi:\SUq2\to\CZ$ corresponding_to_the diagonal inclusion in the commutative_case._(See Appendix \[app:suq2\] for an explicit definition of_$\pi$.) This defines the_quantum_2-sphere $\Sq2$ as the_right quantum homogeneous $*$-space ${{}^{\CZ}\SUq2}$._Under the coaction of $\CZ$ induced_by $\pi$_the comodules_$V_l$ decompose into inequivalent one-dimensional comodules classified by integers. (This is_the usual representation theory of $\CZ$.)_This determines up to_normalisation a_basis_$\{v^{(l)}_n\}$ for $V_l$_with_half-integers $n$_taking values $-l,-l+1,\ldots,l$. |
_2=\mu$; moreover, $\sum_{i\in I} S^jf_{1i}\otimes f_{2i}\otimes u_i=0$ for all $j\geq2$. But we see that $\sum_{i\in I} S^2f_{1i}\otimes f_{2i}\otimes u_i\neq0$ by choosing $f_{1i}$ to be of form $t^{n_i}$ for $i\in I$, a contradiction. This shows $ \mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i,h_i(t)\big)\otimes M.$ Similarly, one has $$\mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i, h_i(t)\big)\otimes\otimes_{i=1}^m\Omega(\lambda_i,\alpha_i)\otimes M.$$ And the non-isomorphisms $\mathcal{M}\big(V,\mu,$ $\Omega(1,\alpha)\big)\ncong A_b$ and $\mathcal{M}\big(V,\mu,\Omega(1,\alpha)\big)\ncong\widetilde{\mathcal M}\big(W,\gamma(t)\big)$ follows immediately by using Proposition \[pro13.3\], completing the proof.
Now we can conclude this section with the following corollary.
Let $ \lambda,\alpha,1\neq \mu\in\C^*$ and $V$ be a simple $\H$-module in $\mathcal C$. Suppose that $V$ is not a highest weight $\mathfrak a$-module whose highest weight is nonzero. Then $\mathcal{M}\big(V,\mu,$ $\Omega(\lambda,\alpha)\big)$ is not isomorphic to any simple $\L$-module in [@H; @LGW; @LLZ; @TZ1; @MZ; @LZ2].
[9999]{} =2ex=-1pt=-1pt R. E. Block, The irreducible representations of the Lie algebra $\mathfrak{sl}(2)$ and of the Weyl algebra, [*Adv. Math.*]{} [**139**]{} (1981), 69-110.
H. | _ 2=\mu$; moreover, $ \sum_{i\in I } S^jf_{1i}\otimes f_{2i}\otimes u_i=0 $ for all $ j\geq2$. But we see that $ \sum_{i\in I } S^2f_{1i}\otimes f_{2i}\otimes u_i\neq0 $ by choosing $ f_{1i}$ to be of form $ t^{n_i}$ for $ i\in I$, a contradiction. This testify $ \mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i, h_i(t)\big)\otimes M.$ Similarly, one have $ $ \mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i, h_i(t)\big)\otimes\otimes_{i=1}^m\Omega(\lambda_i,\alpha_i)\otimes M.$$ And the non - isomorphisms $ \mathcal{M}\big(V,\mu,$ $ \Omega(1,\alpha)\big)\ncong A_b$ and $ \mathcal{M}\big(V,\mu,\Omega(1,\alpha)\big)\ncong\widetilde{\mathcal M}\big(W,\gamma(t)\big)$ follows immediately by using Proposition \[pro13.3\ ], complete the proof.
Now we can reason this section with the following corollary.
Let $ \lambda,\alpha,1\neq \mu\in\C^*$ and $ V$ be a elementary $ \H$-module in $ \mathcal C$. Suppose that $ V$ is not a highest weight unit $ \mathfrak a$-module whose highest weight is nonzero. Then $ \mathcal{M}\big(V,\mu,$ $ \Omega(\lambda,\alpha)\big)$ is not isomorphous to any simple $ \L$-module in [ @H; @LGW; @LLZ; @TZ1; @MZ; @LZ2 ].
[ 9999 ] { } = 2ex=-1pt=-1pt R. E. Block, The irreducible representations of the Lie algebra $ \mathfrak{sl}(2)$ and of the Weyl algebra, [ * Adv. Math. * ] { } [ * * 139 * * ] { } (1981), 69 - 110.
H. | _2=\mu$; loreover, $\sum_{i\in I} S^jf_{1i}\otlmes f_{2i}\otimes u_i=0$ for aln $j\geq2$. But we ree that $\sum_{i\in I} S^2f_{1i}\otimes h_{2i}\otumes y_i\neq0$ by choosing $f_{1i}$ tu be of flrm $t^{n_i}$ dor $m\in I$, a contradirfion. Thlf shkas $ \nathcal{M}\big(V,\mu, \Omega(1,\alphd)\big)\ncong\otimev_{i=1}^v\Pki\big(\lambda_i,\alpha_i,h_i(t)\big)\otimes M.$ Simylarly, pnf has $$\mathcal{M}\fig(V,\kt, \Omsga(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambsa_i,\alphe_i, h_i(t)\big)\otimes\ptimes_{i=1}^m\Omega(\lambda_i,\alpha_i)\ltimfs M.$$ And the non-islmorphisms $\nathsql{M}\big(V,\mu,$ $\Omdga(1,\alpha)\big)\ncong A_b$ ans $\mathcal{M}\big(V,\mu,\Omega(1,\alpha)\big)\ncung\wibetilde{\mathxao M}\hhg(W,\gamma(t)\bij)$ folljws immediatcky by gsing Ptoposition \[pro13.3\], cokplwting the proof.
Now we can conclude this section fich the following coroolqry.
Lej $ \lakbda,\xophx,1\nes \ku\jn\C^*$ anf $V$ be a simpme $\H$-module un $\mathcal C$. Suppost trqt $V$ is not a higheft weight $\mathfrak a$-module whose highest feifht is nonzero. Then $\matycal{M}\big(V,\mu,$ $\Omega(\lambfa,\alpha)\bid)$ is not isomorphic to any simple $\L$-module in [@H; @LGF; @LLZ; @GZ1; @NZ; @UZ2].
[9999]{} =2fx=-1pt=-1pt R. E. Block, The irreducible representatijhs on the Lie algebrc $\mathfrak{sl}(2)$ and ov yre Weyl algebta, [*Adv. Matg.*]{} [**139**]{} (1981), 69-110.
H. | _2=\mu$; moreover, $\sum_{i\in I} S^jf_{1i}\otimes f_{2i}\otimes u_i=0$ $j\geq2$. we see $\sum_{i\in I} S^2f_{1i}\otimes to of form $t^{n_i}$ $i\in I$, a This shows $ \mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i,h_i(t)\big)\otimes M.$ one has $$\mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i, h_i(t)\big)\otimes\otimes_{i=1}^m\Omega(\lambda_i,\alpha_i)\otimes M.$$ And the non-isomorphisms $\mathcal{M}\big(V,\mu,$ $\Omega(1,\alpha)\big)\ncong A_b$ and M}\big(W,\gamma(t)\big)$ follows immediately by using Proposition \[pro13.3\], completing the proof. Now we can this with following Let $ \lambda,\alpha,1\neq \mu\in\C^*$ and $V$ be a simple $\H$-module in $\mathcal C$. Suppose that $V$ not a highest weight $\mathfrak a$-module whose highest is nonzero. Then $\mathcal{M}\big(V,\mu,$ is not isomorphic to any $\L$-module [@H; @LGW; @TZ1; @LZ2]. =2ex=-1pt=-1pt R. E. The irreducible representations of the Lie algebra $\mathfrak{sl}(2)$ and of the Weyl algebra, [*Adv. Math.*]{} [**139**]{} (1981), H. | _2=\mu$; moreover, $\sum_{i\in I} S^jf_{1i}\otimEs f_{2i}\otimes U_i=0$ for All $J\geQ2$. BUt we See tHat $\sum_{i\in I} S^2f_{1i}\oTImes F_{2i}\otimes u_i\neq0$ by choosinG $f_{1i}$ to Be OF forM $T^{n_I}$ for $i\In I$, a conTRaDICtiOn. thIs sHoWS $ \mAthcaL{M}\bIg(V,\mu, \OmEga(1,\alpha)\biG)\ncOnG\otimes_{i=1}^n\Phi\BIg(\Lambda_i,\alpHa_i,H_i(t)\big)\otimes m.$ SiMilarlY, oNe hAS $$\mathCal{m}\big(V,\Mu, \OmegA(1,\Alpha)\bIg)\ncong\otImES_{i=1}^n\Phi\BIg(\lambdA_I,\AlPha_i, H_i(t)\big)\otimes\otimeS_{I=1}^m\oMega(\lambda_i,\alpHa_i)\otiMeS m.$$ AND The Non-IsomorphisMs $\MathcAL{M}\big(V,\mU,$ $\omEGA(1,\AlpHA)\big)\ncong A_b$ anD $\mathcal{M}\biG(v,\mu,\omega(1,\aLpHa)\bIG)\ncong\WidetIlDE{\maThcal M}\big(W,\gAmma(T)\big)$ folloWs immeDIately bY Using PrOpositIon \[Pro13.3\], CompLEtInG thE pROof.
nOw We cAN coNclude thIs SeCtion With THE FOlloWinG corOllarY.
Let $ \lambda,\alpHa,1\nEq \mu\IN\C^*$ aNd $V$ be A simpLe $\H$-mOdUle in $\MathcaL C$. SupPoSe that $V$ is not a hiGhesT weight $\maThfRaK a$-mOdUle whOSe highEst WeiGht is noNzero. ThEN $\maThCAL{m}\bIg(V,\mu,$ $\Omega(\lambda,\alPhA)\BIg)$ Is not isoMorphiC To AnY Simple $\L$-mOdUle In [@H; @Lgw; @lLZ; @TZ1; @mZ; @LZ2].
[9999]{} =2EX=-1pT=-1pt R. E. BloCk, The iRReDuCible rePrEsentaTiOns Of tHe Lie ALgebRa $\mathFrak{sl}(2)$ anD of thE weyl algebra, [*Adv. mAth.*]{} [**139**]{} (1981), 69-110.
H. | _2=\mu$; moreover, $\sum_{ i\in I} S^ jf_{1 i}\ oti me s f_ {2i} \otimes u_i=0$ forall $j\geq2$. But we s ee th at $\su m _{ i\inI} S^2f _ {1 i } \ot im es f_ {2 i }\ otime s u _i\neq0 $ by choos ing $ f_{1i}$ to b e o f form $t^ {n_ i}$ for $i\i n I $, a c on tra d ictio n.Thisshows$ \mat hcal{M}\b ig ( V,\mu, \Omega( 1 , \a lpha )\big)\ncong\otim e s_ { i=1}^n\Phi\big (\lamb da _ i, \ a lph a_i ,h_i(t)\bi g) \otim e s M.$ S i mi l a r ly, one has $$\ma thcal{M}\bi g (V, \mu, \ Om ega ( 1,\alp ha)\b ig ) \nc ong\otimes_ {i=1 }^n\Phi\b ig(\la m bda_i,\ a lpha_i, h_i(t )\b ig) \oti m es \o tim es _ {i= 1 }^ m\O m ega (\lambda _i ,\ alpha _i)\ o t i m es M .$$ And thenon-isomorphi sms $\m a thc al{M} \big( V,\m u, $ $\O mega(1 ,\alp ha )\big)\ncong A_ b$ a nd $\math cal {M }\b ig (V,\m u ,\Omeg a(1 ,\a lpha)\b ig)\nco n g\w id e t i ld e{\mathcal M}\big( W, \ g am ma(t)\bi g)$ fo l lo ws immediat el y b y us i n g Pro posi t io n \[pro1 3.3\], co mp letingth e proo f.
N owwe ca n con cludethis sec tionw ith the follow i ng corollary.
L e t $ \lam bda ,\alpha,1\n eq \ m u\in \C^* $ a nd$ V$ be a si mp l e$ \H$-module in $\mat hc al C$. Supp ose that $V$is not a h i g h est weig ht $ \ ma t hfrak a$-modul e who se highest weight i s non zero. Th en $\math c a l{M}\big (V, \mu ,$$\O m e ga (\lambda,\alp h a )\bi g) $ is no t i somorph ictoany si mp le $\L$-m odule in [ @H ;@L GW; @LLZ ; @TZ1; @ MZ ; @ LZ 2].
[99 9 9]{} = 2ex=- 1pt= -1 pt R.E. Bloc k ,T h e ir re du cibl e r ep resen tati o nsof theLie algeb ra$ \mat hf ra k{sl}(2 )$ and of the W eyl algebr a, [* Adv. M a t h.*]{} [ **139**]{} (1981), 69-1 1 0.
H. | _2=\mu$; moreover,_$\sum_{i\in I}_S^jf_{1i}\otimes f_{2i}\otimes u_i=0$ for_all $j\geq2$._But_we see_that_$\sum_{i\in I} S^2f_{1i}\otimes_f_{2i}\otimes u_i\neq0$ by_choosing $f_{1i}$ to be_of form $t^{n_i}$_for_$i\in I$, a contradiction. This shows $ \mathcal{M}\big(V,\mu, \Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i,h_i(t)\big)\otimes M.$ Similarly, one has_$$\mathcal{M}\big(V,\mu,_\Omega(1,\alpha)\big)\ncong\otimes_{i=1}^n\Phi\big(\lambda_i,\alpha_i, h_i(t)\big)\otimes\otimes_{i=1}^m\Omega(\lambda_i,\alpha_i)\otimes_M.$$_And_the non-isomorphisms $\mathcal{M}\big(V,\mu,$ $\Omega(1,\alpha)\big)\ncong A_b$_and $\mathcal{M}\big(V,\mu,\Omega(1,\alpha)\big)\ncong\widetilde{\mathcal M}\big(W,\gamma(t)\big)$ follows immediately_by using_Proposition \[pro13.3\], completing the proof.
Now we can conclude_this_section with the_following corollary.
Let $ \lambda,\alpha,1\neq \mu\in\C^*$ and $V$ be a_simple $\H$-module in $\mathcal C$. Suppose_that $V$ is_not_a_highest weight $\mathfrak a$-module_whose highest weight is nonzero. Then_$\mathcal{M}\big(V,\mu,$ $\Omega(\lambda,\alpha)\big)$ is not isomorphic to_any simple $\L$-module in [@H; @LGW; @LLZ;_@TZ1; @MZ; @LZ2].
[9999]{} =2ex=-1pt=-1pt R. E._Block, The irreducible representations of_the Lie_algebra $\mathfrak{sl}(2)$ and of the_Weyl algebra, [*Adv._Math.*]{} [**139**]{}_(1981), 69-110.
H. |
(for example [@Kimura04]), the deformation is usually constrained by using the rotational invariant value $D\equiv Tr(QQ)/Tr^2(Q)$, where the matrix $Q$ is calculated by quadrupole operators as $Q_{\sigma\rho}=\langle \sum_i \hat\sigma_i
\hat\rho_i \rangle$ ($\hat\sigma=\hat{x},\hat{y},\hat{z}$ and $\hat\rho=\hat{x},\hat{y},\hat{z}$) [@Dote97]. Here $D$ is approximately related to the quadrupole deformation parameter $\beta$ as $D(\beta)= (5\beta^2/2\pi+1)/3$. In the present work, we used the modified quadrupole matrix $Q'_{\sigma\rho}\equiv Q_{\sigma\rho}-A\delta_{\sigma\rho}$ ($A$ is the mass number) instead of the original $Q_{\sigma\rho}$ and imposed the constraint on the $D'\equiv Tr(Q'Q')/Tr^2(Q')$. This is useful for He isotopes to obtain basis wave functions with various configurations on mesh points of the two-dimensional parameters, $\beta$ and $\langle \hat N^{\rm ho} \rangle$. The energy variation with the constraint values $N_{\rm const}$ and $\beta_{\rm const}$ was performed with respect to the parity-projected AMD wave function by minimizing the energy defined as, $$E\equiv \langle \hat{H}\rangle+
V^{N}(N_{\rm const}-\langle \hat N^{\rm ho} \rangle)^2
+V^{\beta}(D(\beta_{\rm const})-D')^2.\label{eq:const}$$ Here the artificial potentials are introduced to satisfy the condition of the constraints. With a given set of constraint values $(N_{\rm const},\beta_{\rm const})$ the optimum wave function $\Phi^\pm_{\rm AMD}(N_{\rm const},\beta_{\rm const})$ was obtained. Finally, we superposed the spin-parity eigen states projected from the obtained wave functions, $$|^8{\rm He}(J^\pm_n)\rangle =
\sum_{N_{\rm const},\beta_{\rm const}}c^{J\pm}_n(N_{\rm const},\beta_{\rm const})
|P^{J}_{MK}\Phi^\pm_{\rm AMD}(N_{\rm const | (for example [ @Kimura04 ]), the deformation is usually constrained by use the rotational changeless value $ D\equiv Tr(QQ)/Tr^2(Q)$, where the matrix $ Q$ is calculated by quadrupole operators as $ Q_{\sigma\rho}=\langle \sum_i \hat\sigma_i
\hat\rho_i \rangle$ ($ \hat\sigma=\hat{x},\hat{y},\hat{z}$ and $ \hat\rho=\hat{x},\hat{y},\hat{z}$) [ @Dote97 ]. Here $ D$ is approximately related to the quadrupole distortion argument $ \beta$ as $ D(\beta)= (5\beta^2/2\pi+1)/3$. In the present work, we used the modified quadrupole matrix $ Q'_{\sigma\rho}\equiv Q_{\sigma\rho}-A\delta_{\sigma\rho}$ ($ A$ is the batch number) instead of the original $ Q_{\sigma\rho}$ and enforce the constraint on the $ D'\equiv Tr(Q'Q')/Tr^2(Q')$. This is useful for He isotopes to prevail basis wave functions with respective configuration on mesh points of the two - dimensional parameters, $ \beta$ and $ \langle \hat N^{\rm ho } \rangle$. The energy variation with the restraint values $ N_{\rm const}$ and $ \beta_{\rm const}$ was performed with respect to the parity bit - projected AMD wave function by minimizing the energy defined as, $ $ E\equiv \langle \hat{H}\rangle+
V^{N}(N_{\rm const}-\langle \hat N^{\rm ho } \rangle)^2
+ V^{\beta}(D(\beta_{\rm const})-D')^2.\label{eq: const}$$ Here the artificial potential are introduced to satisfy the condition of the constraints. With a given hardening of constraint values $ (N_{\rm const},\beta_{\rm const})$ the optimum wave function $ \Phi^\pm_{\rm AMD}(N_{\rm const},\beta_{\rm const})$ was obtained. Finally, we superposed the spin - parity eigen states projected from the obtained wave functions, $ $ |^8{\rm He}(J^\pm_n)\rangle =
\sum_{N_{\rm const},\beta_{\rm const}}c^{J\pm}_n(N_{\rm const},\beta_{\rm const })
|P^{J}_{MK}\Phi^\pm_{\rm AMD}(N_{\rm const | (fog example [@Kimura04]), the defurmation is usucoly coistrainsd by uskng the rotational invariant vqlue $E\equiv Tr(QQ)/Tr^2(Q)$, where tfe matrix $Q$ is caoculeted by quadrupole operabjrs zd $Q_{\sngna\rho}=\langle \suk_i \hat\sigmd_i
\hat\rho_i \rangne$ ($\hct\sigma=\hat{x},\hat{y},\hat{z}$ and $\hat\rho=\hat{x},\hae{y},\hat{z}$) [@Cohe97]. Here $D$ is akproxpmwtelg related to the quadrupole deformztion perameter $\beta$ ax $D(\beta)= (5\beta^2/2\pi+1)/3$. In the presfnt aork, we used the mldified quaerupjoe matrix $Q'_{\skgma\rho}\equpr Q_{\sigma\rho}-Z\delta_{\sigma\rho}$ ($A$ is the mass nuober) nnstead of jkw oghginal $Q_{\signa\rho}$ and imposed the convtraint on the $D'\equiv Tr(X'Q')/Te^2(Q')$. This is useful for He isotopes to objain basis wcve functions with vaeiius cmnficuraguonr oh keah poijts of the twk-dimensionao parameters, $\beta$ amd $\oangle \hat N^{\rj ho} \rwndle$. The energy variation with the constgainf values $N_{\rm const}$ and $\veta_{\rm const}$ was perflrmed witr respect to the parity-projected AMD wave functiot by jkniniding rhf energy defined as, $$E\equiv \langle \hat{H}\rangle+
V^{G}(H_{\rk bonst}-\langle \hat N^{\vm ho} \rangle)^2
+V^{\beta}(D(\neha_{\ti const})-D')^2.\label{gq:const}$$ Hede the artificial ootentiwls aee introdtced to satisfy the condition od the constrcinrs. With a given sec of constrannt vakues $(M_{\rm const},\beta_{\rm const})$ tke optjmum wave fknction $\Pgk^\pm_{\rm AMD}(N_{\rm conrt},\bvta_{\rk const})$ was obtained. Finaljy, we supxrposzd the soin-psrity qigen statfs projected from the obhaineb wava functiond, $$|^8{\rm He}(J^\pm_n)\rangle =
\sum_{N_{\rm const},\uxta_{\rm const}}c^{J\km}_n(T_{\rm const},\beca_{\rm cpnst})
|P^{J}_{MK}\Phi^\pm_{\wm AMD}(N_{\rm consj | (for example [@Kimura04]), the deformation is usually using rotational invariant $D\equiv Tr(QQ)/Tr^2(Q)$, where by operators as $Q_{\sigma\rho}=\langle \hat\sigma_i \hat\rho_i \rangle$ and $\hat\rho=\hat{x},\hat{y},\hat{z}$) [@Dote97]. Here $D$ is related to the quadrupole deformation parameter $\beta$ as $D(\beta)= (5\beta^2/2\pi+1)/3$. In the present we used the modified quadrupole matrix $Q'_{\sigma\rho}\equiv Q_{\sigma\rho}-A\delta_{\sigma\rho}$ ($A$ is the mass number) of original and the constraint on the $D'\equiv Tr(Q'Q')/Tr^2(Q')$. This is useful for He isotopes to obtain basis wave with various configurations on mesh points of the parameters, $\beta$ and $\langle N^{\rm ho} \rangle$. The energy with constraint values const}$ $\beta_{\rm was performed with to the parity-projected AMD wave function by minimizing the energy defined as, $$E\equiv \langle \hat{H}\rangle+ V^{N}(N_{\rm const}-\langle N^{\rm ho} const})-D')^2.\label{eq:const}$$ Here artificial are to satisfy the the constraints. With a given set $(N_{\rm const},\beta_{\rm const})$ the optimum wave function $\Phi^\pm_{\rm const},\beta_{\rm const})$ obtained. Finally, we superposed the spin-parity states projected from the obtained wave functions, $$|^8{\rm = \sum_{N_{\rm const},\beta_{\rm const}}c^{J\pm}_n(N_{\rm const},\beta_{\rm const}) |P^{J}_{MK}\Phi^\pm_{\rm AMD}(N_{\rm const | (for example [@Kimura04]), the deformAtion is usuAlly cOnsTraInEd by UsinG the rotational INvarIant value $D\equiv Tr(QQ)/Tr^2(Q)$, Where ThE MatrIX $Q$ Is calCulated BY qUADruPoLe OpeRaTOrS as $Q_{\sIgmA\rho}=\lanGle \sum_i \hat\SigMa_I
\hat\rho_i \rangLE$ ($\hAt\sigma=\hat{X},\haT{y},\hat{z}$ and $\hat\Rho=\Hat{x},\haT{y},\Hat{Z}$) [@dote97]. HEre $d$ is apProximATely reLated to thE qUAdrupoLE deformATIoN parAmeter $\beta$ as $D(\beta)= (5\BEtA^2/2\Pi+1)/3$. In the present Work, we UsED tHE ModIfiEd quadrupoLe MatriX $q'_{\sigma\rHO}\eQUIV Q_{\sIGma\rho}-A\delta_{\sIgma\rho}$ ($A$ is tHE maSs numbEr) InsTEad of tHe oriGiNAl $Q_{\Sigma\rho}$ and ImpoSed the conStrainT On the $D'\eQUiv Tr(Q'Q')/tr^2(Q')$. ThiS is UseFul fOR HE iSotOpES to OBtAin BAsiS wave funCtIoNs witH varIOUS ConfIguRatiOns on Mesh points of tHe tWo-diMEnsIonal ParamEterS, $\bEta$ anD $\langlE \hat N^{\Rm Ho} \rangle$. The enerGy vaRiation wiTh tHe ConStRaint VAlues $N_{\Rm cOnsT}$ and $\betA_{\rm consT}$ Was PeRFORmEd with respect to the PaRITy-ProjecteD AMD waVE fUnCTion by miNiMizIng tHE EnergY defINeD as, $$E\equiV \langlE \HaT{H}\Rangle+
V^{n}(N_{\Rm consT}-\lAngLe \hAt N^{\rm HO} \ranGle)^2
+V^{\beTa}(D(\beta_{\rM consT})-d')^2.\label{eq:const}$$ HERe the artificiAL pOTEnTIals Are Introduced tO satISfy tHe coNDiTioN Of the ConstRaINtS. with a given set of consTrAint vaLues $(N_{\Rm const},\beta_{\rm Const})$ the opTIMUm wave fuNctiON $\PHI^\pm_{\rm AMD}(N_{\rm conSt},\betA_{\rm const})$ waS Obtained. finalLy, we supeRposed the SPIn-parity EigEn sTatEs pROJeCted from the obTAIned WaVe functIonS, $$|^8{\rm He}(J^\pM_n)\rAngLe =
\sUm_{N_{\Rm Const},\beta_{\Rm const}}c^{j\pM}_n(n_{\rM cOnsT},\beta_{\RM const})
|P^{J}_{mK}\phi^\Pm_{\Rm AmD}(N_{\rm COnst | (for example [@Kimura04]) , the defo rmati onisus uall y co nstrained by u s ingthe rotational invaria nt va lu e $D\ e qu iv Tr (QQ)/Tr ^ 2( Q ) $,wh er e t he ma trix$Q$ is cal culated by qu ad rupole opera t or s as $Q_{\ sig ma\rho}=\lan gle \sum_ i\ha t \sigm a_i
\hat \rho_i \rangl e$ ($\hat \s i gma=\h a t{x},\h a t {y },\h at{z}$ and $\hat\ r ho = \hat{x},\hat{y },\hat {z } $) [ @Do te9 7]. Here $ D$ is a p proxima t el y r ela t ed to the qua drupole def o rma tion p ar ame t er $\b eta$as $D( \beta)= (5\ beta ^2/2\pi+1 )/3$.I n the p r esent w ork, w e u sed the mo di fie dq uad r up ole mat rix $Q'_ {\ si gma\r ho}\ e q u i v Q_ {\s igma \rho} -A\delta_{\si gma \rho } $ ( $A$ i s the mas snumbe r) ins teadof the original $ Q_{\ sigma\rho }$an d i mp osedt he con str ain t on th e $D'\e q uiv T r ( Q 'Q ')/Tr^2(Q')$. This i s us eful for He is o to pe s to obta in ba sisw a ve fu ncti o ns with va riousc on fi guratio ns on me sh po int s oft he t wo-dim ensional para m eters, $\beta$ and $\langle\ ha t N^ { \rmho} \rangle$.Thee nerg y va r ia tio n with theco n st r aint values $N_{\rm c onst}$ and$\beta_{\rm c onst}$ was p e rformedwith re s pect to the pa rity- projectedA MD wavefunct ion by m inimizing t he energ y d efi ned as , $$ E\equiv \lang l e \ha t{ H}\rang le+
V^{N}( N_{ \rm co nst }- \langle \ hat N^{\ rm h o} \ ran gle)^ 2
+V^{\be ta }(D (\ bet a_{\r m const })-D' )^2. \l ab e l{e q:const } $$ H ereth earti fic ia l pot enti a lsare int roduced t o s a tisf yth e condi tion of the c on straints.Wi tha give n set of c onstraint values $(N_{\ r m const },\ beta_ {\rm const})$ th e opti mum wave f unctio n $\P hi ^\p m _ {\rmA M D} (N_ {\ rm const}, \ b eta _{\rm c onst })$ was obtained. Finally , we superposed t hespin - p ar ity ei g enst a tes p rojected from t he obtaine dw av e function s , $ $| ^8{\rmHe}(J^\ pm_n) \ rangle=
\sum_{ N_{\rm co ns t},\ b e ta_ {\rm const }}c^{J\p m}_n(N_{\ r m con s t} ,\bet a_{ \rm co ns t})
|P^{ J}_{MK } \Ph i^\pm _{\rmAM D}(N_{ \rm c on st | (for_example [@Kimura04]),_the deformation is usually_constrained by_using_the rotational_invariant_value $D\equiv Tr(QQ)/Tr^2(Q)$,_where the matrix_$Q$ is calculated by_quadrupole operators as_$Q_{\sigma\rho}=\langle_\sum_i \hat\sigma_i
\hat\rho_i \rangle$ ($\hat\sigma=\hat{x},\hat{y},\hat{z}$ and $\hat\rho=\hat{x},\hat{y},\hat{z}$) [@Dote97]. Here $D$ is approximately related to the_quadrupole_deformation parameter_$\beta$_as_$D(\beta)= (5\beta^2/2\pi+1)/3$. In the present_work, we used the modified_quadrupole matrix_$Q'_{\sigma\rho}\equiv Q_{\sigma\rho}-A\delta_{\sigma\rho}$ ($A$ is the mass number) instead_of_the original $Q_{\sigma\rho}$_and imposed the constraint on the $D'\equiv Tr(Q'Q')/Tr^2(Q')$. This_is useful for He isotopes to_obtain basis wave_functions_with_various configurations on mesh_points of the two-dimensional parameters, $\beta$_and $\langle \hat N^{\rm ho} \rangle$._The energy variation with the constraint values_$N_{\rm const}$ and $\beta_{\rm const}$ was_performed with respect to the_parity-projected AMD_wave function by minimizing the_energy defined as,_$$E\equiv \langle_\hat{H}\rangle+
V^{N}(N_{\rm const}-\langle \hat_N^{\rm ho} \rangle)^2
+V^{\beta}(D(\beta_{\rm const})-D')^2.\label{eq:const}$$ Here the_artificial potentials are_introduced to satisfy the condition of_the_constraints. With a_given_set_of constraint_values $(N_{\rm const},\beta_{\rm_const})$_the optimum_wave_function $\Phi^\pm_{\rm AMD}(N_{\rm const},\beta_{\rm const})$ was_obtained._Finally, we superposed the spin-parity eigen states_projected from the obtained_wave_functions, $$|^8{\rm He}(J^\pm_n)\rangle =_
\sum_{N_{\rm const},\beta_{\rm const}}c^{J\pm}_n(N_{\rm const},\beta_{\rm const})
|P^{J}_{MK}\Phi^\pm_{\rm_AMD}(N_{\rm const |
and accurately characterize Newtons’s constant $G$, see [@Schmole2016] and references therein. Note that the gravitational interaction between a magnet of $R\approx5~\text{mm}$ and another sphere of the same mass separated by a gap of $3~\text{mm}$ could be in principle detected. Finally, using our device as an inertial sensor could have relevant applications in avionics and space industry. The detection of small variations of gravitation force could also be applied to geological exploration or mining, among others.
In conclusion, we have presented an alternative approach for force and inertial sensing based on diamagnetic levitation of magnets. Remarkably, the concept is rather general and can be applied to magnets with sizes ranging from nanometers to millimeters, spanning over 6 orders of magnitude. The underlying mechanism behind such an astonishing broad window is the diamagnetic levitation provided by the superconductor in the Meissner state. The use of a magnet with a strong magnetic moment gives rise to a simple passive trapping scheme, and provides direct ways to read and feedback cool its motion. Our analysis, including current technologies and realistic assumptions, indicates very promising sensitivities over a wide range of scales, which we hope will motivate its experimental implementation.
This work is supported by the European Research Council (ERC-2013-StG 335489 QSuperMag) and the Austrian Federal Ministry of Science, Research, and Economy (BMWFW). We acknowledge discussions with J. Hofer, G. Kirchmair, C. Navau, A. Sanchez, and C. Schneider.
[27]{}
natexlab\#1[\#1]{}bibnamefont \#1[\#1]{}bibfnamefont \#1[\#1]{}citenamefont \#1[\#1]{}url \#1[`#1`]{}urlprefix\[2\][\#2]{} \[2\]\[\][[\#2](#2)]{}
R. Bogue [Sens. Rev. [**33**]{}, 300 (2013)](http://www.emeraldinsight.com/doi/full/10.1108/SR-05-2013-678).
D. Rugar, R. Budakian, H. J. Mamin, and B. W. Chui, [Nature [**430**]{}, 329 (2004)](http://dx.doi.org/10.1038/nature02658).
C. L. D | and accurately characterize Newtons ’s constant $ G$, see [ @Schmole2016 ] and references therein. notice that the gravitational interaction between a attraction of $ R\approx5~\text{mm}$ and another celestial sphere of the same bulk separate by a gap of $ 3~\text{mm}$ could be in principle detect. Finally, using our device as an inertial sensor could suffer relevant applications in avionics and space diligence. The detection of small variations of gravitation effect could also be apply to geological exploration or mining, among others.
In conclusion, we have presented an alternate approach for force and inertial sensing base on diamagnetic levitation of magnets. Remarkably, the concept is preferably general and can be use to magnets with sizes ranging from nanometers to millimeters, cross over 6 orders of magnitude. The underlying mechanism behind such an astonishing broad window is the diamagnetic levitation provide by the superconductor in the Meissner state. The use of a magnet with a strong magnetic moment gives rise to a simple passive trapping dodge, and provides direct room to understand and feedback cool its motion. Our analysis, including current technology and realistic assumptions, indicates very promising sensitivity over a wide range of scales, which we hope will motivate its experimental implementation.
This study is supported by the European Research Council (ERC-2013 - StG 335489 QSuperMag) and the Austrian Federal Ministry of Science, Research, and Economy (BMWFW). We acknowledge discussions with J. Hofer, G. Kirchmair, C. Navau, A. Sanchez, and C. Schneider.
[ 27 ] { }
natexlab\#1[\#1]{}bibnamefont \#1[\#1]{}bibfnamefont \#1[\#1]{}citenamefont \#1[\#1]{}url \#1[`#1`]{}urlprefix\[2\][\#2 ] { } \[2\]\[\][[\#2](#2) ] { }
R. Bogue [ Sens. Rev. [ * * 33 * * ] { }, 300 (2013)](http://www.emeraldinsight.com / doi / full/10.1108 / SR-05 - 2013 - 678).
D. Rugar, R. Budakian, H. J. Mamin, and B. W. Chui, [ Nature [ * * 430 * * ] { }, 329 (2004)](http://dx.doi.org/10.1038 / nature02658).
C. L. D | anf accurately characterizt Newtons’s constaur $G$, sex [@Schmols2016] and rewerences therein. Note that tie geavitqtional interaction begween a mwgnet of $R\apkrox5~\text{mm}$ and another spmzre or the wame mass sepatated by a gdp of $3~\text{mm}$ cmuud be in principle detected. Finally, ufing out fevice as an igertpaj sehsor could have relevant applicatikns in evionics and spsce industry. The detection of dmall variations ov gravitatiin fjece could alro be applpzd to geolofical exploration or mining, amovg otkers.
In concouwioj, we have prxsenteq an alternabpve appsoach fpr force and ikertiel swnsing based on diamajnetic levitation of magnets. Sejarkably, the concwpr is tathes gevwrau ahd czn be wppmied to mafnets with wizes ranging from magimeters to mimlimetqrf, spanning over 6 orders of magnitude. Tht undsrlying mechanism behine such an astonishing broad wigdow is the diamagnetic levitation provided by tha supxrzonbmgtor un the Meissner state. The use of a magnet with w surokg magnetic momekt gives rise to a slmlje passive trxpping scgeme, and provides firect rays ro read agd fredback cool its motion. Our analysis, inbludung current technopogies and xealisjic asxumptions, indicates verv promjsing sensihivities kxer a wide range of swales, which we hope will mjtivate ivs ex'erimentxl ikplemegtation.
Thid work is supported by tje Eutopean Research Founcil (ERC-2013-StG 335489 QSuperMag) and tix Austrian Fecesal Ministrv of Sgience, Research, and Economy (BKWFW). We acknuwledge diacussiois with J. Hoser, G. Kirchmaht, C. Navau, A. Senchez, anq C. Wchnwider.
[27]{}
nagdxlab\#1[\#1]{}bibnamefomt \#1[\#1]{}bibfnamefont \#1[\#1]{}citwnamefont \#1[\#1]{}url \#1[`#1`]{}urlpvefix\[2\][\#2]{} \[2\]\[\][[\#2](#2)]{}
D. Bogue [Sens. Rev. [**33**]{}, 300 (2013)](htup://qww.emeraldinsibht.zom/qop/fukl/10.1108/FS-05-2013-678).
D. Rugar, R. Buddkiav, H. G. Kamin, and B. W. Chui, [Vatute [**430**]{}, 329 (2004)](http://dx.doi.org/10.1038/natgre02658).
C. M. D | and accurately characterize Newtons’s constant $G$, see references Note that gravitational interaction between another of the same separated by a of $3~\text{mm}$ could be in principle Finally, using our device as an inertial sensor could have relevant applications in and space industry. The detection of small variations of gravitation force could also applied geological or among others. In conclusion, we have presented an alternative approach for force and inertial sensing based diamagnetic levitation of magnets. Remarkably, the concept is general and can be to magnets with sizes ranging nanometers millimeters, spanning 6 of The underlying mechanism such an astonishing broad window is the diamagnetic levitation provided by the superconductor in the Meissner state. use of with a magnetic gives to a simple scheme, and provides direct ways to cool its motion. Our analysis, including current technologies realistic assumptions, very promising sensitivities over a wide of scales, which we hope will motivate its implementation. This work is supported by the European Research Council (ERC-2013-StG 335489 QSuperMag) and the Ministry of Science, Research, Economy (BMWFW). We discussions J. G. C. Navau, Sanchez, and C. Schneider. [27]{} natexlab\#1[\#1]{}bibnamefont \#1[\#1]{}bibfnamefont \#1[\#1]{}citenamefont \#1[\#1]{}url \#1[`#1`]{}urlprefix\[2\][\#2]{} \[2\]\[\][[\#2](#2)]{} Bogue [Sens. Rev. [**33**]{}, 300 (2013)](http://www.emeraldinsight.com/doi/full/10.1108/SR-05-2013-678). D. Rugar, R. Budakian, Mamin, B. W. Chui, [**430**]{}, 329 (2004)](http://dx.doi.org/10.1038/nature02658). C. D | and accurately characterize newtons’s coNstanT $G$, sEe [@SChMole2016] And rEferences thereIN. NotE that the gravitational iNteraCtIOn beTWeEn a maGnet of $R\APpROX5~\teXt{Mm}$ And AnOThEr sphEre Of the saMe mass sepaRatEd By a gap of $3~\text{MM}$ cOuld be in prIncIple detected. finAlly, usInG ouR DevicE as An ineRtial sENsor coUld have reLeVAnt appLIcationS IN aVionIcs and space industRY. THE detection of smAll varIaTIoNS Of gRavItation forCe Could ALso be apPLiED TO geOLogical explorAtion or miniNG, amOng othErS.
In COnclusIon, we HaVE prEsented an alTernAtive apprOach foR Force anD InertiaL sensiNg bAseD on dIAmAgNetIc LEviTAtIon OF maGnets. RemArKaBly, thE conCEPT Is raTheR genEral aNd can be applieD to MagnETs wIth siZes raNginG fRom naNometeRs to mIlLimeters, spanninG oveR 6 orders of MagNiTudE. THe undERlying MecHanIsm behiNd such aN AstOnISHInG broad window is the dIaMAGnEtic leviTation PRoViDEd by the sUpErcOnduCTOr in tHe MeISsNer state. the use OF a MaGnet witH a Strong MaGneTic MomenT GiveS rise tO a simple PassiVE trapping schemE, And provides diREcT WAyS To reAd aNd feedback cOol iTS motIon. OUR aNalYSis, inCludiNg CUrREnt technologies and rEaListic AssumPtions, indicatEs very promISINg sensitIvitIEs OVer a wide range oF scalEs, which we hOPe will moTivatE its expeRimental iMPLementatIon.
thiS woRk iS SUpPorted by the EuROPean reSearch COunCil (ERC-2013-STG 335489 QsupErMAg) aNd The AustriAn FederaL MInIsTrY of scienCE, ResearcH, aNd ECoNomY (BMWFw). we acknOwledGe diScUsSIonS with J. HOFeR, g. kircHmAiR, C. NaVau, a. SAncheZ, and c. schNeider.
[27]{}
nAtexlab\#1[\#1]{}biBnaMEfonT \#1[\#1]{}bIbFnamefoNt \#1[\#1]{}citenamefonT \#1[\#1]{}uRl \#1[`#1`]{}urlprefiX\[2\][\#2]{} \[2\]\[\][[\#2](#2)]{}
R. bogUe [Sens. rEV. [**33**]{}, 300 (2013)](http://www.Emeraldinsight.com/doi/fulL/10.1108/sR-05-2013-678).
D. RugaR, R. BUdakiAn, H. J. mamin, and B. w. ChUi, [NatuRe [**430**]{}, 329 (2004)](hTTp://dx.doI.org/10.1038/naTure02658).
C. l. D | and accurately characteri ze Newtons ’s co nst ant $ G$,see[@Schmole2016] andreferences therein. No te th at theg ra vitat ional i n te r a cti on b etw ee n a magn etof $R\a pprox5~\te xt{ mm }$ and anoth e rsphere ofthe same mass s epa ratedby ag ap of $3 ~\tex t{mm}$ couldbe in pri nc i ple de t ected.F i na lly, using our device as an inertial se nsor c ou l dh a verel evant appl ic ation s in avi o ni c s and space industr y. The dete c tio n of s ma llv ariati ons o fg rav itation for ce c ould also be ap p lied to geologi cal ex plo rat iono rmi nin g, amo n goth e rs.
In con cl us ion,we h a v e pres ent ed a n alt ernative appr oac h fo r fo rce a nd in erti al sens ing ba sed o ndiamagnetic lev itat ion of ma gne ts . R em arkab l y, the co nce pt is r ather g e ner al a n dcan be applied toma g n et s with s izes r a ng in g from na no met erst o mill imet e rs , spanni ng ove r 6 o rders o fmagnit ud e.The unde r lyin g mech anism be hinds uch an astonis h ing broad win d ow i st he d iam agnetic lev itat i on p rovi d ed by the s uperc on d uc t or in the Meissnerst ate. T he us e of a magnet with a st r o n g magnet ic m o me n t gives rise t o a s imple pass i ve trapp ing s cheme, a nd provid e s directway s t o r ead a nd feedback coo l itsmo tion. O uranalysi s,inc lud ing c urrent te chnologi es a nd r eal istic assumpti on s,in dic atesv ery pr omisi ng s en si t ivi ties ov e ra wide r an ge o f s ca les,whic h we hope w ill motiv ate itsex pe rimenta l implementat io n.
This w or k i s supp o r ted by t he European Research Co u ncil (E RC- 2013- StG335489 QS upe rMag)and the Au strian Fede ra l M i n istry o fSci en ce, Resear c h , a nd Ec on omy(BMWFW) . We acknowledge d i scu ssions with J . H ofer , G. Ki r ch m air ,C . N a v au, A. Sanchez, and C. Sc hn e id er.
[27]{ }
n at exlab\# 1[\#1]{ }bibn a mefont\#1[\#1]{ }bibfname fo nt \ # 1 [\# 1]{}citena mefont \ #1[\#1]{} u rl \# 1 [` #1`]{ }ur lprefi x\ [2\ ][\#2 ]{} \[ 2 \]\ [\][[ \#2](# 2) ]{}
R . Bog ue [Sens.Rev. [**33**]{}, 300 (2 013)]( http: //w ww.emeral din s igh t.com/doi /ful l/10.1108/ SR- 05- 2013- 678 ) .
D. Rug a r, R. Budak ian, H. J. Mam i n, an d B. W. Chui, [ N a t ure [**4 30* * ]{}, 3 29 ( 2004)](http://dx. d oi.org/10.1038 /nat u r e02 658 ) .
C .L. D | and_accurately characterize_Newtons’s constant $G$, see [@Schmole2016]_and references_therein._Note that_the_gravitational interaction between_a magnet of_$R\approx5~\text{mm}$ and another sphere_of the same_mass_separated by a gap of $3~\text{mm}$ could be in principle detected. Finally, using our_device_as an_inertial_sensor_could have relevant applications in_avionics and space industry. The_detection of_small variations of gravitation force could also be_applied_to geological exploration_or mining, among others.
In conclusion, we have presented an_alternative approach for force and inertial_sensing based on_diamagnetic_levitation_of magnets. Remarkably, the_concept is rather general and can_be applied to magnets with sizes_ranging from nanometers to millimeters, spanning over_6 orders of magnitude. The underlying_mechanism behind such an astonishing_broad window_is the diamagnetic levitation provided_by the superconductor_in the_Meissner state. The_use of a magnet with a_strong magnetic moment_gives rise to a simple passive_trapping_scheme, and provides_direct_ways_to read_and feedback cool_its_motion. Our_analysis,_including current technologies and realistic assumptions,_indicates_very promising sensitivities over a wide range_of scales, which we_hope_will motivate its experimental_implementation.
This work is supported by_the European Research Council (ERC-2013-StG 335489_QSuperMag) and_the Austrian_Federal Ministry of Science, Research, and Economy (BMWFW). We acknowledge discussions_with J. Hofer, G. Kirchmair, C._Navau, A. Sanchez, and_C. Schneider.
[27]{}
natexlab\#1[\#1]{}bibnamefont_\#1[\#1]{}bibfnamefont_\#1[\#1]{}citenamefont \#1[\#1]{}url \#1[`#1`]{}urlprefix\[2\][\#2]{}_\[2\]\[\][[\#2](#2)]{}
R. Bogue_[Sens. Rev. [**33**]{}, 300_(2013)](http://www.emeraldinsight.com/doi/full/10.1108/SR-05-2013-678).
D. Rugar, R. Budakian, H. J. Mamin, and B. W. Chui, [Nature [**430**]{},_329 (2004)](http://dx.doi.org/10.1038/nature02658).
C. L. D |
\leq T'$ is derived in the pure system of Def. \[mmb04\].
.5em For distinction, let us write $T\leq_1T'$ if we use the system including reflexivity and transitivity. We define the third relation $\leq_2$. First $\bigcup_{i\in I}S_i\leq_2\bigcup_{j\in J}S'_j$ if there is a mapping $I\mor fJ$ such that $S_i\leq_2S'_{f(i)}$ for every $i$. Second $\bigcap_{i\in I}R_i\leq_2\bigcap_{j\in J}R'_j$ if there is a mapping $J\mor fI$ such that $R_{f(j)}\leq_2R'_j$ for every $j$. Finally $R\leq_2R'$ if both $R$ and $R'$ are the same atomic type or $R=S\rightarrow T$ and $R'=S'\rightarrow T'$ with $S'\leq_2S$ and $T\leq_2T'$. Now it is easy to verify that $T\leq_1T'$ implies $T\leq_2T'$ by induction on the construction of the derivation trees. It is also easy to show that $T\leq_2T'$ implies $T\leq T'$. (As a consequence, all of $\leq,\leq_1$ and $\leq_2$ turn out to be the same.)
3ex
\[seh62\] We give the inference rules of typing judgements. A [*typing judgement*]{} has the form $\Gamma\;\vdash\;M\mathbin:T\;\mathbin|
\;\Delta$. Here the typing environment $\Gamma$ is a finite sequence of $x_i\mathbin:S_i$, and the typing environment $\Delta$ is a finite sequence of $k_j\mathbin:T_j$. As usual, we assume that all variables in the typing environments are distinct from each other. We stress that only subsidiary types are assigned to ordinary variables while continuation variables have no limitation. We assume a special type $\dbot$ for typing jumps. The following are inference rules:
2ex 5em
2ex 5em
2ex 5em
2ex 5em
2ex 5em | \leq T'$ is derived in the pure system of Def. \[mmb04\ ].
.5em For distinction, lease us spell $ T\leq_1T'$ if we use the system including reflexivity and transitivity. We specify the third relation $ \leq_2$. First $ \bigcup_{i\in I}S_i\leq_2\bigcup_{j\in J}S'_j$ if there be a function $ I\mor fJ$ such that $ S_i\leq_2S'_{f(i)}$ for every $ i$. Second $ \bigcap_{i\in I}R_i\leq_2\bigcap_{j\in J}R'_j$ if there is a mapping $ J\mor fI$ such that $ R_{f(j)}\leq_2R'_j$ for every $ j$. Finally $ R\leq_2R'$ if both $ R$ and $ R'$ are the same atomic character or $ R = S\rightarrow T$ and $ R'=S'\rightarrow T'$ with $ S'\leq_2S$ and $ T\leq_2T'$. Now it is easy to verify that $ T\leq_1T'$ incriminate $ T\leq_2T'$ by induction on the structure of the derivation tree. It is besides easy to show that $ T\leq_2T'$ implies $ T\leq T'$. (As a consequence, all of $ \leq,\leq_1 $ and $ \leq_2 $ become out to be the same .)
3ex
\[seh62\ ] We give the inference rule of type judgements. A [ * type judgement * ] { } has the form $ \Gamma\;\vdash\;M\mathbin: T\;\mathbin|
\;\Delta$. Here the typing environment $ \Gamma$ is a finite succession of $ x_i\mathbin: S_i$, and the typing environment $ \Delta$ is a finite sequence of $ k_j\mathbin: T_j$. As usual, we assume that all variables in the typing environment are distinct from each other. We stress that only subsidiary types are assigned to ordinary variables while continuation variable have no limitation. We wear a special type $ \dbot$ for typing jumps. The follow are inference rules:
2ex 5em
2ex 5em
2ex 5em
2ex 5em
2ex 5em | \leq T'$ is derived in the purt system of Def. \[mmy04\].
.5wm For distihction, ldt us write $T\leq_1T'$ if we use vhe wysten including reflexivith and trajsitivitt. We eefine the third rcjatikk $\leq_2$. Hirst $\bigcup_{i\in I}S_i\leq_2\bigwup_{j\in J}S'_j$ if dhdrz is a mapping $I\mor fJ$ such that $S_i\lqq_2S'_{f(i)}$ fpr every $i$. Seconq $\bibsap_{i\jn I}R_i\leq_2\bigcap_{j\in J}R'_j$ if there is a mapppng $J\mor fI$ such yhat $R_{f(j)}\leq_2R'_j$ for every $j$. Vinaply $R\leq_2R'$ if both $G$ and $R'$ are the wame atomic gype or $R=S\gnghtarrow T$ and $R'=S'\rightarrow T'$ with $S'\leq_2S$ xnd $T\keq_2T'$. Now ij js gasy to verihy than $T\leq_1T'$ implicx $T\leq_2D'$ by incuction on the coistryction of the derivatmon trees. It is also easy to vhkw that $T\leq_2T'$ impoiws $T\lgq T'$. (Ds a xonreqheice, all ov $\lxq,\leq_1$ and $\lsq_2$ turn out to be the same.)
3ex
\[sen62\] Rv give the inrerencq wules of typing judgements. A [*typing judcemsnt*]{} has the form $\Gamma\;\veash\;M\mathbin:T\;\mathbin|
\;\Felta$. Herq the typing environment $\Gamma$ is a finite sequenwe of $b_i\mcbmbin:R_u$, wnd the typing environment $\Delta$ is a finite fsqiekce of $k_j\mathbin:B_j$. As usual, we assimf yrat all variacles iu tge typing environmfnts arg distunct from eacn other. We stress that only subsidiary nypew are assigned to lrdinary vaxiablex whike continuation variablzs havs no limitahion. We aarume a special thpe $\dtot$ for uhping jumps. The fjllowing ere iuference rulgs:
2ex 5em
2qx 5em
2ex 5em
2fx 5em
2cf 5em | \leq T'$ is derived in the pure Def. .5em For let us write system reflexivity and transitivity. define the third $\leq_2$. First $\bigcup_{i\in I}S_i\leq_2\bigcup_{j\in J}S'_j$ if is a mapping $I\mor fJ$ such that $S_i\leq_2S'_{f(i)}$ for every $i$. Second $\bigcap_{i\in J}R'_j$ if there is a mapping $J\mor fI$ such that $R_{f(j)}\leq_2R'_j$ for every Finally if $R$ $R'$ are the same atomic type or $R=S\rightarrow T$ and $R'=S'\rightarrow T'$ with $S'\leq_2S$ and $T\leq_2T'$. it is easy to verify that $T\leq_1T'$ implies by induction on the of the derivation trees. It also to show $T\leq_2T'$ $T\leq (As a consequence, of $\leq,\leq_1$ and $\leq_2$ turn out to be the same.) 3ex \[seh62\] We give the inference rules typing judgements. judgement*]{} has form \;\Delta$. the typing environment a finite sequence of $x_i\mathbin:S_i$, and $\Delta$ is a finite sequence of $k_j\mathbin:T_j$. As we assume all variables in the typing environments distinct from each other. We stress that only types are assigned to ordinary variables while continuation variables have no limitation. We assume a $\dbot$ for typing jumps. following are inference 2ex 2ex 2ex 2ex 5em 5em | \leq T'$ is derived in the pure sysTem of Def. \[mmB04\].
.5em FoR diStiNcTion, Let uS write $T\leq_1T'$ if wE Use tHe system including refleXivitY aND traNSiTivitY. We defiNE tHE ThiRd ReLatIoN $\LeQ_2$. FirsT $\biGcup_{i\in i}S_i\leq_2\bigcUp_{j\In j}S'_j$ if there is A MaPping $I\mor fj$ suCh that $S_i\leq_2S'_{F(i)}$ fOr everY $i$. secONd $\bigCap_{I\in I}R_I\leq_2\biGCap_{j\in j}R'_j$ if therE iS A mappiNG $J\mor fI$ SUCh That $r_{f(j)}\leq_2R'_j$ for every $j$. fInALly $R\leq_2R'$ if both $r$ and $R'$ aRe THe SAMe aTomIc type or $R=S\RiGhtarROw T$ and $R'=s'\RiGHTArrOW T'$ with $S'\leq_2S$ anD $T\leq_2T'$. Now it IS eaSy to veRiFy tHAt $T\leq_1t'$ implIeS $t\leQ_2T'$ by inductiOn on The constrUction OF the derIVation tRees. It Is aLso Easy TO sHoW thAt $t\Leq_2t'$ ImPliES $T\lEq T'$. (As a coNsEqUence, All oF $\LEQ,\Leq_1$ aNd $\lEq_2$ tuRn out To be the same.)
3ex
\[Seh62\] we giVE thE infeRence RuleS oF typiNg judgEmentS. A [*Typing judgement*]{} Has tHe form $\GamMa\;\vDaSh\;M\MaThbin:t\;\MathbiN|
\;\DeLta$. here the Typing eNVirOnMENT $\GAmma$ is a finite sequeNcE OF $x_I\mathbin:s_i$, and tHE tYpINg enviroNmEnt $\deltA$ IS a finIte sEQuEnce of $k_j\MathbiN:t_j$. as Usual, we AsSume thAt All VarIableS In thE typinG environMents ARe distinct from EAch other. We strESs THAt ONly sUbsIdiary types Are aSSignEd to ORdInaRY variAbles WhILe COntinuation variableS hAve no lImitaTion. We assume a Special typE $\DBOt$ for typIng jUMpS. the following arE infeRence rules:
2EX 5em
2ex 5em
2eX 5em
2ex 5Em
2ex 5em | \leq T'$ is derived in the pure syst em of De f.\[ mmb0 4\].
.5em For dis t inct ion, let us write $T\l eq_1T '$ if w e u se th e syste m i n c lud in gref le x iv ity a ndtransit ivity. Wedef in e the thirdr el ation $\le q_2 $. First $\b igc up_{i\ in I} S _i\le q_2 \bigc up_{j\ i n J}S' _j$ if th er e is am apping$ I \m or f J$ such that $S_i \ le q _2S'_{f(i)}$ f or eve ry $i $ . Se con d $\bigcap _{ i\inI }R_i\le q _2 \ b i gca p _{j\in J}R'_j $ if therei s a mappi ng $J \ mor fI $ suc ht hat $R_{f(j)}\ leq_ 2R'_j$ fo r ever y $j$. F i nally $ R\leq_ 2R' $ i f bo t h$R $ a nd $R' $ a ret hesame ato mi ctypeor $ R = S \ righ tar rowT$ an d $R'=S'\righ tar rowT '$with$S'\l eq_2 S$ and$T\leq _2T'$ .Now it is easyto v erify tha t $ T\ leq _1 T'$ i m plies$T\ leq _2T'$ b y induc t ion o n t he construction of t he d er ivationtrees. It i s also ea sy to sho w that$T\l e q_ 2T'$ imp lies $ T \l eq T'$. ( As a con se que nce , all of $ \leq,\ leq_1$ a nd $\ l eq_2$ turn out to be the sam e .) 3e x
\[ seh 62\] We giv e th e inf eren c erul e s oftypin gj ud g ements. A [*typingju dgemen t*]{} has the form $\Gamma\; \ v d ash\;M\m athb i n: T \;\mathbin|
\ ;\Del ta$. Heret he typin g env ironment $\Gamma$ i s a fini teseq uen ceo f $ x_i\mathbin:S _ i $, a nd the ty pin g envir onm ent $\ Del ta $ is a fi nite seq ue nc eof $k _j\ma t hbin:T_j $. As u sua l, we assume that all v ar i abl es in t h et y ping e nv iron men ts aredist i nct from e ach other . W e str es sthat on ly subsidiary t ypes are a ss ign ed too r dinary v ariables while continua t ion var iab les h aveno limita tio n. Weass u me a s pecial type $ \db o t $ for t yp ing j umps. Thef o llo wingar e in ference rules:
2ex 5em
2 ex5em
2ex 5em
2e x 5e m
2 ex5 em | \leq T'$_is derived_in the pure system_of Def. \[mmb04\].
.5em_For_distinction, let_us_write $T\leq_1T'$ if_we use the_system including reflexivity and_transitivity. We define_the_third relation $\leq_2$. First $\bigcup_{i\in I}S_i\leq_2\bigcup_{j\in J}S'_j$ if there is a mapping $I\mor fJ$_such_that $S_i\leq_2S'_{f(i)}$_for_every_$i$. Second $\bigcap_{i\in I}R_i\leq_2\bigcap_{j\in J}R'_j$_if there is a mapping_$J\mor fI$_such that $R_{f(j)}\leq_2R'_j$ for every $j$. Finally $R\leq_2R'$_if_both $R$ and_$R'$ are the same atomic type or $R=S\rightarrow T$_and $R'=S'\rightarrow T'$ with $S'\leq_2S$ and_$T\leq_2T'$. Now it_is_easy_to verify that $T\leq_1T'$_implies $T\leq_2T'$ by induction on the_construction of the derivation trees. It_is also easy to show that $T\leq_2T'$_implies $T\leq T'$. (As a consequence,_all of $\leq,\leq_1$ and $\leq_2$_turn out_to be the same.)
3ex
\[seh62\] We_give the inference_rules of_typing judgements. A_[*typing judgement*]{} has the form $\Gamma\;\vdash\;M\mathbin:T\;\mathbin|
_\;\Delta$. Here the_typing environment $\Gamma$ is a finite_sequence_of $x_i\mathbin:S_i$, and_the_typing_environment $\Delta$_is a finite_sequence_of $k_j\mathbin:T_j$._As_usual, we assume that all variables_in_the typing environments are distinct from each_other. We stress that_only_subsidiary types are assigned_to ordinary variables while continuation_variables have no limitation. We assume_a special_type $\dbot$_for typing jumps. The following are inference rules:
2ex 5em
2ex 5em
2ex 5em
2ex_5em
2ex 5em |
-\sigma(t,u^0(t))|^2_{L_Q}dt.
\end{aligned}$$ Defining $\tau_N=\inf\{t:|V^\e(t)-V^0(t)|^2+\int_0^t\|V^\e(s)-V^0(s)\|^2ds>N\}$, we have $$\begin{aligned}
\label{II0}
&|V^\e(t\wedge\tau_N)-V^0(t\wedge\tau_N)|^2+2 \int_0^{t\wedge\tau_N}\|V^\e(s)-V^0(s)\|^2ds\notag\\
\le& 2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s)-V^0(s), u^0(s)), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s),u^\e(s)-u^0(s)), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\int_0^{t\wedge\tau_N}\left|\Big(\frac{1}{\sqrt{\e}}[\tilde{R}(s,u^\e(s))-\tilde{R}(s,u^0(s))]
-\tilde{R}'(s,u^0(s)) V^0(s), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\left|\int_0^{t\wedge\tau_N}\Big([\sigma(s,u^\e(s))-\sigma(s,u^0(s))]dW(s), V^\e(s)-V^0(s)\Big)\right|\notag\\
&+\int_0^{t\wedge\tau_N}|\sigma(s,u^\e(s))-\sigma(s,u^0(s))|^2_{L_Q}ds\notag\\
=:&\sum\limits^{5}_{k=1}I_k.\end{aligned}$$
Because of and, we have | -\sigma(t, u^0(t))|^2_{L_Q}dt.
\end{aligned}$$ Defining $ \tau_N=\inf\{t:|V^\e(t)-V^0(t)|^2+\int_0^t\|V^\e(s)-V^0(s)\|^2ds > N\}$, we have $ $ \begin{aligned }
\label{II0 }
& |V^\e(t\wedge\tau_N)-V^0(t\wedge\tau_N)|^2 + 2 \int_0^{t\wedge\tau_N}\|V^\e(s)-V^0(s)\|^2ds\notag\\
\le & 2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s)-V^0(s), u^0(s) ), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
& +2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s),u^\e(s)-u^0(s) ), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
& +2\int_0^{t\wedge\tau_N}\left|\Big(\frac{1}{\sqrt{\e}}[\tilde{R}(s, u^\e(s))-\tilde{R}(s, u^0(s) ) ]
-\tilde{R}'(s, u^0(s) ) V^0(s), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
& +2\left|\int_0^{t\wedge\tau_N}\Big([\sigma(s, u^\e(s))-\sigma(s, u^0(s))]dW(s), V^\e(s)-V^0(s)\Big)\right|\notag\\
& + \int_0^{t\wedge\tau_N}|\sigma(s, u^\e(s))-\sigma(s, u^0(s))|^2_{L_Q}ds\notag\\
=: & \sum\limits^{5}_{k=1}I_k.\end{aligned}$$
Because of and, we have | -\sigla(t,u^0(t))|^2_{L_Q}dt.
\end{aligned}$$ Deflning $\tau_N=\inf\{t:|V^\e(j)-V^0(r)|^2+\int_0^t\|V^\x(s)-V^0(s)\|^2ds>N\}$, we have $$\begin{aligned}
\label{II0}
&|V^\e(t\wedge\vau_N)-C^0(t\wedte\tau_N)|^2+2 \int_0^{t\wedge\tau_N}\|V^\e(r)-V^0(s)\|^2ds\notah\\
\le& 2\int_0^{t\qedgt\tau_N}\left|\Big(B(V^\e(s)-V^0(s), u^0(s)), V^\e(s)-Y^0(f)\Big)\dlght|dv\iotag\\
&+2\int_0^{t\wedge\tsu_N}\left|\Big(T(V^\e(s),u^\e(s)-u^0(s)), V^\e(s)-V^0(v)\Bkg)\xight|ds\notag\\
&+2\int_0^{t\wedge\tau_N}\left|\Big(\frac{1}{\fqrt{\e}}[\tikdf{R}(s,u^\e(s))-\tilde{R}(s,u^0(f))]
-\tilcq{R}'(s,u^0(a)) V^0(s), V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\left|\int_0^{t\wsdge\tau_I}\Big([\sigma(s,u^\e(s))-\sibma(s,u^0(s))]dW(s), V^\e(s)-V^0(s)\Big)\right|\notwg\\
&+\inh_0^{t\wedge\tau_N}|\sigma(s,u^\f(s))-\sigma(s,u^0(s))|^2_{L_W}ds\njrag\\
=:&\sum\limitr^{5}_{k=1}I_k.\end{aligned}$$
Because kf and, we have | -\sigma(t,u^0(t))|^2_{L_Q}dt. \end{aligned}$$ Defining $\tau_N=\inf\{t:|V^\e(t)-V^0(t)|^2+\int_0^t\|V^\e(s)-V^0(s)\|^2ds>N\}$, we have $$\begin{aligned} \int_0^{t\wedge\tau_N}\|V^\e(s)-V^0(s)\|^2ds\notag\\ 2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s)-V^0(s), u^0(s)), &+2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s),u^\e(s)-u^0(s)), V^\e(s)-V^0(s)\Big)\right|ds\notag\\ &+2\int_0^{t\wedge\tau_N}\left|\Big(\frac{1}{\sqrt{\e}}[\tilde{R}(s,u^\e(s))-\tilde{R}(s,u^0(s))] &+\int_0^{t\wedge\tau_N}|\sigma(s,u^\e(s))-\sigma(s,u^0(s))|^2_{L_Q}ds\notag\\ Because of and, have | -\sigma(t,u^0(t))|^2_{L_Q}dt.
\end{aligned}$$ DefIning $\tau_N=\iNf\{t:|V^\e(T)-V^0(t)|^2+\Int_0^T\|V^\E(s)-V^0(s)\|^2Ds>N\}$, wE have $$\begin{aligNEd}
\laBel{II0}
&|V^\e(t\wedge\tau_N)-V^0(t\wedGe\tau_n)|^2+2 \iNT_0^{t\weDGe\Tau_N}\|V^\E(s)-V^0(s)\|^2ds\nOTaG\\
\LE& 2\inT_0^{t\WeDge\TaU_n}\lEft|\BiG(B(V^\E(s)-V^0(s), u^0(s)), V^\E(s)-V^0(s)\Big)\rigHt|dS\nOtag\\
&+2\int_0^{t\wedgE\TaU_N}\left|\Big(B(v^\e(s),U^\e(s)-u^0(s)), V^\e(s)-V^0(s)\BiG)\riGht|ds\nOtAg\\
&+2\iNT_0^{t\wedGe\tAu_N}\leFt|\Big(\fRAc{1}{\sqrt{\E}}[\tilde{R}(s,u^\E(s))-\TIlde{R}(s,U^0(S))]
-\tilde{R}'(S,U^0(S)) V^0(S), V^\e(s)-v^0(s)\Big)\right|ds\notag\\
&+2\LEfT|\Int_0^{t\wedge\tau_N}\BIg([\sigmA(s,U^\E(s))-\SIGma(S,u^0(s))]DW(s), V^\e(s)-V^0(s)\BiG)\rIght|\nOTag\\
&+\int_0^{t\WEdGE\TAu_N}|\SIgma(s,u^\e(s))-\sigma(S,u^0(s))|^2_{L_Q}ds\notaG\\
=:&\Sum\Limits^{5}_{K=1}I_K.\enD{AligneD}$$
BecaUsE Of aNd, we have | -\sigma(t,u^0(t))|^2_{L_Q} dt.
\end{ align ed} $$De fini ng $ \tau_N=\inf\{t : |V^\ e(t)-V^0(t)|^2+\int_0^ t\|V^ \e ( s)-V ^ 0( s)\|^ 2ds>N\} $ ,w e ha ve $ $\b eg i n{ align ed}
\label {II0}
&|V^ \e( t\ wedge\tau_N) - V^ 0(t\wedge\ tau _N)|^2+2 \in t_0 ^{t\we dg e\t a u_N}\ |V^ \e(s) -V^0(s ) \|^2ds \notag\\\l e & 2\in t _0^{t\w e d ge \tau _N}\left|\Big(B(V ^ \e ( s)-V^0(s), u^0 (s)),V^ \ e( s ) -V^ 0(s )\Big)\rig ht |ds\n o tag\\
& + 2\ i n t _0^ { t\wedge\tau_N }\left|\Big ( B(V ^\e(s) ,u ^\e ( s)-u^0 (s)), V ^ \e( s)-V^0(s)\B ig)\ right|ds\ notag\ \
&+2\in t _0^{t\w edge\t au_ N}\ left | \B ig (\f ra c {1} { \s qrt { \e} }[\tilde {R }( s,u^\ e(s) ) - \ t ilde {R} (s,u ^0(s) )]
-\tilde{R} '(s ,u^0 ( s)) V^0( s), V ^\e( s) -V^0( s)\Big )\rig ht |ds\notag\\
&+2 \lef t|\int_0^ {t\ we dge \t au_N} \ Big([\ sig ma( s,u^\e( s))-\si g ma( s, u ^ 0 (s ))]dW(s), V^\e(s)- V^ 0 ( s) \Big)\ri ght|\n o ta g\ \
&+\int_ 0^ {t\ wedg e \ tau_N }|\s i gm a(s,u^\e (s))-\ s ig ma (s,u^0( s) )|^2_{ L_ Q}d s\n otag\ \
=: &\sum\ limits^{ 5}_{k = 1}I_k.\end{ali g ned}$$
Becau s eo f a n d, w e h ave | -\sigma(t,u^0(t))|^2_{L_Q}dt.
\end{aligned}$$_Defining $\tau_N=\inf\{t:|V^\e(t)-V^0(t)|^2+\int_0^t\|V^\e(s)-V^0(s)\|^2ds>N\}$,_we have $$\begin{aligned}
\label{II0}
&|V^\e(t\wedge\tau_N)-V^0(t\wedge\tau_N)|^2+2 \int_0^{t\wedge\tau_N}\|V^\e(s)-V^0(s)\|^2ds\notag\\
\le&_2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s)-V^0(s), u^0(s)),_V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\int_0^{t\wedge\tau_N}\left|\Big(B(V^\e(s),u^\e(s)-u^0(s)),_V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\int_0^{t\wedge\tau_N}\left|\Big(\frac{1}{\sqrt{\e}}[\tilde{R}(s,u^\e(s))-\tilde{R}(s,u^0(s))]
-\tilde{R}'(s,u^0(s)) V^0(s),_V^\e(s)-V^0(s)\Big)\right|ds\notag\\
&+2\left|\int_0^{t\wedge\tau_N}\Big([\sigma(s,u^\e(s))-\sigma(s,u^0(s))]dW(s),_V^\e(s)-V^0(s)\Big)\right|\notag\\
&+\int_0^{t\wedge\tau_N}|\sigma(s,u^\e(s))-\sigma(s,u^0(s))|^2_{L_Q}ds\notag\\
=:&\sum\limits^{5}_{k=1}I_k.\end{aligned}$$
Because of_and, we have |
vanishing Euler characteristic too and it is noncyclic since the only virtually cyclic torsion-free group is $\mathbb Z$, yet $H$ is not cyclic.
\[nongeomcomm\]Suppose $\phi: F \to F$ and $\psi:F' \to F'$ are minimal injective endomorphisms such that $F *_\phi$ and $F' *_\psi$ are commensurable. Then $\phi$ is irreducible and atoroidal if and only if $\psi$ is irreducible and atoroidal.
Finally, we combine this theorem with Corollary \[irredtorgrp\] to get:
\[irredgrp\]Suppose $\phi: F \to F$ and $\psi:F' \to F'$ are minimal injective endomorphisms such that $F *_\phi \cong F' *_\psi$. Then $\phi$ is irreducible if and only if $\psi$ is irreducible.
Q.I.-Invariance Conjecture {#qi}
==========================
Very little is known about the geometry of mapping tori whose monodromies are irreducible and atoroidal. Brinkmann proved that atoroidal automorphisms of free groups have word-hyperbolic mapping tori [@Bri00] and gave an explicit description for when such mapping tori have non-trivial splittings over cyclic subgroups [@Bri02]: an atoroidal automorphism of $F$ has a mapping torus that splits over $\mathbb Z$ if and only if the (outer) automorphism fixes a [*free splitting*]{} of $F$; in particular, mapping tori of irreducible and atoroidal automorphisms never split over $\mathbb Z$.
By work of Kapovich-Kleiner [@KK00], word-hyperbolic free-by-cyclic groups that do not split over $\mathbb Z$ have Gromov boundary homeomorphic to a menger curve. Unfortunately, the topology of the boundary is not sufficient to detect irreducible and atoroidal monodromies as there are reducible and atoroidal automorphisms that do not fix a free splitting.
The Gromov boundary of a word-hyperbolic group can be given a [*visual metric*]{} that is unique up to [*quasi-symmetry*]{} and its quasi-symmetry class is a complete geometric invariant of the group, i.e., word-hyperbolic groups are quasi-isometric if and only their visual boundaries are quasi-symmetric [@P | vanishing Euler characteristic too and it is noncyclic since the only virtually cyclic tortuosity - spare group is $ \mathbb Z$, yet $ H$ is not cyclic.
\[nongeomcomm\]Suppose $ \phi: F \to F$ and $ \psi: F' \to F'$ are minimal injective endomorphisms such that $ F * _ \phi$ and $ F' * _ \psi$ are commensurable. Then $ \phi$ is irreducible and atoroidal if and merely if $ \psi$ is irreducible and atoroidal.
last, we combine this theorem with Corollary \[irredtorgrp\ ] to get:
\[irredgrp\]Suppose $ \phi: F \to F$ and $ \psi: F' \to F'$ are minimal injective endomorphisms such that $ F * _ \phi \cong F' * _ \psi$. Then $ \phi$ is irreducible if and entirely if $ \psi$ is irreducible.
Q.I.-Invariance guess { # qi }
= = = = = = = = = = = = = = = = = = = = = = = = = =
Very little is know about the geometry of mapping tori whose monodromies are irreducible and atoroidal. Brinkmann testify that atoroidal automorphisms of free groups have word - hyperbolic mapping tori [ @Bri00 ] and contribute an explicit description for when such mapping tori have non - superficial splittings over cyclic subgroups [ @Bri02 ]: an atoroidal automorphism of $ F$ has a mapping torus that splits over $ \mathbb Z$ if and merely if the (outer) automorphism fixes a [ * free splitting * ] { } of $ F$; in particular, mapping torus of irreducible and atoroidal automorphisms never split over $ \mathbb Z$.
By work of Kapovich - Kleiner [ @KK00 ], word - hyperbolic free - by - cyclic groups that do not split over $ \mathbb Z$ have Gromov boundary homeomorphic to a menger curve. Unfortunately, the topology of the boundary is not sufficient to detect irreducible and atoroidal monodromies as there are reducible and atoroidal automorphisms that do not fix a free splitting.
The Gromov boundary of a news - hyperbolic group can be given a [ * visual measured * ] { } that is alone up to [ * quasi - isotropy * ] { } and its quasi - symmetry class is a complete geometric invariant of the group, i.e., word - hyperbolic groups are quasi - isometric if and only their visual boundaries are quasi - symmetrical [ @P | vajishing Euler characterirtic too and it is nonryclic aince thd only virtually cyclic torsmon-feee geoup is $\mathbb Z$, yet $H$ is not cjclic.
\[nongwomcinm\]Suppose $\'gi: F \to F$ and $\isi:F' \co F'$ are minimal injective endomorphisms sjck that $F *_\phi$ and $F' *_\psi$ are commensurwble. Thrn $\phi$ is irredusiblt agd aforoidal if and only if $\psi$ is irrsduciblt and atoroidal.
Finslly, we combine this theorfm wlth Corollary \[irredhorgrp\] to ggf:
\[irwwdgrp\]Suppose $\phi: F \to F$ and $\psi:F' \tk F'$ are minimal injective endomurphixms such tyar $F *_\khi \cong F' *_\pwi$. Thvn $\phi$ is irrccucibla if anc only if $\psi$ ls irceduxible.
Q.I.-Invariance Conoecture {#qi}
==========================
Very littlg is known ayout the geometry of napping torh whuwe oonkdcomjes arf icreducible znd atoroidql. Brinkmann proved trqt atoroidal zutomowprisms of free groups have word-hyperbolib malping tori [@Bri00] and gave an explicit descriptlon for wren such mapping tori have non-trivial splittings mver ryzlie subgfiuos [@Bri02]: an atoroidal automorphism of $F$ has a mwlponn torus that spllts over $\mathbb Z$ of amq only if the (outer) aufomorphism fixes a [*free sklittibg*]{} of $F$; ig patticular, mapping tori of ireeducible anb aroroidal automorphnsms never s'lit ofer $\mayhbb Z$.
By work of Kapovieh-Kleiher [@KK00], word-jyperbolid free-by-cyclic gruupx dhat do not split over $\mathfb Z$ have Grokov bouvdari homeoiorphic to a mekcer curve. Unfortunwtely, tve topologj of the boundary is not sufficmxnt to detect isregucible cnd atproidal monodwomies as therg are redbcible and atoropdal autokorphisms trat do not fif a free splivting.
The Dromiv biundary uf a word-hypernolic grolp can be given a [*visual mejrjc*]{} that is uniqbt uk ro [*quasi-symmetty*]{} xnd ins auwvi-symmetry cnass is s comolete gtjmctrkc imvariant of the grou[, i.e., word-hyperbolic grpuis are quqsi-isomeeric if and omly their visual blundacies ace quaxi-simmetric [@P | vanishing Euler characteristic too and it is the virtually cyclic group is $\mathbb cyclic. $\phi: F \to and $\psi:F' \to are minimal injective endomorphisms such that *_\phi$ and $F' *_\psi$ are commensurable. Then $\phi$ is irreducible and atoroidal if only if $\psi$ is irreducible and atoroidal. Finally, we combine this theorem with \[irredtorgrp\] get: $\phi: \to F$ and $\psi:F' \to F'$ are minimal injective endomorphisms such that $F *_\phi \cong F' Then $\phi$ is irreducible if and only if is irreducible. Q.I.-Invariance Conjecture ========================== Very little is known the of mapping whose are and atoroidal. Brinkmann that atoroidal automorphisms of free groups have word-hyperbolic mapping tori [@Bri00] and gave an explicit description for such mapping non-trivial splittings cyclic [@Bri02]: atoroidal automorphism of a mapping torus that splits over and only if the (outer) automorphism fixes a splitting*]{} of in particular, mapping tori of irreducible atoroidal automorphisms never split over $\mathbb Z$. By of Kapovich-Kleiner [@KK00], word-hyperbolic free-by-cyclic groups that do not split over $\mathbb Z$ have Gromov to a menger curve. the topology of boundary not to irreducible and monodromies as there are reducible and atoroidal automorphisms that do not a free splitting. The Gromov boundary of a word-hyperbolic group given [*visual metric*]{} that unique up to [*quasi-symmetry*]{} its class is a complete of group, are if only their visual boundaries quasi-symmetric [@P | vanishing Euler characterisTic too and iT is noNcyCliC sInce The oNly virtually cyCLic tOrsion-free group is $\mathbB Z$, yet $h$ iS Not cYClIc.
\[nonGeomcomM\]suPPOse $\PhI: F \To F$ AnD $\PsI:F' \to F'$ Are Minimal Injective eNdoMoRphisms such tHAt $f *_\phi$ and $F' *_\psI$ arE commensurabLe. THen $\phi$ Is IrrEDucibLe aNd atoRoidal IF and onLy if $\psi$ is IrREducibLE and atoROIdAl.
FiNally, we combine thiS ThEOrem with CorollAry \[irrEdTOrGRP\] to Get:
\[Irredgrp\]SuPpOse $\phI: f \to F$ and $\PSi:f' \TO f'$ arE Minimal injectIve endomorpHIsmS such tHaT $F *_\pHI \cong F' *_\Psi$. ThEn $\PHi$ iS irreduciblE if aNd only if $\pSi$ is irREduciblE.
q.I.-InvarIance COnjEctUre {#qI}
==========================
veRy LitTlE Is kNOwN abOUt tHe geometRy Of MappiNg toRI WHOse mOnoDromIes arE irreducible aNd aToroIDal. brinkMann pRoveD tHat atOroidaL autoMoRphisms of free grOups Have word-hYpeRbOliC mAppinG Tori [@BrI00] anD gaVe an expLicit deSCriPtION FoR when such mapping toRi HAVe Non-triviAl spliTTiNgS Over cyclIc SubGrouPS [@bri02]: an AtorOIdAl automoRphism OF $F$ HaS a mappiNg Torus tHaT spLitS over $\MAthbB Z$ if anD only if tHe (outER) automorphism fIXes a [*free splitTInG*]{} OF $F$; IN parTicUlar, mapping Tori OF irrEducIBlE anD AtoroIdal aUtOMoRPhisms never split oveR $\mAthbb Z$.
by worK of Kapovich-KlEiner [@KK00], worD-HYPerbolic Free-BY-cYClic groups that Do not Split over $\mAThbb Z$ havE GromOv boundaRy homeomoRPHic to a meNgeR cuRve. unfORTuNately, the topoLOGy of ThE boundaRy iS not sufFicIenT to DetEcT irreduciBle and atOrOiDaL mOnoDromiES as there ArE reDuCibLe and AToroidAl autOmorPhIsMS thAt do not FIx A FRee sPlItTing.
the grOmov bOundARy oF a word-hYperbolic GroUP can Be GiVen a [*visUal metric*]{} that Is Unique up to [*QuAsi-SymmetRY*]{} And its quAsi-symmetry class is a compLEte geomEtrIc invAriaNt of the grOup, I.e., word-HypERbolic Groups Are quAsI-isOMEtric IF AnD onLy Their visuaL BOunDarieS aRe quAsi-symmEtric [@P | vanishing Euler character istic tooand i t i s n on cycl ic s ince the onlyv irtu ally cyclic torsion-fr ee gr ou p is$ \m athbb Z$, ye t $ H $ is n ot cy cl i c.
\[n ong eomcomm \]Suppose$\p hi : F \to F$ a n d$\psi:F' \ toF'$ are mini mal injec ti vee ndomo rph ismssuch t h at $F*_\phi$ a nd $F' *_ \ psi$ ar e co mmen surable. Then $\p h i$ is irreducible and a to r oi d a l i f a nd only if $ \psi$ is irre d uc i b l e a n d atoroidal.
Finally, w e co mbineth ist heorem with C o rol lary \[irre dtor grp\] toget:
\ [irredg r p\]Supp ose $\ phi : F \to F$ a nd$\ p si: F '\to F'$ are min im al inje ctiv e e n domo rph isms such that $F *_\p hi\con g F' *_\p si$.Then $ \phi$ is ir reduc ib le if and onlyif $ \psi$ isirr ed uci bl e.
Q . I.-Inv ari anc e Conje cture { # qi}
= = = = == ================== == Ve ry littl e is k n ow na bout the g eom etry o f map ping to ri whose monod r om ie s are i rr educib le an d a toroi d al.Brinkm ann prov ed th a t atoroidal au t omorphisms of fr e e g r oups ha ve word-hyp erbo l ic m appi n gtor i [@Br i00]an d g a ve an explicit desc ri ptionfor w hen such mapp ing tori h a v e non-tri vial sp l ittings over c yclic subgroups [@Bri02] : anatoroida l automor p h ism of $ F$has amap p i ng torus that s p l itsov er $\ma thb b Z$ if an d o nly if t he (outer ) automo rp hi sm f ixe s a [ * free spl it tin g* ]{} of $ F $; inparti cula r, m a ppi ng tori of i rred uc ib le a ndat oroid al a u tom orphism s never s pli t ove r$\ mathbbZ$.
By workof Kapovich- Kl ein er [@K K 0 0], word -hyperbolic free-by-cyc l ic grou psthatdo n ot splitove r $\ma thb b Z$ ha ve Gro mov b ou nda r y home o m or phi cto a menge r cur ve. U nf ortu nately, the topology of t h e b oundary is no t s uffi c i en t t o d e tec ti rre d u cible and atoro idal monod ro m ie s as there are r educibl e and a toroi d al auto morphisms that dono t fi x a f ree splitt ing.
Th e Gromovb ounda r yof awor d-hype rb oli c gro up can begiven a [*v is ual me tric* ]{ } that i s unique up to [*quasi- symmet ry*]{ } a nd its qu asi - sym metry cla ss i s a comple tegeo metri c i n varia nt o f t heg roup, i.e . , word-hy p er bol i c g roups are q u a s i-i somet ric if and onl y their visual bo u ndaries are qu asi- s y mme tri c [@P | vanishing_Euler characteristic_too and it is_noncyclic since_the_only virtually_cyclic_torsion-free group is_$\mathbb Z$, yet_$H$ is not cyclic.
\[nongeomcomm\]Suppose_$\phi: F \to_F$_and $\psi:F' \to F'$ are minimal injective endomorphisms such that $F *_\phi$ and $F'_*_\psi$_are commensurable._Then_$\phi$_is irreducible and atoroidal if_and only if $\psi$ is_irreducible and_atoroidal.
Finally, we combine this theorem with Corollary \[irredtorgrp\] to_get:
\[irredgrp\]Suppose_$\phi: F \to_F$ and $\psi:F' \to F'$ are minimal injective endomorphisms_such that $F *_\phi \cong F'_*_\psi$. Then $\phi$_is_irreducible_if and only if_$\psi$ is irreducible.
Q.I.-Invariance Conjecture {#qi}
==========================
Very little_is known about the geometry of_mapping tori whose monodromies are irreducible and_atoroidal. Brinkmann proved that atoroidal automorphisms_of free groups have word-hyperbolic_mapping tori_[@Bri00] and gave an explicit_description for when_such mapping_tori have non-trivial_splittings over cyclic subgroups [@Bri02]: an_atoroidal automorphism of_$F$ has a mapping torus that_splits_over $\mathbb Z$_if_and_only if_the (outer) automorphism_fixes_a [*free_splitting*]{}_of $F$; in particular, mapping tori_of_irreducible and atoroidal automorphisms never split over $\mathbb_Z$.
By work of Kapovich-Kleiner_[@KK00],_word-hyperbolic free-by-cyclic groups that_do not split over $\mathbb Z$_have Gromov boundary homeomorphic to a_menger curve._Unfortunately, the_topology of the boundary is not sufficient to detect irreducible and_atoroidal monodromies as there are reducible_and atoroidal automorphisms that_do not_fix_a free splitting.
The_Gromov_boundary of_a word-hyperbolic group can be given a_[*visual metric*]{}_that is unique up to [*quasi-symmetry*]{}_and its quasi-symmetry class_is_a complete geometric invariant of the_group, i.e., word-hyperbolic groups are quasi-isometric_if and only their visual_boundaries_are_quasi-symmetric [@P |
\[ssec:ap2\] and \[ssec:arch\]). We now quantify how the size of the aperture affects the observed ratio of the dynamical and luminous masses of a cluster, $M_{dyn}/M_{lum}$. In other words, we derive $M_{dyn}/M_{lum}$ as would be done by an observer.
Quantifying the impact of the aperture {#ssec:ap2}
--------------------------------------
When the aperture is large, the cylinder it defines (i.e. the cylinder with the line-of-sight as its main axis) intercepts a greater fraction of the outer shells containing the stars expanding in the vicinity of the tidal radius and beyond. As a result, the measured radial velocity dispersion rises. This simple exercise demonstrates the importance of the size of the field-of-view: the smaller the aperture, the smaller the observed departure from virial equilibrium, and the shorter the time-span required for the cluster targeted region to return to equilibrium.
Figure \[fig:mdml\] shows the post-gas-expulsion evolution of $M_{dyn}/M_{lum}$. Each panel corresponds to a different set of model parameters, i.e. star formation efficiency ($\epsilon$), gas expulsion time-scale ($\tau_{gexp}/\tau_{cross}$), tidal field impact ($r_h/r_t$) and their corresponding final bound fraction ($F_{bound}$). The cases covered by Fig. \[fig:mdml\] range from the complete disruption of the cluster ($F_{bound}=0$ in top panel \[a\]) to a cluster barely affected by gas expulsion ($F_{bound} \lesssim 1$ in bottom panel \[d\]). Note the degeneracy between panels \[b\] and \[c\]: the lower SFE of panel \[b\] (SFE=0.33 compared to 0.40 in panel \[c\]) is compensated by its longer gas-expulsion time-scale ($\tau_{gexp}=\tau_{cross}$ compared to $\tau_{gexp}=0$ in panel \[c\]). Each panel considers in turn four aperture radii, $r_{ap}=$ 3.5, 5, 10 and 20pc, which correspond to 4.5, 6.4, 12.8 and 25.6 cluster initial half-mass radii ($r_{hm, init}=0.8$pc; see key in top panel). Measuring in our simulations the observed half-light radius for apertures smaller | \[ssec: ap2\ ] and \[ssec: arch\ ]). We now quantify how the size of the aperture affects the observed ratio of the dynamic and aglow bulk of a cluster, $ M_{dyn}/M_{lum}$. In early words, we deduce $ M_{dyn}/M_{lum}$ as would be done by an observer.
quantify the shock of the aperture { # ssec: ap2 }
--------------------------------------
When the aperture is large, the cylinder it specify (i.e. the cylinder with the line - of - sight as its independent axis) intercepts a big fraction of the out shells containing the stars expanding in the vicinity of the tidal radius and beyond. As a consequence, the measured radial velocity dispersion rise. This simple exercise demonstrates the importance of the size of the battlefield - of - view: the smaller the aperture, the smaller the observed departure from virial chemical equilibrium, and the shorter the time - span required for the cluster targeted region to return to equilibrium.
Figure \[fig: mdml\ ] shows the post - gas - expulsion evolution of $ M_{dyn}/M_{lum}$. Each panel corresponds to a different set of model parameter, i.e. star formation efficiency ($ \epsilon$), natural gas ejection time - scale ($ \tau_{gexp}/\tau_{cross}$), tidal field impingement ($ r_h / r_t$) and their corresponding final bound fraction ($ F_{bound}$). The cases cover by Fig. \[fig: mdml\ ] range from the complete disruption of the cluster ($ F_{bound}=0 $ in top panel \[a\ ]) to a cluster scantily affected by gas expulsion ($ F_{bound } \lesssim 1 $ in bottom panel \[d\ ]). Note the degeneracy between panels \[b\ ] and \[c\ ]: the low SFE of panel \[b\ ] (SFE=0.33 compare to 0.40 in panel \[c\ ]) is compensated by its longer accelerator - expulsion time - scale ($ \tau_{gexp}=\tau_{cross}$ compare to $ \tau_{gexp}=0 $ in panel \[c\ ]). Each panel considers in turn four aperture radius, $ r_{ap}=$ 3.5, 5, 10 and 20pc, which correspond to 4.5, 6.4, 12.8 and 25.6 bunch initial half - mass radii ($ r_{hm, init}=0.8$pc; see key in top panel). measure in our simulations the observe half - light spoke for apertures smaller | \[ssef:ap2\] and \[ssec:arch\]). We now duantify how thg wize oh the alerture xffects the observed ratio oh thw dynqmical and luminous marses of a cluster, $M_{dyi}/M_{lum}$. In other words, we derive $L_{dyn}/K_{oum}$ as would bg done by an observer.
Quanthfhiug the impact of the aperture {#ssec:ap2}
--------------------------------------
Rhen thr wperture is latge, tnq cympnber it defines (i.e. the cylinder sith tht line-of-sight as iys main axis) intercepts a hreaher fraction of thf outer sheols sintaining thd stars exkauding in thg vicinity of the tidal radius avd bevond. As a rgshlh, the measurxd radpal velocity dispersimn risex. This simple cxercmse eemonstrates the impoctance of the size os the fiend-kf-view: the smallee rhe akertuse, tfw soalmec tge obsfrvxd departurs from viriql equilibrium, and uhe whorter the tjme-spag wequired for the cluster targeted regiot tk return to equilibrium.
Digure \[fig:mdml\] shows jhe post-gaf-expulsion evolution of $M_{dyn}/M_{lum}$. Each panel correvponda to a aufverent set of model parameters, i.e. star formatykn enficiency ($\epsilok$), gas expulsion tike-dcsje ($\tau_{gexp}/\tau_{zross}$), cjdzl field impact ($r_h/g_t$) and jheir xorresponqing final bound fraction ($F_{boune}$). The cases boveeed by Fig. \[fig:mdml\] xange from tke comklete cisruption of the clustzr ($F_{bohnd}=0$ in top oanel \[a\]) tk a cluster barelh anfewted by gas expulsion ($F_{bougd} \lesssin 1$ iu bottom pangl \[d\]). Noee the degfneragf between panels \[b\] and \[e\]: the lower SFE of panel \[b\] (SFE=0.33 compared to 0.40 in panel \[c\]) is cokpansdted by nts lokger gas-expulsijn time-scale ($\tcu_{gexp}=\tab_{cross}$ compared no $\tau_{gex'}=0$ in panel \[c\]). Each panel cmjsiders in tnrn four wperrure radii, $f_{xp}=$ 3.5, 5, 10 and 20pc, wnich corrvskond to 4.5, 6.4, 12.8 and 25.6 cluster initlal hxmf-mass radii ($r_{hn, ibit}=0.8$pc; see key im tup [ajek). Iaasuring in mur rimjkatiovs the obsevvea hakf-light radius for a[erthres smaller | \[ssec:ap2\] and \[ssec:arch\]). We now quantify how of aperture affects observed ratio of of cluster, $M_{dyn}/M_{lum}$. In words, we derive as would be done by an Quantifying the impact of the aperture {#ssec:ap2} -------------------------------------- When the aperture is large, cylinder it defines (i.e. the cylinder with the line-of-sight as its main axis) a fraction the shells containing the stars expanding in the vicinity of the tidal radius and beyond. As a the measured radial velocity dispersion rises. This simple demonstrates the importance of size of the field-of-view: the the the smaller observed from equilibrium, and the the time-span required for the cluster targeted region to return to equilibrium. Figure \[fig:mdml\] shows the post-gas-expulsion of $M_{dyn}/M_{lum}$. corresponds to different of parameters, i.e. star ($\epsilon$), gas expulsion time-scale ($\tau_{gexp}/\tau_{cross}$), tidal and their corresponding final bound fraction ($F_{bound}$). The covered by \[fig:mdml\] range from the complete disruption the cluster ($F_{bound}=0$ in top panel \[a\]) to cluster barely affected by gas expulsion ($F_{bound} \lesssim 1$ in bottom panel \[d\]). Note the panels \[b\] and \[c\]: lower SFE of \[b\] compared 0.40 panel \[c\]) compensated by its longer gas-expulsion time-scale ($\tau_{gexp}=\tau_{cross}$ compared to $\tau_{gexp}=0$ in \[c\]). Each panel considers in turn four aperture radii, $r_{ap}=$ 10 20pc, which correspond 4.5, 6.4, 12.8 and cluster half-mass radii ($r_{hm, init}=0.8$pc; in panel). simulations observed radius for apertures smaller | \[ssec:ap2\] and \[ssec:arch\]). We now quaNtify how thE size Of tHe aPeRturE affEcts the observeD RatiO of the dynamical and lumiNous mAsSEs of A ClUster, $m_{dyn}/M_{luM}$. in OTHer WoRdS, we DeRIvE $M_{dyn}/m_{luM}$ as woulD be done by aN obSeRver.
QuantifyINg The impact oF thE aperture {#sseC:ap2}
--------------------------------------
when thE aPerTUre is LarGe, the CylindER it defInes (i.e. the CyLInder wITh the liNE-Of-SighT as its main axis) intERcEPts a greater fraCtion oF tHE oUTEr sHelLs containiNg The stARs expanDInG IN The VIcinity of the tIdal radius aND beYond. As A rEsuLT, the meAsureD rADiaL velocity diSperSion rises. this siMPle exerCIse demoNstratEs tHe iMporTAnCe Of tHe SIze OF tHe fIEld-Of-view: thE sMaLler tHe apERTURe, thE smAlleR the oBserved departUre From VIriAl equIlibrIum, aNd The shOrter tHe timE-sPan required for tHe clUster targEteD rEgiOn To retURn to eqUilIbrIum.
FiguRe \[fig:mdML\] shOwS THE pOst-gas-expulsion evoLuTIOn Of $M_{dyn}/M_{lUm}$. Each PAnEl COrresponDs To a DiffEREnt seT of mODeL parametErs, i.e. sTAr FoRmation EfFicienCy ($\EpsIloN$), gas eXPulsIon timE-scale ($\taU_{gexp}/\TAu_{cross}$), tidal fiELd impact ($r_h/r_t$) aND tHEIr COrreSpoNding final bOund FRactIon ($F_{BOuNd}$). THE caseS coveReD By fIg. \[fig:mdml\] range from tHe CompleTe disRuption of the cLuster ($F_{bouND}=0$ IN top paneL \[a\]) to A ClUSter barely affeCted bY gas expulsIOn ($F_{bound} \LesssIm 1$ in bottOm panel \[d\]). NOTE the degeNerAcy BetWeeN PAnEls \[b\] and \[c\]: the loWER SFE Of Panel \[b\] (SfE=0.33 cOmpared To 0.40 iN paNel \[C\]) is CoMpensated By its lonGeR gAs-ExPulSion tIMe-scale ($\tAu_{GexP}=\tAu_{cRoss}$ cOMpared To $\tau_{Gexp}=0$ In PaNEl \[c\]). each panEL cONSideRs In Turn FouR aPertuRe raDIi, $r_{Ap}=$ 3.5, 5, 10 and 20pc, Which corrEspONd to 4.5, 6.4, 12.8 AnD 25.6 cLuster iNitial half-masS rAdii ($r_{hm, iniT}=0.8$pC; seE key in TOP panel). MeAsuring in our simulations THe obserVed Half-lIght Radius for ApeRtures SmaLLer | \[ssec:ap2\] and \[ssec:ar ch\]). Wenow q uan tif yhowthesize of the ap e rtur e affects the observed rati oo f th e d ynami cal and lu m i nou sma sse so fa clu ste r, $M_{ dyn}/M_{lu m}$ .In other wor d s, we derive $M _{dyn}/M_{lu m}$ as wo ul d b e done by an o bserve r .
Qua ntifyingth e impac t of the a pe rtur e {#ssec:ap2}
--- - -- - -------------- ------ -- - -- - - --- -
When the a pe rture is larg e ,t h e cy l inder it defi nes (i.e. t h e c ylinde rwit h the l ine-o f- s igh t as its ma in a xis) inte rcepts a great e r fract ion of th e o uter sh el lsco n tai n in g t h e s tars exp an di ng in the v i c init y o f th e tid al radius and be yond . As a re sult, the m easur ed rad ial v el ocity dispersio n ri ses. This si mp leex ercis e demon str ate s the i mportan c e o ft h e s ize of the field-o f- v i ew : the sm allert he a p erture,th e s mall e r theobse r ve d depart ure fr o mvi rial eq ui libriu m, an d t he sh o rter the t ime-span requ i red for the cl u ster targeted re g i on to r etu rn to equil ibri u m.
Figu r e\[f i g:mdm l\] s ho w st he post-gas-expulsi on evolu tionof $M_{dyn}/M _{lum}$. E a c h panel c orre s po n ds to a differ ent s et of mode l paramet ers,i.e. sta r formati o n efficie ncy ($ \ep sil o n $) , gas expulsi o n tim e- scale ( $\t au_{gex p}/ \ta u_{ cro ss }$), tida l fieldim pa ct ( $r_ h/r_t $ ) and th ei r c or res pondi n g fina l bou nd f ra ct i on($F_{bo u nd } $ ). T he c ases co ve red b y Fi g . \ [fig:md ml\] rang e f r om t he c omplete disruption o fthe cluste r($F _{boun d } =0$ in t op panel \[a\]) to a cl u ster ba rel y aff ecte d by gasexp ulsion ($ F _{boun d} \le sssim 1 $ i n botto m pa nel \ [d\]). Not e the dege ne racy betwee n panels \[b\] and \[c \]: the lower SF E of p an el\ [b \ ] ( SF E =0. 3 3 compared to 0. 40 in pane l\ [c \]) is com p ens at ed by i ts long er ga s -expuls ion time- scale ($\ ta u_{g e x p}= \tau_{cros s}$ comp ared to $ \ tau_{ g ex p}=0$ in panel \ [c\ ]). E ach pa n elconsi ders i nturn f our a pe rture ra dii, $r_{ap}=$ 3.5, 5,10 and 20pc , w hich corr esp o ndto 4.5, 6 .4,12.8 and 2 5.6 cl uster in i tialhalf - ma ssr adii($r_ { hm, init} = 0. 8$p c ; s ee key in t o p pan el).Mea s uringin o ur simulations th e observed half -lig h t ra diu s for a pertures small er | \[ssec:ap2\] and_\[ssec:arch\]). We_now quantify how the_size of_the_aperture affects_the_observed ratio of_the dynamical and_luminous masses of a_cluster, $M_{dyn}/M_{lum}$. In_other_words, we derive $M_{dyn}/M_{lum}$ as would be done by an observer.
Quantifying the impact of_the_aperture {#ssec:ap2}
--------------------------------------
When_the_aperture_is large, the cylinder it_defines (i.e. the cylinder with_the line-of-sight_as its main axis) intercepts a greater fraction_of_the outer shells_containing the stars expanding in the vicinity of the_tidal radius and beyond. As a_result, the measured_radial_velocity_dispersion rises. This simple_exercise demonstrates the importance of the_size of the field-of-view: the smaller_the aperture, the smaller the observed departure_from virial equilibrium, and the shorter_the time-span required for the_cluster targeted_region to return to equilibrium.
Figure_\[fig:mdml\] shows the_post-gas-expulsion evolution_of $M_{dyn}/M_{lum}$. Each_panel corresponds to a different set_of model parameters,_i.e. star formation efficiency ($\epsilon$), gas_expulsion_time-scale ($\tau_{gexp}/\tau_{cross}$), tidal_field_impact_($r_h/r_t$) and_their corresponding final_bound_fraction ($F_{bound}$)._The_cases covered by Fig. \[fig:mdml\] range from_the_complete disruption of the cluster ($F_{bound}=0$ in_top panel \[a\]) to_a_cluster barely affected by_gas expulsion ($F_{bound} \lesssim 1$_in bottom panel \[d\]). Note the_degeneracy between_panels \[b\]_and \[c\]: the lower SFE of panel \[b\] (SFE=0.33 compared to_0.40 in panel \[c\]) is compensated_by its longer gas-expulsion_time-scale ($\tau_{gexp}=\tau_{cross}$_compared_to $\tau_{gexp}=0$ in_panel_\[c\]). Each_panel considers in turn four aperture radii,_$r_{ap}=$ 3.5,_5, 10 and 20pc, which correspond_to 4.5, 6.4, 12.8_and_25.6 cluster initial half-mass radii ($r_{hm,_init}=0.8$pc; see key in top panel)._Measuring in our simulations the_observed_half-light_radius for apertures smaller |
percolating sets can have different sizes. For example, another minimal percolating set with $5$ vertices appears in Figure \[ex1\] (b).
![(a) In this tree, having nodes $2,4,5$ infected (shaded in red) initially is sufficient to ensure that the whole tree is infected. (b) This minimal percolating set shaded in red is of size $5$.[]{data-label="ex1"}](Fig8.jpg)
In what follows we shall work with general finite trees $T(V,E)$ with set of vertices $V$ and set of edges $E$. In particular, we shall consider the smallest minimal percolating sets in the following section.
Algorithms for Finding Smallest Minimal Percolating Set {#minimal2}
=======================================================
Consider $F(t)$-bootstrap percolation on a tree $T(V,E)$ with initially infected set $A_0\subset V$. As before, we shall denote by $A_t$ be the set of nodes infected at time $t$. For simplicity, we shall use here the word “infected” synonymously with “infected”. In order to build an algorithm to find smallest percolating sets, we first need to introduce a few definitions that will simplify the notation at later stages.
We shall denote by $L(a)$ the largest time $t$ such that $a \leq F(t),$ and if there does not exist such a time $t$, then set $L(a)=\infty$. Similarly, define $B(a)$ as the smallest time $t$ such that $a \leq F(t)$, and if such a time $t$ does not exist, set $B(a)=\infty$.
Given $a,b\in \mathbb{N}$, if $a<b$ then $L(a) \geq L(b)$. Indeed, this holds because if a node can be infected to with $b$ neighbors, it can with $a$ neighbors where $a<b$. Note that in general, a smallest percolating set $A_0$ must be a minimal percolating set. To see this, suppose not. Then there exists some $v$ in $A_0$ such that $A_0 -\{v\}$ percolates the graph. That means that $A_0 -\{v\}$, a smaller set that $A_0$, is a percolating set | percolating sets can have different size. For case, another minimal percolating set with $ 5 $ vertices appears in Figure \[ex1\ ] (b-complex vitamin).
! [ (a) In this tree, having nodes $ 2,4,5 $ infected (shade in loss) initially is sufficient to ensure that the unharmed tree is infected. (b) This minimal percolating set up shaded in red is of size $ 5$.[]{data - label="ex1"}](Fig8.jpg)
In what follow we shall work with general finite trees $ T(V, E)$ with set of vertices $ V$ and set of edge $ E$. In particular, we shall consider the smallest minimal percolating sets in the be section.
Algorithms for Finding Smallest Minimal Percolating Set { # minimal2 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Consider $ F(t)$-bootstrap percolation on a tree $ T(V, E)$ with initially infected put $ A_0\subset V$. As before, we shall denote by $ A_t$ be the set of nodes infected at meter $ t$. For simplicity, we shall use here the word “ infected ” synonymously with “ infected ”. In order to build an algorithm to find smallest percolating sets, we first need to introduce a few definitions that will simplify the notation at later stages.
We shall announce by $ L(a)$ the large time $ t$ such that $ a \leq F(t),$ and if there does not exist such a prison term $ t$, then specify $ L(a)=\infty$. Similarly, define $ B(a)$ as the smallest time $ t$ such that $ a \leq F(t)$, and if such a meter $ t$ does not exist, set $ B(a)=\infty$.
Given $ a, b\in \mathbb{N}$, if $ a < b$ then $ L(a) \geq L(b)$. Indeed, this holds because if a node can be infect to with $ b$ neighbors, it can with $ a$ neighbors where $ a < b$. Note that in general, a smallest percolating set $ A_0 $ must be a minimal percolating bent. To see this, suppose not. Then there exists some $ v$ in $ A_0 $ such that $ A_0 -\{v\}$ percolates the graph. That mean that $ A_0 -\{v\}$, a smaller stage set that $ A_0 $, is a percolating set | pegcolating sets can have aifferent sizes. For exemple, ahother mknimal percolating set with $5$ vwrtictf appears in Figurd \[ex1\] (b).
![(a) Ij this teee, iaving nodes $2,4,5$ inhscted (smcded jk red) mnitially is sunficient to ensure that tve wkole tree is infected. (b) This minimal percolstlng set shaded in geq is of size $5$.[]{data-label="ex1"}](Fig8.jpg)
In what rollows we shall worl with general finite treed $T(V,F)$ with set of vertlces $V$ and wet jd edges $E$. In particulag, we shall donsider the smallest minimal pdrcolcting sets un thf following wectijn.
Algorithms for Finging Smsllest Minimal Peccolqting Set {#minimal2}
=======================================================
Consmder $F(t)$-bootstrap persolation mn a tree $T(V,E)$ with unutialny itfecgwd ret $A_0\xugset V$. As before, we shall denore by $A_t$ be the set os nodes infectsd at eiie $t$. For simplicity, we shall use here tve sord “infected” synonymouwly with “infected”. In lrder to fuild an algorithm to find smallest percolating sats, wx wirwt vwef to introduce a few definitions that will siillofj the notation at later stagrs.
Ae fhall denote cy $L(a)$ cge largest time $t$ sufh that $a \lew F(t),$ and yf tnere does not exist such a rime $t$, then fwt $L(a)=\infty$. Similarpy, define $B(c)$ as tne smsllest time $t$ such that $a \lsq F(t)$, and iv such a fkme $t$ does not ebisn, sed $B(a)=\infty$.
Given $a,b\in \mathbb{G}$, if $a<b$ tien $L(c) \geq L(b)$. Indged, thif holds befause if a node can be ijfectgd to fith $b$ neihhbors, it can with $a$ neighbors xiere $a<b$. Note jhad it generaj, a skallest percojating set $A_0$ most be a iinimxl percolaning set. Vo see this, fuppose not. Tvgn there exisvs some $v$ in $Q_0$ suxh that $X_0 -\{v\}$ percolates the grapk. Thar means that $A_0 -\{v\}$, a smxmler set that $A_0$, is a percolating xet | percolating sets can have different sizes. For minimal set with vertices appears in this having nodes $2,4,5$ (shaded in red) is sufficient to ensure that the tree is infected. (b) This minimal percolating set shaded in red is of $5$.[]{data-label="ex1"}](Fig8.jpg) In what follows we shall work with general finite trees $T(V,E)$ with of $V$ set edges $E$. In particular, we shall consider the smallest minimal percolating sets in the following section. for Finding Smallest Minimal Percolating Set {#minimal2} ======================================================= $F(t)$-bootstrap percolation on a $T(V,E)$ with initially infected set V$. before, we denote $A_t$ the set of infected at time $t$. For simplicity, we shall use here the word “infected” synonymously with “infected”. In to build to find percolating we need to introduce definitions that will simplify the notation We shall denote by $L(a)$ the largest time such that \leq F(t),$ and if there does exist such a time $t$, then set $L(a)=\infty$. define $B(a)$ as the smallest time $t$ such that $a \leq F(t)$, and if such $t$ does not exist, $B(a)=\infty$. Given $a,b\in if then \geq Indeed, this because if a node can be infected to with $b$ neighbors, can with $a$ neighbors where $a<b$. Note that in general, percolating $A_0$ must be minimal percolating set. To this, not. Then there exists in such percolates graph. means that $A_0 -\{v\}$, smaller set that $A_0$, is percolating set | percolating sets can have difFerent sizeS. For eXamPle, AnOtheR minImal percolatinG Set wIth $5$ vertices appears in FiGure \[eX1\] (b).
![(A) in thIS tRee, haVing nodES $2,4,5$ iNFEctEd (ShAdeD iN ReD) initIalLy is sufFicient to eNsuRe That the whole TReE is infecteD. (b) THis minimal peRcoLating SeT shADed in Red Is of sIze $5$.[]{datA-Label="eX1"}](Fig8.jpg)
In WhAT folloWS we shalL WOrK witH general finite treES $T(v,e)$ with set of vertIces $V$ aNd SEt OF EdgEs $E$. in particulAr, We shaLL considER tHE SMalLEst minimal perColating setS In tHe follOwIng SEction.
algorItHMs fOr Finding SmAlleSt Minimal percolATing Set {#MInimal2}
=======================================================
COnsideR $F(t)$-BooTstrAP pErColAtIOn oN A tRee $t(v,E)$ wIth initiAlLy InfecTed sET $a_0\SUbseT V$. AS befOre, we Shall denote by $a_t$ bE the SEt oF nodeS infeCted At Time $t$. for simPliciTy, We shall use here tHe woRd “infecteD” syNoNymOuSly wiTH “infecTed”. in oRder to bUild an aLGorItHM TO fInd smallest percolaTiNG SeTs, we firsT need tO InTrODuce a few DeFinItioNS That wIll sIMpLify the nOtatioN At LaTer stagEs.
we shalL dEnoTe bY $L(a)$ thE LargEst timE $t$ such thAt $a \leQ f(t),$ and if there doES not exist such A TiME $T$, tHEn seT $L(a)=\Infty$. SimilaRly, dEFine $b(a)$ as THe SmaLLest tIme $t$ sUcH ThAT $a \leq F(t)$, and if such a tiMe $T$ does nOt exiSt, set $B(a)=\infty$.
GIven $a,b\in \maTHBB{N}$, if $a<b$ thEn $L(a) \GEq l(B)$. Indeed, this holDs becAuse if a nodE Can be infEcted To with $b$ nEighbors, iT CAn with $a$ nEigHboRs wHerE $A<B$. NOte that in geneRAL, a smAlLest perColAting seT $A_0$ mUst Be a MinImAl percolaTing set. TO sEe ThIs, SupPose nOT. Then theRe ExiStS soMe $v$ in $a_0$ Such thAt $A_0 -\{v\}$ pErcoLaTeS The Graph. ThAT mEANs thAt $a_0 -\{v\}$, A smaLleR sEt thaT $A_0$, is A PerColatinG set | percolating sets can have different size s.For e xamp le,another minima l per colating set with $5$verti ce s app e ar s inFigure\ [e x 1 \](b ).
! [( a )In th istree, h aving node s $ 2, 4,5$ infecte d ( shaded inred ) initiallyissuffic ie ntt o ens ure that the w h ole tr ee is inf ec t ed. (b ) This m i n im al p ercolating set sh a de d in red is ofsize $ 5$ . [] { d ata -la bel="ex1"} ]( Fig8. j pg)
In wh a t fol l ows we shallwork with g e ner al fin it e t r ees $T (V,E) $w ith set of ver tice s $V$ and set o f edges$ E$. Inpartic ula r,we s h al lcon si d ert he sm a lle st minim al p ercol atin g s e ts i n t he f ollow ing section.
Al gori t hms forFindi ng S ma llest Minim al Pe rc olating Set {#m inim al2}
==== === == === == ===== = ====== === === ======= ======= = === == = = =
Consider $F(t)$-bo ot s t ra p percol ationo nat ree $T(V ,E )$with i nitia llyi nf ected se t $A_0 \ su bs et V$.As befor e, we sh all d e note by $A _t$ be t he se t of nodes infe c ted at time $ t $. F or simp lic ity, we sha ll u s e he re t h ewor d “inf ected ”s yn o nymously with “infe ct ed”. I n ord er to build a n algorith m t o find s mall e st percolating se ts, w e first ne e d to int roduc e a fewdefinitio n s that wi llsim pli fyt h enotation at l a t er s ta ges.
W e s hall de not e b y $ L(a )$ the larg est time $ t$ s uc h t hat $ a \leq F( t) ,$an d i f the r e does notexis tsu c h a time $ t $, t hense t$L(a )=\ in fty$. Sim i lar ly, def ine $B(a) $ a s the s ma llest t ime $t$ suchth at $a \leq F (t) $, and i f such a time $t$ does not exis t , set $ B(a )=\in fty$ .
Given$a, b\in \ mat h bb{N}$ , if $ a<b$th en$ L (a) \ g e qL(b )$ . Indeed,t h isholds b ecau se if a node can be infec t edto with $b$ n eig hbor s , i t c a nw ith $ a $ n e i ghbors where $a <b$. Noteth a tin general , asm allestpercola tings et $A_0 $ must be a minima lperc o l ati ng set. To see thi s, suppos e not. Th en th ere exist ssom e $v$ in $A _ 0$suchthat $ A_ 0 -\{v \}$ p er colatesthe graph. That means t hat $A _0 -\ {v\ }$, a sma lle r se t that $A _0$, is a perc ola tin g set | percolating_sets can_have different sizes. For_example, another_minimal_percolating set_with_$5$ vertices appears_in Figure \[ex1\]_(b).
![(a) In this tree,_having nodes $2,4,5$_infected_(shaded in red) initially is sufficient to ensure that the whole tree is infected._(b)_This minimal_percolating_set_shaded in red is of_size $5$.[]{data-label="ex1"}](Fig8.jpg)
In what follows we_shall work_with general finite trees $T(V,E)$ with set of_vertices_$V$ and set_of edges $E$. In particular, we shall consider the_smallest minimal percolating sets in the_following section.
Algorithms for_Finding_Smallest_Minimal Percolating Set {#minimal2}
=======================================================
Consider_$F(t)$-bootstrap percolation on a tree $T(V,E)$_with initially infected set $A_0\subset V$._As before, we shall denote by $A_t$_be the set of nodes infected_at time $t$. For simplicity,_we shall_use here the word “infected”_synonymously with “infected”._In order_to build an_algorithm to find smallest percolating sets,_we first need_to introduce a few definitions that_will_simplify the notation_at_later_stages.
We shall_denote by $L(a)$_the_largest time_$t$_such that $a \leq F(t),$ and_if_there does not exist such a time_$t$, then set $L(a)=\infty$._Similarly,_define $B(a)$ as the_smallest time $t$ such that_$a \leq F(t)$, and if such_a time_$t$ does_not exist, set $B(a)=\infty$.
Given $a,b\in \mathbb{N}$, if $a<b$ then $L(a) \geq_L(b)$. Indeed, this holds because if_a node can be_infected to_with_$b$ neighbors, it_can_with $a$_neighbors where $a<b$. Note that in general,_a smallest_percolating set $A_0$ must be a_minimal percolating set. To_see_this, suppose not. Then there exists_some $v$ in $A_0$ such that_$A_0 -\{v\}$ percolates the graph._That_means_that $A_0 -\{v\}$, a smaller_set that $A_0$, is a percolating_set |
${{\mathbb{F}}}_{q^n}^\times \cong {{\mathbb{Z}}}/(q^n-1){{\mathbb{Z}}}$ under any embedding $
{{\mathbb{F}}}_{q^n}^\times \hookrightarrow GL_{{{\mathbb{F}}}_q}({{\mathbb{F}}}_{q^n}) \cong GL_n({{\mathbb{F}}}_q)
$ that comes from a choice of ${{\mathbb{F}}}_q$-vector space isomorphism ${{\mathbb{F}}}_{q^n} \cong {{\mathbb{F}}}_q^n$. The analogy between Singer cycles in $GL_n({{\mathbb{F}}}_q)$ and $n$-cycles in ${\mathfrak{S}}_n$ is reasonably well-established [@StantonWebbR §7], [@StantonWhiteR §§8-9]. Fixing such a Singer cycle $c$, denote by $t_q(n,\ell)$ the number of ordered factorizations $(t_1,\ldots,t_\ell)$ of $c=t_1 t_2 \cdots t_\ell$ in which each $t_i$ is a [*reflection*]{} in $GL_n({{\mathbb{F}}}_q)$, that is, the fixed space $({{\mathbb{F}}}_q^n)^{t_i}$ is a hyperplane in ${{\mathbb{F}}}_q^n$.
\[q-Denes-theorem\] For $n \geq 2$, one has $
t_q(n,n)=(q^n-1)^{n-1}.
$
\[q-factorization-theorem\] For $n \geq 2$, more generally $t_q(n,\ell)$ has ordinary generating function $$\label{q-Jackson-ordinary-gf-product}
\sum_{\ell \geq 0} t_q(n,\ell) x^\ell
= (q^n-1)^{n-1} x^n \cdot
\left(1+x[n]_q \right)^{-1} \prod_{k=0}^{n-1} \left(1+x[n]_q(1+q^k-q^{k+1})\right)^{-1}$$ and explicit formulas $$\begin{aligned}
\label{q-Jackson-sum-formula}
\qquad
t_q(n,\ell)
&= \frac{(-[n]_q)^\ell}{q^{\binom{n}{2}}( | $ { { \mathbb{F}}}_{q^n}^\times \cong { { \mathbb{Z}}}/(q^n-1){{\mathbb{Z}}}$ under any embedding $
{ { \mathbb{F}}}_{q^n}^\times \hookrightarrow GL_{{{\mathbb{F}}}_q}({{\mathbb{F}}}_{q^n }) \cong GL_n({{\mathbb{F}}}_q)
$ that comes from a choice of $ { { \mathbb{F}}}_q$-vector space isomorphism $ { { \mathbb{F}}}_{q^n } \cong { { \mathbb{F}}}_q^n$. The doctrine of analogy between Singer cycle in $ GL_n({{\mathbb{F}}}_q)$ and $ n$-cycles in $ { \mathfrak{S}}_n$ is reasonably well - established [ @StantonWebbR § 7 ], [ @StantonWhiteR § § 8 - 9 ]. specify such a Singer bicycle $ c$, denote by $ t_q(n,\ell)$ the number of order factorizations $ (t_1,\ldots, t_\ell)$ of $ c = t_1 t_2 \cdots t_\ell$ in which each $ t_i$ is a [ * reflection * ] { } in $ GL_n({{\mathbb{F}}}_q)$, that is, the sterilize space $ ({ { \mathbb{F}}}_q^n)^{t_i}$ is a hyperplane in $ { { \mathbb{F}}}_q^n$.
\[q - Denes - theorem\ ] For $ n \geq 2 $, one has $
t_q(n, n)=(q^n-1)^{n-1 }.
$
\[q - factorization - theorem\ ] For $ n \geq 2 $, more broadly $ t_q(n,\ell)$ has ordinary generating routine $ $ \label{q - Jackson - average - gf - product }
\sum_{\ell \geq 0 } t_q(n,\ell) x^\ell
= (q^n-1)^{n-1 } x^n \cdot
\left(1+x[n]_q \right)^{-1 } \prod_{k=0}^{n-1 } \left(1+x[n]_q(1+q^k - q^{k+1})\right)^{-1}$$ and explicit formula $ $ \begin{aligned }
\label{q - Jackson - sum - formula }
\qquad
t_q(n,\ell)
& = \frac{(-[n]_q)^\ell}{q^{\binom{n}{2 } } ( | ${{\mahhbb{F}}}_{q^n}^\times \cong {{\mathbb{E}}}/(q^n-1){{\mathbb{Z}}}$ under cby embxdding $
{{\jathbb{F}}}_{q^v}^\times \hookrightarrow GL_{{{\mathub{F}}}_q}({{\nathbv{F}}}_{q^n}) \cong GL_n({{\mathbb{F}}}_q)
$ ghat comed from a chomce of ${{\mathbb{F}}}_q$-vxdtor space isklorpkiwm ${{\mathbb{F}}}_{q^n} \cpng {{\mathbb{X}}}_q^n$. The analogf cecween Singer cycles in $GL_n({{\mathbb{F}}}_q)$ agd $n$-cycked in ${\mathfrak{S}}_g$ is weasknably well-established [@StantonWebbD §7], [@StanuonWhiteR §§8-9]. Fixing xuch a Singer cycle $c$, denohe bj $t_q(n,\ell)$ the numbeg of orderee fasrorizations $(g_1,\ldots,t_\ell)$ of $c=t_1 t_2 \cdojs t_\ell$ in which each $t_i$ is a [*rewlectnon*]{} in $GL_n({{\mqtybb{V}}}_x)$, that is, tie fixvd space $({{\mathnn{F}}}_q^n)^{t_i}$ is a huperplane in ${{\msthub{F}}}_q^b$.
\[q-Denes-theorem\] For $n \jeq 2$, one has $
t_q(n,n)=(q^n-1)^{n-1}.
$
\[z-factorizdtnon-theorem\] For $n \geq 2$, mire ggneranly $g_w(n,\eul)$ gax krdinagy jenerating runction $$\lavel{q-Jackson-ordinary-bf-[gpduct}
\sum_{\ell \feq 0} t_z(n,\qll) x^\ell
= (q^n-1)^{n-1} x^n \cdot
\left(1+x[n]_q \right)^{-1} \prog_{k=0}^{n-1} \left(1+x[n]_q(1+q^k-q^{k+1})\right)^{-1}$$ and ezplicit formulas $$\begij{aligned}
\jabel{q-Jackson-sum-formula}
\qquad
t_q(n,\ell)
&= \frac{(-[n]_q)^\ell}{q^{\bitom{n}{2}}( | ${{\mathbb{F}}}_{q^n}^\times \cong {{\mathbb{Z}}}/(q^n-1){{\mathbb{Z}}}$ under any embedding $ GL_{{{\mathbb{F}}}_q}({{\mathbb{F}}}_{q^n}) GL_n({{\mathbb{F}}}_q) $ comes from a ${{\mathbb{F}}}_{q^n} {{\mathbb{F}}}_q^n$. The analogy Singer cycles in and $n$-cycles in ${\mathfrak{S}}_n$ is reasonably [@StantonWebbR §7], [@StantonWhiteR §§8-9]. Fixing such a Singer cycle $c$, denote by $t_q(n,\ell)$ number of ordered factorizations $(t_1,\ldots,t_\ell)$ of $c=t_1 t_2 \cdots t_\ell$ in which each is [*reflection*]{} $GL_n({{\mathbb{F}}}_q)$, is, the fixed space $({{\mathbb{F}}}_q^n)^{t_i}$ is a hyperplane in ${{\mathbb{F}}}_q^n$. \[q-Denes-theorem\] For $n \geq 2$, one $ t_q(n,n)=(q^n-1)^{n-1}. $ \[q-factorization-theorem\] For $n \geq 2$, generally $t_q(n,\ell)$ has ordinary function $$\label{q-Jackson-ordinary-gf-product} \sum_{\ell \geq 0} x^\ell (q^n-1)^{n-1} x^n \left(1+x[n]_q \prod_{k=0}^{n-1} and explicit formulas \label{q-Jackson-sum-formula} \qquad t_q(n,\ell) &= \frac{(-[n]_q)^\ell}{q^{\binom{n}{2}}( | ${{\mathbb{F}}}_{q^n}^\times \cong {{\mathbb{Z}}}/(Q^n-1){{\mathbb{Z}}}$ uNder aNy eMbeDdIng $
{{\mAthbB{F}}}_{q^n}^\times \hookrIGhtaRrow GL_{{{\mathbb{F}}}_q}({{\mathbb{F}}}_{q^N}) \cong gL_N({{\MathBB{F}}}_Q)
$ that Comes frOM a CHOicE oF ${{\mAthBb{f}}}_Q$-vEctor SpaCe isomoRphism ${{\mathBb{F}}}_{Q^n} \Cong {{\mathbb{F}}}_q^N$. thE analogy beTweEn Singer cyclEs iN $GL_n({{\maThBb{F}}}_Q)$ And $n$-cYclEs in ${\mAthfraK{s}}_n$ is reAsonably wElL-EstablIShed [@StaNTOnwebbr §7], [@StantonWhiteR §§8-9]. FixINg SUch a Singer cyclE $c$, denoTe BY $t_Q(N,\Ell)$ The Number of orDeRed faCTorizatIOnS $(T_1,\LDotS,T_\ell)$ of $c=t_1 t_2 \cdotS t_\ell$ in whicH EacH $t_i$ is a [*ReFleCTion*]{} in $gL_n({{\maThBB{F}}}_q)$, That is, the fiXed sPace $({{\mathbB{F}}}_q^n)^{t_i}$ IS a hyperPLane in ${{\mAthbb{F}}}_Q^n$.
\[q-denEs-thEOrEm\] for $N \gEQ 2$, onE HaS $
t_q(N,N)=(q^n-1)^{N-1}.
$
\[q-factorIzAtIon-thEoreM\] fOR $N \geq 2$, MorE genErallY $t_q(n,\ell)$ has ordInaRy geNEraTing fUnctiOn $$\laBeL{q-JacKson-orDinarY-gF-product}
\sum_{\ell \gEq 0} t_q(N,\ell) x^\ell
= (q^N-1)^{n-1} x^N \cDot
\LeFt(1+x[n]_q \RIght)^{-1} \prOd_{k=0}^{N-1} \leFt(1+x[n]_q(1+q^k-Q^{k+1})\right)^{-1}$$ ANd eXpLICIt Formulas $$\begin{alignEd}
\LABeL{q-JacksoN-sum-foRMuLa}
\QQuad
t_q(n,\eLl)
&= \FraC{(-[n]_q)^\eLL}{Q^{\binoM{n}{2}}( | ${{\mathbb{F}}}_{q^n}^\ti mes \cong{{\ma thb b{Z }} }/(q ^n-1 ){{\mathbb{Z}} } $ un der any embedding $
{{ \math bb { F}}} _ {q ^n}^\ times \ h oo k r igh ta rr owGL _ {{ {\mat hbb {F}}}_q }({{\mathb b{F }} }_{q^n}) \co n gGL_n({{\ma thb b{F}}}_q)
$tha t come sfro m a ch oic e of${{\ma t hbb{F} }}_q$-vec to r space isomorp h i sm ${{ \mathbb{F}}}_{q^n } \ c ong {{\mathbb{ F}}}_q ^n $ .T h e a nal ogy betwee nSinge r cycles in $ G L_n ( {{\mathbb{F}} }_q)$ and $ n $-c yclesin ${ \ mathfr ak{S} }_ n $ i s reasonabl y we ll-establ ished[ @Stanto n WebbR § 7], [@ Sta nto nWhi t eR § §8- 9] . Fi x in g s u cha Singer c yc le $c $, d e n o t e by $t _q(n ,\ell )$ the number of ord e red fact oriza tion s$(t_1 ,\ldot s,t_\ el l)$ of $c=t_1 t _2 \ cdots t_\ ell $inwh ich e a ch $t_ i$isa [*ref lection * ]{} i n $ GL _n({{\mathbb{F}}}_ q) $ , t hat is,the fi x ed s p ace $({{ \m ath bb{F } } }_q^n )^{t _ i} $ is a h yperpl a ne i n ${{\m at hbb{F} }} _q^ n$.
\[q - Dene s-theo rem\] Fo r $n\ geq 2$, one ha s $
t_q(n,n)=( q ^n - 1 )^ { n-1} .
$
\[q-facto riza t ion- theo r em \]F or $n \geq 2 $ ,m ore generally $t_q( n, \ell)$ hasordinary gene rating fun c t i on $$\la bel{ q -J a ckson-ordinary -gf-p roduct}
\s u m_{\ell\geq0} t_q(n ,\ell) x^ \ e ll
= (q^ n-1 )^{ n-1 } x ^ n\cdot
\left( 1 + x[n] _q \right )^{ -1} \pr od_ {k= 0}^ {n- 1} \left(1+ x[n]_q(1 +q ^k -q ^{ k+1 })\ri g ht)^{-1} $$ an dexp licit formul as $$ \beg in {a l ign ed}
\l a be l { q-Ja ck so n-su m-f or mula}
\q q uad
t_q(n, \ell)
&= \f r ac{( -[ n] _q)^\el l}{q^{\binom{ n} {2}}( | ${{\mathbb{F}}}_{q^n}^\times_\cong {{\mathbb{Z}}}/(q^n-1){{\mathbb{Z}}}$_under any embedding $
{{\mathbb{F}}}_{q^n}^\times_\hookrightarrow GL_{{{\mathbb{F}}}_q}({{\mathbb{F}}}_{q^n})_\cong_GL_n({{\mathbb{F}}}_q)
$ that_comes_from a choice_of ${{\mathbb{F}}}_q$-vector space_isomorphism ${{\mathbb{F}}}_{q^n} \cong {{\mathbb{F}}}_q^n$._The analogy between_Singer_cycles in $GL_n({{\mathbb{F}}}_q)$ and $n$-cycles in ${\mathfrak{S}}_n$ is reasonably well-established [@StantonWebbR §7], [@StantonWhiteR §§8-9]._Fixing_such a_Singer_cycle_$c$, denote by $t_q(n,\ell)$ the_number of ordered factorizations $(t_1,\ldots,t_\ell)$_of $c=t_1_t_2 \cdots t_\ell$ in which each $t_i$ is_a_[*reflection*]{} in $GL_n({{\mathbb{F}}}_q)$,_that is, the fixed space $({{\mathbb{F}}}_q^n)^{t_i}$ is a hyperplane_in ${{\mathbb{F}}}_q^n$.
\[q-Denes-theorem\] For $n \geq 2$,_one has $
t_q(n,n)=(q^n-1)^{n-1}.
$
\[q-factorization-theorem\]_For_$n_\geq 2$, more generally_$t_q(n,\ell)$ has ordinary generating function $$\label{q-Jackson-ordinary-gf-product}
\sum_{\ell_\geq 0} t_q(n,\ell) x^\ell
= (q^n-1)^{n-1} _x^n \cdot
\left(1+x[n]_q \right)^{-1} \prod_{k=0}^{n-1} \left(1+x[n]_q(1+q^k-q^{k+1})\right)^{-1}$$ and_explicit formulas $$\begin{aligned}
\label{q-Jackson-sum-formula}
\qquad
t_q(n,\ell)
&=_\frac{(-[n]_q)^\ell}{q^{\binom{n}{2}}( |
the desired inequality in.
Proof of Proposition \[prop:stability\] {#sec:proofpropstability}
---------------------------------------
We prove Proposition \[prop:stability\] for the monopoly and the proof is identical for the symmetric duopoly. To prove Proposition \[prop:stability\], we first formulate the static optimization problem via a network flow model that characterizes the *capacity region* of the network for a given set of prices $\ell_{ij}^1(t)=\ell_{ij}^1\; \forall t$ (Hence, $\Theta_{ij}(t)=\Theta_{ij} \; \forall t$). The capacity region is defined as the set of all arrival rates $[\Lambda_{ij}]_{i,j \in \mathcal M}$, where there exists a charging and routing policy under which the queueing network of the system is stable. Let $x_i^e$ be the number of vehicles available at node $i$, $\alpha_{ij}^e$ be the fraction of vehicles at node $i$ with energy level $v$ being routed to node $j$, and $\alpha_{ic}^e$ be the fraction of vehicles charging at node $i$ starting with energy level $v$. We say the static vehicle allocation for node $i$ and energy level $v$ is feasible if $\alpha_{ic}^e + \sum_{\substack{j\in{\cal M}\\ j\neq i}} \alpha_{ij}^e \leq 1$. The optimization problem that characterizes the capacity region of the network ensures that the total number of vehicles routed from $i$ to $j$ is at least as large as the nominal arrival rate to the queue $(i,j)$. Namely, the vehicle allocation problem can be formulated as follows:
\[eq:capacityregionproblem\] $$\begin{aligned}
\label{eq:capobj}&\underset{x_i^e,\alpha_{ij}^e,\alpha_{ic}^e}{\text{min}}& & \rho\\
& \text{subject to}
\label{eq:capconst1}& & \Theta_{ij} \leq \sum_{e=e_{ij}}^{e_{\max}}x_i^e\alpha_{ij}^e \quad\forall i,j\in \mathcal M,\\
\label{eq:capconst2}& & & \rho \geq \alpha_{ic}^e + \sum_{\substack{j\in{\cal M}\\ j\ | the desired inequality in.
Proof of Proposition \[prop: stability\ ] { # sec: proofpropstability }
---------------------------------------
We prove Proposition \[prop: stability\ ] for the monopoly and the proof is identical for the symmetric duopoly. To prove Proposition \[prop: stability\ ], we foremost give voice the static optimization problem via a net flow mannequin that characterizes the * capacity area * of the net for a given set of prices $ \ell_{ij}^1(t)=\ell_{ij}^1\; \forall t$ (Hence, $ \Theta_{ij}(t)=\Theta_{ij } \; \forall t$). The capacity area is defined as the set of all arrival rates $ [ \Lambda_{ij}]_{i, j \in \mathcal M}$, where there exists a charging and spread-eagle policy under which the queueing network of the system is stable. Let $ x_i^e$ be the act of vehicles available at node $ i$, $ \alpha_{ij}^e$ be the fraction of vehicles at node $ i$ with energy horizontal surface $ v$ being routed to node $ j$, and $ \alpha_{ic}^e$ be the fraction of vehicles charging at node $ i$ get down with energy level $ v$. We say the static vehicle allotment for node $ i$ and energy level $ v$ is feasible if $ \alpha_{ic}^e + \sum_{\substack{j\in{\cal M}\\ j\neq i } } \alpha_{ij}^e \leq 1$. The optimization problem that characterizes the capacity region of the network ensures that the total number of vehicles routed from $ i$ to $ j$ is at least as large as the nominal arrival rate to the queue $ (iodine, j)$. Namely, the vehicle allocation trouble can be give voice as follow:
\[eq: capacityregionproblem\ ] $ $ \begin{aligned }
\label{eq: capobj}&\underset{x_i^e,\alpha_{ij}^e,\alpha_{ic}^e}{\text{min } } & & \rho\\
& \text{subject to }
\label{eq: capconst1 } & & \Theta_{ij } \leq \sum_{e = e_{ij}}^{e_{\max}}x_i^e\alpha_{ij}^e \quad\forall i, j\in \mathcal M,\\
\label{eq: capconst2 } & & & \rho \geq \alpha_{ic}^e + \sum_{\substack{j\in{\cal M}\\ j\ | thf desired inequality in.
Pvoof of Propositnin \[pro':stabiljty\] {#sec:pfoofpropstability}
---------------------------------------
We prove Prlpisitiin \[prop:stability\] for thd monopolj and the proid is identmdal for the sglmetxir duopoly. To prpve Proposhtion \[prop:stabhlktv\], we first formulate the static optiiizatiom oroblem via a getwpwk fmow model that characterizes the *czpacity region* of thr network for a given set lf pgices $\ell_{ij}^1(t)=\ell_{ij}^1\; \vorall t$ (Hebce, $\Eyeta_{ij}(t)=\Theta_{kj} \; \forall t$). The capacjty region is defined as the seg of cll arrival rqted $[\Lambda_{ij}]_{i,j \in \mwthcal M}$, whevv there exists a charging anc rmutung policy under whici the queueing netwotk of the vyatem is stable. Ler $z_i^e$ bg the numcwr uf beiicmes avwileble at nods $i$, $\alpha_{ij}^w$ be the fraction og dvnicles at nose $i$ wytr energy level $v$ being routed to node $j$, ans $\alpha_{ic}^e$ be the fractuon of vehicles charglng at noqe $i$ starting with energy level $v$. We say the stathc veiizle ajuicwtion for node $i$ and energy level $v$ is feasibjs of $\alpha_{ic}^e + \sum_{\fubstack{j\in{\vap K}\\ j\neq i}} \alpha_{kj}^e \leq 1$. Fhe optimization pgoblem jhat cyaracteriees tne capacity region of the nwtwork ensurvs tyat the total numbzr of vehiclzs roujed frpm $i$ to $j$ is at least ar ladge as the jominal adfival rate to thd qleue $(i,j)$. Namely, the vehicle alljcation pcoblek can bd fotmulateq as folloas:
\[eq:capacityregionproblel\] $$\begnn{alicned}
\label{ee:capobj}&\underset{x_i^e,\alpha_{ij}^e,\alpha_{mr}^e}{\text{min}}& & \rhp\\
& \dexn{subject co}
\labek{eq:capconst1}& & \Eheta_{ij} \leq \suk_{e=e_{ij}}^{e_{\mcx}}x_i^e\aupha_{ij}^e \quzd\foraln i,j\in \mathsal M,\\
\label{eq:cdkconst2}& & & \rho \jeq \alpha_{yc}^e + \sum_{\wubstacy{g\in{\cal M}\\ j\ | the desired inequality in. Proof of Proposition --------------------------------------- prove Proposition for the monopoly for symmetric duopoly. To Proposition \[prop:stability\], we formulate the static optimization problem via network flow model that characterizes the *capacity region* of the network for a set of prices $\ell_{ij}^1(t)=\ell_{ij}^1\; \forall t$ (Hence, $\Theta_{ij}(t)=\Theta_{ij} \; \forall t$). The capacity is as set all arrival rates $[\Lambda_{ij}]_{i,j \in \mathcal M}$, where there exists a charging and routing policy under the queueing network of the system is stable. $x_i^e$ be the number vehicles available at node $i$, be fraction of at $i$ energy level $v$ routed to node $j$, and $\alpha_{ic}^e$ be the fraction of vehicles charging at node $i$ starting with level $v$. the static allocation node and energy level feasible if $\alpha_{ic}^e + \sum_{\substack{j\in{\cal M}\\ \leq 1$. The optimization problem that characterizes the region of network ensures that the total number vehicles routed from $i$ to $j$ is at as large as the nominal arrival rate to the queue $(i,j)$. Namely, the vehicle allocation be formulated as follows: $$\begin{aligned} \label{eq:capobj}&\underset{x_i^e,\alpha_{ij}^e,\alpha_{ic}^e}{\text{min}}& & & to} & \leq \sum_{e=e_{ij}}^{e_{\max}}x_i^e\alpha_{ij}^e i,j\in \mathcal M,\\ \label{eq:capconst2}& & & \rho \geq \alpha_{ic}^e + \sum_{\substack{j\in{\cal j\ | the desired inequality in.
ProOf of ProposItion \[ProP:stAbIlitY\] {#sec:ProofpropstabiLIty}
---------------------------------------
WE prove Proposition \[prop:sTabilItY\] For tHE mOnopoLy and thE PrOOF is IdEnTicAl FOr The syMmeTric duoPoly. To provE PrOpOsition \[prop:sTAbIlity\], we firSt fOrmulate the sTatIc optiMiZatIOn proBleM via a NetworK Flow moDel that chArACterizES the *capACItY regIon* of the network foR A gIVen set of prices $\Ell_{ij}^1(t)=\ElL_{Ij}^1\; \FORalL t$ (HEnce, $\Theta_{iJ}(t)=\theta_{IJ} \; \forall T$). thE CAPacITy region is defIned as the seT Of aLl arriVaL raTEs $[\LambDa_{ij}]_{i,J \iN \MatHcal M}$, where tHere Exists a chArging ANd routiNG policy Under wHicH thE queUEiNg NetWoRK of THe SysTEm iS stable. LEt $X_i^E$ be thE numBER OF vehIclEs avAilabLe at node $i$, $\alphA_{ij}^E$ be tHE frActioN of veHiclEs At nodE $i$ with EnergY lEvel $v$ being routeD to nOde $j$, and $\alPha_{Ic}^E$ be ThE fracTIon of vEhiCleS chargiNg at nodE $I$ stArTING wIth energy level $v$. We sAy THE sTatic vehIcle alLOcAtIOn for nodE $i$ And EnerGY Level $V$ is fEAsIble if $\alPha_{ic}^e + \SUm_{\SuBstack{j\In{\Cal M}\\ j\nEq I}} \alPha_{Ij}^e \leQ 1$. the oPtimizAtion proBlem tHAt characterizeS The capacity reGIoN OF tHE netWorK ensures thaT the TOtal NumbER oF veHIcles RouteD fROm $I$ To $j$ is at least as large As The nomInal aRrival rate to tHe queue $(i,j)$. NAMELy, the vehIcle ALlOCation problem cAn be fOrmulated aS Follows:
\[eQ:capaCityregiOnproblem\] $$\BEGin{alignEd}
\lAbeL{eq:CapOBJ}&\uNderset{x_i^e,\alpHA_{Ij}^e,\aLpHa_{ic}^e}{\teXt{mIn}}& & \rho\\
& \teXt{sUbjEct To}
\lAbEl{eq:capcoNst1}& & \Theta_{Ij} \LeQ \sUm_{E=e_{iJ}}^{e_{\max}}X_I^e\alpha_{iJ}^e \QuaD\fOraLl i,j\iN \MathcaL M,\\
\labEl{eq:CaPcONst2}& & & \Rho \geq \aLPhA_{IC}^e + \suM_{\sUbStacK{j\iN{\cAl M}\\ j\ | the desired inequality in .
Proof o f Pro pos iti on \[p rop: stability\] {# s ec:p roofpropstability}
--- ----- -- - ---- - -- ----- ------- - -- - - --- -
W e p ro v ePropo sit ion \[p rop:stabil ity \] for the mon o po ly and the pr oof is ident ica l forth e s y mmetr icduopo ly. To provePropositi on \[prop : stabili t y \] , we first formulatet he static optimiz ationpr o bl e m vi a a network f lo w mod e l thatc ha r a c ter i zes the *capa city region * of the n et wor k for a give ns etof prices $ \ell _{ij}^1(t )=\ell _ {ij}^1\ ; \foral l t$ ( Hen ce, $\T h et a_ {ij }( t )=\ T he ta_ { ij} \; \for al lt$).Thec a p a city re gion is d efined as the se t of all arri val r ates $ [\Lam bda_{i j}]_{ i, j \in \mathcalM}$, where th ere e xis ts a ch a rgingand ro uting p olicy u n der w h i c hthe queueing netwo rk o fthe syst em iss ta bl e . Let $x _i ^e$ bet h e num bero fvehicles avail a bl eat node $ i$, $\ al pha _{i j}^e$ be t he fra ction of vehi c les at node $i $ with energyl ev e l $ v $ be ing routed tonode $j$, and $\ alp h a_{ic }^e$be th e fraction of vehicl es charg ing a t node $i$ st arting wit h e nergy le vel$ v$ . We say the st aticvehicle al l ocationfor n ode $i$and energ y level $v $ i s f eas ibl e if $\alpha_{ic} ^ e + \ su m_{\sub sta ck{j\in {\c alM}\ \ j \n eq i}} \a lpha_{ij }^ e\l eq 1$ . The optimiza ti onpr obl em th a t char acter izes t he cap acity r e gi o n ofth enetw ork e nsure s th a t t he tota l numberofv ehic le sroutedfrom $i$ to $ j$ is at lea st as large a s the no minal arrival rate to t h e queue $( i,j)$ . Na mely, the ve hicleall o cation probl em ca nbef o rmula t e dasfo llows:
\[ e q :ca pacit yr egio nproble m\] $$\begin{align e d}\label{eq:cap obj }&\u n d er set { x_ i ^e, \a l pha _ { ij}^e,\alpha_{i c}^e}{\tex t{ m in }}& & \rho \ \
& \ text{su bject t o}
\l a bel{eq: capconst1 }& & \The ta _{ij } \le q \sum_{e= e_{ij}}^ {e_{\max} } x_i^e \ al pha_{ ij} ^e \qu ad \fo ralli,j\in \ma thcal M,\\\l abel{e q:cap co nst2}& & & \rho \geq \alpha_{ic }^e +\sum_ {\s ubstack{j \in { \ca l M}\\ j\ | the_desired inequality_in.
Proof of Proposition \[prop:stability\]_{#sec:proofpropstability}
---------------------------------------
We prove_Proposition \[prop:stability\]_for the_monopoly_and the proof_is identical for_the symmetric duopoly. To_prove Proposition \[prop:stability\],_we_first formulate the static optimization problem via a network flow model that characterizes the_*capacity_region* of_the_network_for a given set of_prices $\ell_{ij}^1(t)=\ell_{ij}^1\; \forall t$ (Hence,_$\Theta_{ij}(t)=\Theta_{ij} \;_\forall t$). The capacity region is defined as_the_set of all_arrival rates $[\Lambda_{ij}]_{i,j \in \mathcal M}$, where there exists_a charging and routing policy under_which the queueing_network_of_the system is stable._Let $x_i^e$ be the number of_vehicles available at node $i$, $\alpha_{ij}^e$_be the fraction of vehicles at node_$i$ with energy level $v$ being_routed to node $j$, and_$\alpha_{ic}^e$ be_the fraction of vehicles charging_at node $i$_starting with_energy level $v$._We say the static vehicle allocation_for node $i$_and energy level $v$ is feasible_if_$\alpha_{ic}^e + \sum_{\substack{j\in{\cal_M}\\_j\neq_i}} \alpha_{ij}^e_\leq 1$. The_optimization_problem that_characterizes_the capacity region of the network_ensures_that the total number of vehicles routed_from $i$ to $j$_is_at least as large_as the nominal arrival rate_to the queue $(i,j)$. Namely, the_vehicle allocation_problem can_be formulated as follows:
\[eq:capacityregionproblem\] $$\begin{aligned}
\label{eq:capobj}&\underset{x_i^e,\alpha_{ij}^e,\alpha_{ic}^e}{\text{min}}& & \rho\\
& \text{subject to}
\label{eq:capconst1}& & \Theta_{ij}_\leq \sum_{e=e_{ij}}^{e_{\max}}x_i^e\alpha_{ij}^e \quad\forall i,j\in \mathcal M,\\
\label{eq:capconst2}&_& & \rho \geq_\alpha_{ic}^e +_\sum_{\substack{j\in{\cal_M}\\ j\ |
0\] Let $f\colon \mathbb{R} \to [0,\infty)$ be a continuous function verifying $(F_1)-(F_3)$. Then, there exists $\bar{\lambda}>0$ such that for $\lambda>\bar{\lambda}$ the problem $(P_{\lambda})$ admits two nonnegative solutions $u_{\lambda}, v_{\lambda} \in X^s_0(\Omega)$. Moreover, $0\le u_\lambda < v_{\lambda}\le 1$.
The main ideas behind of the proof relies on known arguments for solving this type of problems. However, this issue is non-trivial, by which we need to treat the problema under an appropriate approach that allows to apply such arguments and, in this way, to save several technical difficulties that arise in the nonlocal case.
Specifically, we obtain the first solution $u _{\lambda}$ by truncating the nonlinearity and taking $\lambda$ large enough, where we have assumed only the local condition at zero ($F_1$). The behavior of the norms, according to the parameter $\lambda $ of the possible solutions, is based on the De Giorgi-Nash-Moser theory, precisely on Moser’s iterative scheme. Such procedure was introduced in the mid 1950s and early 1960s, we refer to [@DeGiorgi; @Moser]. An important fact in the proof is that the functional energy associated with the truncated problem has the Mountain Pass (MP) geometry. Thus, we can control, in terms of the parameter $\lambda$, the MP level and the norm of the MP solution.
Respect to the second solution, we impose the conditions ($F_2$) and ($F_3$) on $f$, which widely known when applying the sub- and super-solution method in problems like (\[P\]) with $s=1$. Here we show that this idea still remains hold by defining properly sub- and super-solutions in our nonlocal context, it is when $0<s<1$.
The document is organized as follows: in Section 2 we offer a brief review of the fractional spaces of Sobolev in the context of our problem and remember some useful results. In Section 3 we look for the first solution, while in Section 3 we find the second solution, which ends the proof of Theorem \[main-theorem0\]
Functional framework and preliminaries {#section2}
======================================
In this section, we offer a | 0\ ] Let $ f\colon \mathbb{R } \to [ 0,\infty)$ be a continuous function verifying $ (F_1)-(F_3)$. Then, there exists $ \bar{\lambda}>0 $ such that for $ \lambda>\bar{\lambda}$ the trouble $ (P_{\lambda})$ admit two nonnegative solutions $ u_{\lambda }, v_{\lambda } \in X^s_0(\Omega)$. Moreover, $ 0\le u_\lambda < v_{\lambda}\le 1$.
The chief mind behind of the proof relies on know arguments for solving this character of trouble. However, this issue is non - superficial, by which we need to treat the problema under an appropriate approach path that admit to apply such arguments and, in this way, to save several technical trouble that arise in the nonlocal case.
Specifically, we obtain the beginning solution $ u _ { \lambda}$ by truncating the nonlinearity and taking $ \lambda$ large enough, where we have assume only the local condition at zero ($ F_1 $). The behavior of the norms, according to the argument $ \lambda $ of the possible solutions, is based on the De Giorgi - Nash - Moser theory, precisely on Moser ’s iterative scheme. Such procedure was introduce in the mid 1950s and early 1960s, we refer to [ @DeGiorgi; @Moser ]. An important fact in the proof is that the functional energy associated with the truncated problem have the Mountain Pass (MP) geometry. Thus, we can control, in terms of the argument $ \lambda$, the MP degree and the norm of the MP solution.
Respect to the second solution, we impose the conditions ($ F_2 $) and ($ F_3 $) on $ f$, which widely known when applying the sub- and ace - solution method in problems like (\[P\ ]) with $ s=1$. Here we show that this idea still remains hold by defining properly sub- and super - solutions in our nonlocal context, it is when $ 0 < s<1$.
The document is organize as watch: in Section 2 we offer a brief review of the fractional spaces of Sobolev in the context of our trouble and remember some useful results. In Section 3 we look for the first solution, while in Section 3 we find the second solution, which ends the validation of Theorem \[main - theorem0\ ]
Functional framework and preliminaries { # section2 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
In this section, we offer a | 0\] Leh $f\colon \mathbb{R} \to [0,\inftn)$ be a continuous functmon verjfying $(F_1)-(W_3)$. Then, there exists $\bar{\lambde}>0$ suxh thqt for $\lambda>\bar{\lambda}$ the probpem $(P_{\lamvda})$ edmits two nonnejztive solutiohd $u_{\lcmuda}, v_{\lambda} \in W^s_0(\Omega)$. Moraover, $0\le u_\lambga < r_{\lambda}\le 1$.
The main ideas behind of tre proog gelies on knowg arbtmenfs for solving this type of problejs. Howener, this issue is non-trivial, by which we nefd tl treat the problela under an appwipriate appruach that allows to apkly such arguments and, in this wxy, to save sevetco tfwhnical difhicultpes that arisc in tha nonloval case.
Speciflcallb, we obtain the first solntion $u _{\lambda}$ by troncating tve nonlinearity and tqking $\lamtda$ uqrgd ehongh, where we have assujed only thw local condition au zqgp ($F_1$). The behabior os ehe norms, according to the parameter $\lakbdz $ of the possible solurions, is based on the De Giorgy-Nash-Moser theory, precisely on Moser’s iterative swheme. Ruck procdeuge was introduced in the mid 1950s and early 1960s, we degeg to [@DeGiorgi; @Moscr]. An important favt im the proof is that cge functional energy associwted qith the uruncsted problem has the Mountaun Pass (MP) gvomerry. Thus, we can coutrol, in teros og the parameter $\lambda$, the M' levem and the nlrm of ths MP solution.
Respdct tm the second solution, we iipose the conbitions ($W_2$) anc ($F_3$) on $f$, which wldely known when applyinh the sgb- and supfr-solution method in problems lmle (\[P\]) with $s=1$. Nese fe show chat tmis idea still wemains hold bi defininy propdrly sub- ahd supec-solutions ig our nonlocan context, it ms when $0<s<1$.
Ehe eocunent is urganized as fpllows: in Section 2 qe offer a brief rcview kf the fractioncu spaces of Sobokev in tje cjttext of our procleo and femember sond usrful results. In Secthon 3 we look for the forft solutuon, whilq in Section 3 we find the seconf solntion, xhich rndf the proof of Theorem \[main-thekrem0\]
Functlonwl framework and preliminarizs {#section2}
======================================
In this section, we offer a | 0\] Let $f\colon \mathbb{R} \to [0,\infty)$ be function $(F_1)-(F_3)$. Then, exists $\bar{\lambda}>0$ such $(P_{\lambda})$ two nonnegative solutions v_{\lambda} \in X^s_0(\Omega)$. $0\le u_\lambda < v_{\lambda}\le 1$. The ideas behind of the proof relies on known arguments for solving this type problems. However, this issue is non-trivial, by which we need to treat the under appropriate that to apply such arguments and, in this way, to save several technical difficulties that arise in nonlocal case. Specifically, we obtain the first solution _{\lambda}$ by truncating the and taking $\lambda$ large enough, we assumed only local at ($F_1$). The behavior the norms, according to the parameter $\lambda $ of the possible solutions, is based on the De theory, precisely iterative scheme. procedure introduced the mid 1950s 1960s, we refer to [@DeGiorgi; @Moser]. in the proof is that the functional energy with the problem has the Mountain Pass (MP) Thus, we can control, in terms of the $\lambda$, the MP level and the norm of the MP solution. Respect to the second impose the conditions ($F_2$) ($F_3$) on $f$, widely when the and super-solution in problems like (\[P\]) with $s=1$. Here we show that this still remains hold by defining properly sub- and super-solutions in context, is when $0<s<1$. document is organized as in 2 we offer a of fractional in context our problem and remember useful results. In Section 3 look for the first we find the second solution, which ends the of Theorem \[main-theorem0\] Functional framework and preliminaries ====================================== In this section, we offer a | 0\] Let $f\colon \mathbb{R} \to [0,\infty)$ be A continuouS funcTioN veRiFyinG $(F_1)-(F_3)$. THen, there exists $\BAr{\laMbda}>0$ such that for $\lambda>\bAr{\lamBdA}$ The pRObLem $(P_{\lAmbda})$ adMItS TWo nOnNeGatIvE SoLutioNs $u_{\Lambda}, v_{\Lambda} \in X^s_0(\omeGa)$. moreover, $0\le u_\lAMbDa < v_{\lambda}\lE 1$.
ThE main ideas beHinD of the PrOof RElies On kNown aRgumenTS for soLving this TyPE of proBLems. HowEVEr, This Issue is non-trivial, BY wHIch we need to treAt the pRoBLeMA UndEr aN appropriaTe ApproACh that aLLoWS TO apPLy such argumenTs and, in this WAy, tO save sEvEraL TechniCal diFfICulTies that ariSe in The nonlocAl case.
sPecificALly, we obTain thE fiRst SoluTIoN $u _{\LamBdA}$ By tRUnCatINg tHe nonlinEaRiTy and TakiNG $\LAMbda$ LarGe enOugh, wHere we have assUmeD onlY The Local CondiTion At Zero ($F_1$). the behAvior Of The norms, accordiNg to The parameTer $\LaMbdA $ oF the pOSsible SolUtiOns, is baSed on thE de GIoRGI-naSh-Moser theory, preciSeLY On moser’s itErativE ScHeME. Such proCeDurE was INTroduCed iN ThE mid 1950s and Early 1960s, WE rEfEr to [@DeGIoRgi; @MosEr]. an iMpoRtant FAct iN the prOof is thaT the fUNctional energy ASsociated with THe TRUnCAted ProBlem has the MOuntAIn PaSs (MP) GEoMetRY. Thus, We can CoNTrOL, in terms of the parameTeR $\lambdA$, the Mp level and the nOrm of the MP SOLUtion.
ResPect TO tHE second solutioN, we imPose the conDItions ($F_2$) aNd ($F_3$) on $F$, which wiDely known WHEn applyiNg tHe sUb- aNd sUPEr-Solution methoD IN proBlEms like (\[p\]) wiTh $s=1$. Here We sHow ThaT thIs Idea still Remains hOlD bY dEfIniNg proPErly sub- aNd SupEr-SolUtionS In our nOnlocAl coNtExT, It iS when $0<s<1$.
THE dOCUmenT iS oRganIzeD aS follOws: iN secTion 2 we oFfer a brieF reVIew oF tHe FractioNal spaces of SoBoLev in the coNtExt Of our pROBlem and rEmember some useful resultS. in SectiOn 3 wE look For tHe first soLutIon, whiLe iN sectioN 3 we finD the sEcOnd SOLutioN, WHiCh eNdS the proof oF tHeoRem \[maIn-TheoRem0\]
FuncTional framework and PRelIminaries {#sectIon2}
======================================
in thIS SeCtiON, wE OffEr A | 0\] Let $f\colon \mathbb{R } \to [0,\ infty )$beacont inuo us function ve r ifyi ng $(F_1)-(F_3)$. Then , the re exis t s$\bar {\lambd a }> 0 $ su ch t hat f o r$\lam bda >\bar{\ lambda}$ t hepr oblem $(P_{\ l am bda})$ adm its two nonnega tiv e solu ti ons $u_{\ lam bda}, v_{\l a mbda}\in X^s_0 (\ O mega)$ . Moreov e r ,$0\l e u_\lambda < v_{ \ la m bda}\le 1$.
T he mai ni de a s be hin d of the p ro of re l ies onk no w n arg u ments for sol ving this t y peof pro bl ems . Howev er, t hi s is sue is non- triv ial, by w hich w e need t o treatthe pr obl ema und e ran ap pr o pri a te ap p roa ch thatal lo ws to app l y s ucharg umen ts an d, in this wa y,to s a vesever al te chni ca l dif ficult ies t ha t arise in thenonl ocal case .
Sp eci fi cally , we ob tai n t he firs t solut i on$u _ { \l ambda}$ by truncat in g th e nonlin earity an dt aking $\ la mbd a$ l a r ge en ough , w here wehave a s su me d onlyth e loca lcon dit ion a t zer o ($F_ 1$). The beha v ior of the nor m s, accordingt ot h ep aram ete r $\lambda$ of theposs i bl e s o lutio ns, i sb as e d on the De Giorgi- Na sh-Mos er th eory, precise ly on Mose r ’ s iterati ve s c he m e. Such proced ure w as introdu c ed in th e mid 1950s a nd early1 9 60s, weref erto[@D e G io rgi; @Moser]. A n im po rtant f act in the pr oof is th at the func tional e ne rg yas soc iated with the t run ca ted prob l em has theMoun ta in Pas s (MP)g eo m e try. T hu s, w e c an cont rol, interms o f the par ame t er $ \l am bda$, t he MP level a nd the normof th e MP s o l ution.
Respect to the second s o lution, we impo se t he condit ion s ($F_ 2$) and ($ F_3$)on $f $, wh i c h wid e l ykno wn when appl y i ngthe s ub - an d super -solution method i n pr oblems like ( \[P \])w i th $s = 1$ . He re wes h ow that this id ea still r em a in s hold byd efi ni ng prop erly su b- an d super- solutions in our n on loca l con text, it i s when $ 0<s<1$.
T he do c um ent i s o rganiz ed as foll ows: i n Se ction 2 weof fer abrief r eview of the fractional spacesof Sob olevinthe conte xto f o ur proble m an d remember so meusefu l r e sults . In Se cti o n 3 w e lo o k for the fi rst s ol ution, whil e i n S ectio n 3 we fin d th e second solution , which ends th e pr o o f o f T h eore m\[main-theorem 0\]
F u nctional f ramework an d prelim in a ries{#sect ion2}======= = = == = ====== ==== === ========= === =In this s ec t ion, w e of fe r a | 0\] Let_$f\colon \mathbb{R}_\to [0,\infty)$ be a_continuous function_verifying_$(F_1)-(F_3)$. Then,_there_exists $\bar{\lambda}>0$ such_that for $\lambda>\bar{\lambda}$_the problem $(P_{\lambda})$ admits_two nonnegative solutions_$u_{\lambda},_v_{\lambda} \in X^s_0(\Omega)$. Moreover, $0\le u_\lambda < v_{\lambda}\le 1$.
The main ideas behind of the_proof_relies on_known_arguments_for solving this type of_problems. However, this issue is_non-trivial, by_which we need to treat the problema under_an_appropriate approach that_allows to apply such arguments and, in this way,_to save several technical difficulties that_arise in the_nonlocal_case.
Specifically,_we obtain the first_solution $u _{\lambda}$ by truncating the_nonlinearity and taking $\lambda$ large enough,_where we have assumed only the local_condition at zero ($F_1$). The behavior_of the norms, according to_the parameter_$\lambda $ of the possible_solutions, is based_on the_De Giorgi-Nash-Moser theory,_precisely on Moser’s iterative scheme. Such_procedure was introduced_in the mid 1950s and early_1960s,_we refer to_[@DeGiorgi;_@Moser]._An important_fact in the_proof_is that_the_functional energy associated with the truncated_problem_has the Mountain Pass (MP) geometry. Thus,_we can control, in_terms_of the parameter $\lambda$,_the MP level and the_norm of the MP solution.
Respect to_the second_solution, we_impose the conditions ($F_2$) and ($F_3$) on $f$, which widely known_when applying the sub- and super-solution_method in problems like_(\[P\]) with_$s=1$._Here we show_that_this idea_still remains hold by defining properly sub-_and super-solutions_in our nonlocal context, it is_when $0<s<1$.
The document is_organized_as follows: in Section 2 we_offer a brief review of the_fractional spaces of Sobolev in_the_context_of our problem and remember_some useful results. In Section 3_we look for_the first solution, while in Section 3_we_find the second solution, which ends_the_proof of Theorem \[main-theorem0\]
Functional framework and_preliminaries_{#section2}
======================================
In_this section, we offer a |
fact is that For $K = 1 $ the local minimum is not realized at all even for weak field strengths. Actually, the value of $K =0.75$ corresponding $m_H \sim 70
$Gev is a bound value for the mass $m_H$ and the field strengths $H
\sim 0.01 - 0.1 M^2_w/e = 0.01 - 0.1 \cdot 10^{24}$ G give upper limits on the magnetic fields in the phase transition epoch. Stronger fields stimulates straight transition to the global minimum.Thus we see that due to $V_{ring}$ term the upper limit on the magnetic field is decreased from $2 \cdot 10^{24} G$ derived with the EP $V^{(1)}$ to $0.1 \cdot 10^{24} G$. Moreover, they restrict the Higgs boson mass: $m_H < 70 Gev$, otherwise the local electroweak minimum is not produced. We also observe that the phase transition is to be of the first type one.
It is very important that the described minima of the EP are stable at high temperatures even for strong fields $h$. Really, typical value of $B$ when the symmetry restoration happens for $K \sim 1$ is $ B \sim
0.1$. For the calculated lower temperature mass we have $
\frac{1}{12} g^2/B^2 \sim 3 $. Hence we find that the effective mass of the unstable mode $ \phi^2 - h + \frac{1}{12} g^2/B^2 $ is positive for all values of $h$ considered. Thus, one has to conclude that classical constant magnetic field must be stable in the electroweak phase transition epoch. No $W$- and $Z$-boson condensates can be produced at high temperatures. These condensates would be realized at lower temperatures as the intermediate states of the vacuum.
6. Discussion {#discussion.unnumbered}
-------------
Let us discuss the vacuum stability in a magnetic field. This problem is of interest because of the presence in the $W$-boson spectrum the mode $p^2_0 = M^2_w - eH$ which becomes unstable for $H >
M^2_w/e$. The evolution of this state and its consequences have been investigated in various aspects by many authors [@AO2],[@Sk2],[@AO]. At high temperatures, | fact is that For $ K = 1 $ the local minimum is not realized at all even for weak sphere strength. Actually, the value of $ K = 0.75 $ corresponding $ m_H \sim 70
$ Gev is a bounce value for the multitude $ m_H$ and the field strengths $ H
\sim 0.01 - 0.1 M^2_w / e = 0.01 - 0.1 \cdot 10^{24}$ thousand collapse upper limits on the charismatic field in the phase conversion epoch. Stronger fields provoke straight transition to the global minimum. Thus we see that due to $ V_{ring}$ condition the upper limit on the magnetic field is decrease from $ 2 \cdot 10^{24 } G$ derived with the EP $ V^{(1)}$ to $ 0.1 \cdot 10^{24 } G$. Moreover, they restrict the Higgs boson mass: $ m_H < 70 Gev$, otherwise the local electroweak minimum is not produce. We also observe that the phase transition is to be of the first type one.
It is very authoritative that the described minima of the EP are stable at high temperatures even for strong fields $ h$. Really, typical value of $ B$ when the symmetry restoration happens for $ K \sim 1 $ is $ B \sim
0.1$. For the calculated lower temperature mass we have $
\frac{1}{12 } g^2 / B^2 \sim 3 $. Hence we recover that the effective mass of the unstable mood $ \phi^2 - h + \frac{1}{12 } g^2 / B^2 $ is positive for all values of $ h$ considered. Thus, one have to conclude that classical constant magnetic airfield must be stable in the electroweak phase transition epoch. No $ W$- and $ Z$-boson condensates can be produced at gamey temperatures. These condensates would be realized at lower temperature as the intermediate state of matter of the vacuum.
6. Discussion { # discussion.unnumbered }
-------------
Let us discuss the vacuum stability in a charismatic field. This problem is of interest because of the presence in the $ W$-boson spectrum the mode $ p^2_0 = M^2_w - eH$ which becomes unstable for $ H >
M^2_w / e$. The evolution of this state and its consequence have been investigated in various aspect by many authors [ @AO2],[@Sk2],[@AO ]. At eminent temperatures, | faft is that For $K = 1 $ the uocal minimum is not rxalized at all dven for weak field strengthd. Qctuaoly, the value of $K =0.75$ cofrespondijg $m_H \sin 70
$Get is a bound valns for tmz masa $m_H$ cnv the field strgngths $H
\sim 0.01 - 0.1 M^2_w/e = 0.01 - 0.1 \cdod 10^{24}$ G give upper limits on the magnetic sields on the phase tragsitpog epkbh. Stronger fields stimulates sfraight transition tp the global minimum.Thus wf sef that due to $V_{rinh}$ term the olpew limit on thd magnetic field is dedreased from $2 \cdot 10^{24} G$ derived wkth tke EP $V^{(1)}$ to $0.1 \ceot 10^{24} G$. Moreover, they restrict thc Higgs boson kass: $m_H < 70 Gev$, pthxrwiwe the local electrowxak minimum is not ptoduced. We amso observe that rhw phave tsanskriov ia vo ge of hhe first typs one.
It is cery important that trv described mjnima jf the EP are stable at high temperatures evsn for strong fields $h$. Eeally, typical value lf $B$ when the symmetry restoration happens for $K \sim 1$ is $ T \sim
0.1$. Wor tme cxocklated lower temperature mass we have $
\frac{1}{12} g^2/B^2 \aik 3 $. Hence we finb that the effecyige iass of the uvstablz mkde $ \phi^2 - h + \frac{1}{12} h^2/B^2 $ is kositice for alj vakues of $h$ considered. Thus, obe has to couclyde that classical constant mcgnetiv fiekd must be stable in thz elecfroweak phade transifkon epoch. No $W$- avd $D$-bosmn condensates can be prodtced at hmgh tzmperatufes. Jhese cjndensates would be realized at loaer tgmperadures as tje intermediate states of the veruum.
6. Discussipn {#divcussion.bnnumbcred}
-------------
Let us disctss the vacuum stabilnty in a magnetib field. Tiis problem ys of interesd because of vhe presegce un tye $W$-bosuv spectrum the mode $p^2_0 = M^2_w - eH$ whixh becomes unstablc for $G >
M^2_w/e$. The evolucnob of this state ana ies cpnfaquences hava bedn kmvestkgated in vcrkous aspects by many autvors [@AO2],[@Sk2],[@AO]. At high tekpcratures, | fact is that For $K = 1 local is not at all even the of $K =0.75$ $m_H \sim 70 is a bound value for the $m_H$ and the field strengths $H \sim 0.01 - 0.1 M^2_w/e = 0.01 0.1 \cdot 10^{24}$ G give upper limits on the magnetic fields in the transition Stronger stimulates transition to the global minimum.Thus we see that due to $V_{ring}$ term the upper limit on magnetic field is decreased from $2 \cdot 10^{24} derived with the EP to $0.1 \cdot 10^{24} G$. they the Higgs mass: < Gev$, otherwise the electroweak minimum is not produced. We also observe that the phase transition is to be of the type one. very important the minima the EP are high temperatures even for strong fields value of $B$ when the symmetry restoration happens $K \sim is $ B \sim 0.1$. For calculated lower temperature mass we have $ \frac{1}{12} \sim 3 $. Hence we find that the effective mass of the unstable mode $ h + \frac{1}{12} g^2/B^2 is positive for values $h$ Thus, has to that classical constant magnetic field must be stable in the electroweak transition epoch. No $W$- and $Z$-boson condensates can be produced temperatures. condensates would be at lower temperatures as intermediate of the vacuum. 6. ------------- us stability a field. This problem is interest because of the presence the $W$-boson spectrum the eH$ which becomes unstable for $H > M^2_w/e$. evolution of this state and its consequences been investigated in various aspects by many authors [@AO2],[@Sk2],[@AO]. At high temperatures, | fact is that For $K = 1 $ the local minImum is not rEalizEd aT alL eVen fOr weAk field strengtHS. ActUally, the value of $K =0.75$ corresPondiNg $M_h \sim 70
$gEv Is a boUnd valuE FoR THe mAsS $m_h$ anD tHE fIeld sTreNgths $H
\sIm 0.01 - 0.1 M^2_w/e = 0.01 - 0.1 \cdot 10^{24}$ G GivE uPper limits on THe Magnetic fiEldS in the phase tRanSition EpOch. sTrongEr fIelds StimulATes strAight tranSiTIon to tHE global MINiMum.THus we see that due to $v_{RiNG}$ term the upper lImit on ThE MaGNEtiC fiEld is decreAsEd froM $2 \Cdot 10^{24} G$ deRIvED WIth THe EP $V^{(1)}$ to $0.1 \cdot 10^{24} G$. MOreover, they REstRict thE HIggS Boson mAss: $m_H < 70 geV$, OthErwise the loCal eLectroweaK minimUM is not pROduced. WE also oBseRve That THe PhAse TrANsiTIoN is TO be Of the firSt TyPe one.
it is VERY ImpoRtaNt thAt the Described miniMa oF the ep arE stabLe at hIgh tEmPeratUres evEn for StRong fields $h$. RealLy, tyPical valuE of $b$ wHen ThE symmETry resTorAtiOn happeNs for $K \sIM 1$ is $ b \sIM
0.1$. fOr The calculated lower TeMPErAture masS we havE $
\FrAc{1}{12} G^2/b^2 \sim 3 $. HencE wE fiNd thAT The efFectIVe Mass of thE unstaBLe MoDe $ \phi^2 - h + \fRaC{1}{12} g^2/B^2 $ is pOsItiVe fOr all VAlueS of $h$ coNsidered. thus, oNE has to conclude THat classical cONsTANt MAgneTic Field must be StabLE in tHe elECtRowEAk phaSe traNsITiON epoch. No $W$- and $Z$-boson cOnDensatEs can Be produced at hIgh temperaTUREs. These cOndeNSaTEs would be realiZed at Lower tempeRAtures as The inTermediaTe states oF THe vacuum.
6. disCusSioN {#diSCUsSion.unnumbereD}
-------------
lEt us DiScuss thE vaCuum staBilIty In a MagNeTic field. THis problEm Is Of InTerEst beCAuse of thE pResEnCe iN the $W$-BOson spEctruM the MoDe $P^2_0 = m^2_w - eh$ which bECoMES unsTaBlE for $h >
M^2_w/E$. THe evoLutiON of This staTe and its cOnsEQuenCeS hAve been Investigated iN vArious aspeCtS by Many auTHOrs [@AO2],[@Sk2],[@Ao]. At high temperatures, | fact is that For $K = 1 $ the local mini mum is n ot r eali zed at all eve n for weak field strengths. Actu al l y, t h evalue of $K= 0. 7 5 $ c or re spo nd i ng $m_H \s im 70
$ Gev is a b oun dvalue for th e m ass $m_H$and the field s tre ngths$H
\s i m 0.0 1 - 0.1M^2_w/ e = 0.0 1 - 0.1 \ cd o t 10^{ 2 4}$ G g i v euppe r limits on the m a gn e tic fields inthe ph as e t r a nsi tio n epoch. S tr onger fieldss ti m u l ate s straight tra nsition tot heglobal m ini m um.Thu s wese e th at due to $ V_{r ing}$ ter m theu pper li m it on t he mag net icfiel d i sdec re a sed fr om$ 2 \ cdot 10^ {2 4} G$ d eriv e d w iththe EP$V^{( 1)}$ to $0.1\cd ot 1 0 ^{2 4} G$ . Mor eove r, they restr ict t he Higgs boson ma ss:$m_H < 70 Ge v$ , o th erwis e the l oca l e lectrow eak min i mum i s n ot produced. We also o b s er ve thatthe ph a se t r ansition i s t o be o f the fir s ttype one .
Iti sve ry impo rt ant th at th e d escri b ed m inimaof the E P are stable at high temperaturese ve n fo r str ong fields $h$ . Re a lly, typ i ca l v a lue o f $B$ w h en the symmetry restor at ion ha ppens for $K \sim1$ is $ B\ s i m
0.1$.Fort he calculated low er te mperaturem ass we h ave $
\frac{1 }{12} g^2 / B ^2 \sim3 $ . H enc e w e fi nd that the e f f ecti ve mass o f t he unst abl e m ode $\p hi^2 - h+ \frac{ 1} {1 2} g ^2/ B^2 $ is posit iv e f or al l val u es of$h$ c onsi de re d . T hus, on e h a s toco nc lude th at clas sica l co nstantmagneticfie l d mu st b e stabl e in the elec tr oweak phas etra nsitio n epoch. N o $W$- and $Z$-boson co n densate s c an be pro duced athig h temp era t ures.Theseconde ns ate s would b erea li zed at low e r te mpera tu resas theintermediate state s of the vacuum.
6. Dis c u ss ion {# d isc us s ion . u nnumbered}
---- ---------
L e tus discuss the v acuum s tabilit y ina magnet ic field. This pro bl em i s ofinterest b ecause o f the pre s encei nthe $ W$- bosonsp ect rum t he mod e $p ^2_0= M^2_ w- eH$which b ecomes u nstable for $H >
M^2_w/ e$. Th e evo lut ion of th iss tat e and its con sequenceshav e b een i nve s tigat ed i n v ari o us as pect s by manya ut hor s [@ AO2],[@Sk2] , [ @ AO] . Athig h tempe ratu res, | fact_is that_For $K = 1_$ the_local_minimum is_not_realized at all_even for weak_field strengths. Actually, the_value of $K_=0.75$_corresponding $m_H \sim 70
$Gev is a bound value for the mass $m_H$ and the_field_strengths $H
\sim_0.01_-_0.1 M^2_w/e = 0.01 -_0.1 \cdot 10^{24}$ G give_upper limits_on the magnetic fields in the phase transition_epoch._Stronger fields stimulates_straight transition to the global minimum.Thus we see that_due to $V_{ring}$ term the upper_limit on the_magnetic_field_is decreased from $2_\cdot 10^{24} G$ derived with the_EP $V^{(1)}$ to $0.1 \cdot 10^{24}_G$. Moreover, they restrict the Higgs boson_mass: $m_H < 70 Gev$, otherwise_the local electroweak minimum is_not produced._We also observe that the_phase transition is_to be_of the first_type one.
It is very important that_the described minima_of the EP are stable at_high_temperatures even for_strong_fields_$h$. Really,_typical value of_$B$_when the_symmetry_restoration happens for $K \sim 1$_is_$ B \sim
0.1$. For the calculated lower_temperature mass we have_$
\frac{1}{12}_g^2/B^2 \sim 3 $._Hence we find that the_effective mass of the unstable mode_$ \phi^2_- h_+ \frac{1}{12} g^2/B^2 $ is positive for all values of $h$_considered. Thus, one has to conclude_that classical constant magnetic_field must_be_stable in the_electroweak_phase transition_epoch. No $W$- and $Z$-boson condensates can_be produced_at high temperatures. These condensates would_be realized at lower_temperatures_as the intermediate states of the_vacuum.
6. Discussion {#discussion.unnumbered}
-------------
Let us discuss the_vacuum stability in a magnetic_field._This_problem is of interest because_of the presence in the $W$-boson_spectrum the mode_$p^2_0 = M^2_w - eH$ which becomes_unstable_for $H >
M^2_w/e$. The evolution of_this_state and its consequences have been_investigated_in_various aspects by many authors_[@AO2],[@Sk2],[@AO]. At high temperatures, |
1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q }(x)\oplus E^c(x)\oplus E^u_1(x)\oplus E^u_2(x)\oplus \dots \oplus E^u_{q }(x)$$ such that:
1. for every $v\in E^s_j(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=-\lambda_j}$;
2. for every $v\in E^c(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=0}$;
3. for every $v\in E^u_j(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=+\lambda_j}$.\
We may ask, too, that: $\forall x\in N, \dim(G_-(x)\cap G_+(x))=p$.
Let us recall that the stable bundle $E^s(x)=E^s_1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q }(x)$ and the unstable one $E^u(x)=E^u_1(x)\oplus E^u_2(x)\oplus\dots\oplus E^u_{q }(x)$ are isotropic (for the symplectic form) and that $E^c(x)$ is a symplectic subspace of $T_x(T^*M)$ that is orthogonal (for $\omega$) to $E^s(x)\oplus E^u(x)$. Moreover, we have: $\dim E^s_i=\dim E^u_i$.
Reduction of the problem
------------------------
As in the statement of theorem \[thdimlyap\], we assume that $\mu$ is a minimizing ergodic measure whose support is not reduced to a point and that $p\in [1, n]$ is so that | 1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q } (x)\oplus E^c(x)\oplus E^u_1(x)\oplus E^u_2(x)\oplus \dots \oplus E^u_{q } (x)$$ such that:
1. for every $ v\in E^s_j(x)\backslash\ { 0\}$; $ \displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left (\| D\varphi_t(x)v\| \right)=-\lambda_j}$;
2. for every $ v\in E^c(x)\backslash\ { 0\}$; $ \displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left (\| D\varphi_t(x)v\| \right)=0}$;
3. for every $ v\in E^u_j(x)\backslash\ { 0\}$; $ \displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left (\| D\varphi_t(x)v\| \right)=+\lambda_j}$.\
We may ask, too, that: $ \forall x\in N, \dim(G_-(x)\cap G_+(x))=p$.
Let us recall that the static package $ E^s(x)=E^s_1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q } (x)$ and the unstable one $ E^u(x)=E^u_1(x)\oplus E^u_2(x)\oplus\dots\oplus E^u_{q } (x)$ are isotropic (for the symplectic form) and that $ E^c(x)$ is a symplectic subspace of $ T_x(T^*M)$ that is extraneous (for $ \omega$) to $ E^s(x)\oplus E^u(x)$. Moreover, we receive: $ \dim E^s_i=\dim E^u_i$.
Reduction of the problem
------------------------
As in the statement of theorem \[thdimlyap\ ], we bear that $ \mu$ is a minimizing ergodic measure whose documentation is not reduced to a point and that $ p\in [ 1, n]$ is so that | 1(x)\oppus E^s_2(x)\oplus\dots\oplus E^s_{d }(x)\oplus E^c(x)\oplos E^u_1(x)\opnus E^u_2(s)\oplus \duts \oplus E^u_{q }(x)$$ such that:
1. flr evert $v\in E^s_j(x)\backslash\{ 0\}$; $\dksplaystype{\lim_{t\rithtacrow\pm \infty}\frac{1}{v}\mog\left( \| D\varlmi_t(x)v\| \cight)=-\lambda_j}$;
2. fpr every $v\hn E^c(x)\backslasv\{ 0\}$; $\dnsplaystyle{\lim_{t\rightarrow\pm \infty}\fras{1}{t}\log\legt( \| D\varphi_t(x)v\| \ryght)=0}$;
3. for vvtry $v\in E^u_j(x)\backslash\{ 0\}$; $\displaystyme{\lim_{t\rpghtarrow\pm \infty}\grac{1}{t}\log\left( \| D\varphi_t(x)v\| \rlght)=+\pambda_j}$.\
We may wsk, too, thaj: $\fowqll x\in N, \dio(G_-(x)\cap G_+(x))=p$.
Let us recalm that the stable bundle $E^s(x)=E^s_1(x)\uplus E^s_2(x)\oplus\ditw\oppos E^s_{q }(x)$ and the lnstable one $C^l(x)=E^u_1(x)\opnus E^u_2(x)\pplus\dots\oplus E^u_{x }(x)$ are isotropic (for thx symplectic form) anq that $E^c(f)$ ns a symplectic subspqcw of $J_x(T^*M)$ dhat us urtgojonzl (for $\omxga$) to $E^s(x)\ollus E^u(x)$. Moeeover, we have: $\dim T^s_i=\qpk E^u_i$.
Reductikn of ehq problem
------------------------
As in the statement of theorem \[thsimlyap\], we assume that $\nu$ is a minimizing erhodic meafure whose support is not reduced to a point and dhat $'\iv [1, u]$ is su tjat | 1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q }(x)\oplus E^c(x)\oplus E^u_1(x)\oplus E^u_2(x)\oplus E^u_{q such that: for every $v\in D\varphi_t(x)v\| 2. for every E^c(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \| D\varphi_t(x)v\| \right)=0}$; 3. for every E^u_j(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=+\lambda_j}$.\ We may ask, too, that: $\forall N, \dim(G_-(x)\cap G_+(x))=p$. Let us recall that the stable bundle $E^s(x)=E^s_1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{q and unstable $E^u(x)=E^u_1(x)\oplus E^u_{q }(x)$ are isotropic (for the symplectic form) and that $E^c(x)$ is a symplectic subspace of that is orthogonal (for $\omega$) to $E^s(x)\oplus E^u(x)$. we have: $\dim E^s_i=\dim Reduction of the problem ------------------------ in statement of \[thdimlyap\], assume $\mu$ is a ergodic measure whose support is not reduced to a point and that $p\in [1, n]$ is so | 1(x)\oplus E^s_2(x)\oplus\dots\oplus E^s_{Q }(x)\oplus E^c(x)\Oplus e^u_1(x)\OplUs e^u_2(x)\oPlus \Dots \oplus E^u_{q }(x)$$ sUCh thAt:
1. for every $v\in E^s_j(x)\backsLash\{ 0\}$; $\dIsPLaysTYlE{\lim_{t\RightarROw\PM \InfTy}\FrAc{1}{t}\LoG\LeFt( \| D\vaRphI_t(x)v\| \rigHt)=-\lambda_j}$;
2. fOr eVeRy $v\in E^c(x)\backSLaSh\{ 0\}$; $\displaysTylE{\lim_{t\rightarRow\Pm \inftY}\fRac{1}{T}\Log\leFt( \| D\VarphI_t(x)v\| \riGHt)=0}$;
3. for eVery $v\in E^u_J(x)\BAckslaSH\{ 0\}$; $\displaYSTyLe{\liM_{t\rightarrow\pm \infTY}\fRAc{1}{t}\log\left( \| D\varPhi_t(x)v\| \RiGHt)=+\LAMbdA_j}$.\
WE may ask, too, ThAt: $\forALl x\in N, \dIM(G_-(X)\CAP G_+(x))=P$.
let us recall thAt the stable BUndLe $E^s(x)=E^S_1(x)\OplUS E^s_2(x)\opLus\doTs\OPluS E^s_{q }(x)$ and the UnstAble one $E^u(X)=E^u_1(x)\opLUs E^u_2(x)\opLUs\dots\oPlus E^u_{Q }(x)$ aRe iSotrOPiC (fOr tHe SYmpLEcTic FOrm) And that $E^C(x)$ Is A sympLectIC SUBspaCe oF $T_x(T^*m)$ that Is orthogonal (fOr $\oMega$) TO $E^s(X)\opluS E^u(x)$. MOreoVeR, we haVe: $\dim E^S_i=\dim e^u_I$.
Reduction of the ProbLem
------------------------
As in thE stAtEmeNt Of theORem \[thdImlYap\], We assumE that $\mu$ IS a mInIMIZiNg ergodic measure whOsE SUpPort is noT reducED tO a POint and tHaT $p\iN [1, n]$ is SO That | 1(x)\oplus E^s_2(x)\oplus\ dots\oplus E^s_ {q}(x )\ oplu s E^ c(x)\oplus E^u _ 1(x) \oplus E^u_2(x)\oplus\dots \ o plus E^ u_{q}(x)$$s uc h tha t:
1. f o revery $v \in E^s _j(x)\back sla sh \{ 0\}$; $\d i sp laystyle{\ lim _{t\rightarr ow\ pm \in ft y}\ f rac{1 }{t }\log \left( \| D\v arphi_t(x )v \ | \rig h t)=-\la m b da _j}$ ;
2. for every$ v\ i n E^c(x)\backs lash\{ 0 \ }$ ; $\d isp laystyle{\ li m_{t\ r ightarr o w\ p m \in f ty}\frac{1}{t }\log\left( \|D\varp hi _t( x )v\| \ right )= 0 }$;
3. for e very $v\in E^ u_j(x) \ backsla s h\{ 0\} $; $\d isp lay styl e {\ li m_{ t\ r igh t ar row \ pm\infty}\ fr ac {1}{t }\lo g \ l e ft(\|D\va rphi_ t(x)v\| \righ t)= +\la m bda _j}$. \
Wema y ask , too, that :$\forall x\in N , \d im(G_-(x) \ca pG_+ (x ))=p$ .
Letusrec all tha t the s t abl eb u n dl e $E^s(x)=E^s_1(x) \o p l us E^s_2(x )\oplu s \d ot s \oplus E ^s _{q }(x ) $ andtheu ns table on e $E^u ( x) =E ^u_1(x) \o plus E ^u _2( x)\ oplus \ dots \oplus E^u_{q}(x)$ are isotropic( for the sympl e ct i c f o rm)and that $E^c( x)$i s asymp l ec tic subsp ace o f$ T_ x (T^*M)$ that is ort ho gonal(for$\omega$) to$E^s(x)\op l u s E^u(x)$ . Mo r eo v er, we have: $ \dimE^s_i=\dim E^u_i$.
Redu ction of the prob l e m
------ --- --- --- --- - - -- --
As in the s tate me nt of t heo rem \[t hdi mly ap\ ],we assume t hat $\mu $is a m ini mizin g ergodic m eas ur e w hoses upport is n ot r ed uc e d t o a poi n ta n d th at $ p\in [1 ,n]$ i s so tha t | 1(x)\oplus E^s_2(x)\oplus\dots\oplus_E^s_{q }(x)\oplus_E^c(x)\oplus E^u_1(x)\oplus E^u_2(x)\oplus \dots_\oplus E^u_{q_}(x)$$_such that:
1.__for every $v\in_E^s_j(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm_\infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=-\lambda_j}$;
2._ for every_$v\in_E^c(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm \infty}\frac{1}{t}\log\left( \| D\varphi_t(x)v\| \right)=0}$;
3. for every $v\in E^u_j(x)\backslash\{ 0\}$; $\displaystyle{\lim_{t\rightarrow\pm_\infty}\frac{1}{t}\log\left(_\| D\varphi_t(x)v\|_\right)=+\lambda_j}$.\
__ We may ask,_too, that: $\forall x\in N,_\dim(G_-(x)\cap G_+(x))=p$.
Let_us recall that the stable bundle $E^s(x)=E^s_1(x)\oplus E^s_2(x)\oplus\dots\oplus_E^s_{q_}(x)$ and the_unstable one $E^u(x)=E^u_1(x)\oplus E^u_2(x)\oplus\dots\oplus E^u_{q }(x)$ are isotropic (for_the symplectic form) and that $E^c(x)$_is a symplectic_subspace_of_$T_x(T^*M)$ that is orthogonal_(for $\omega$) to $E^s(x)\oplus E^u(x)$. Moreover,_we have: $\dim E^s_i=\dim E^u_i$.
Reduction of_the problem
------------------------
As in the statement of theorem_\[thdimlyap\], we assume that $\mu$ is_a minimizing ergodic measure whose_support is_not reduced to a point_and that $p\in_[1, n]$_is so that |
G_+(x)=E^u(x)\oplus R(x) $
Weak K.A.M. solutions and Green bundles {#sec3}
=======================================
In this section, we recall the weak KAM theory and give a relationship between some tangent cones to the pseudographs of the weak KAM solutions and the Green bundles. These results imply theorem \[th1\]. The proofs are given in section \[sec4\].
Weak KAM theory
---------------
We don’t give any proof in this section, but all the results that we give are proved in [@Fa1] or [@Be1].
If $t>0$, the function $A_t~: M\times M\rightarrow {\mathbb {R}}$ is defined by: $$A_t(q_0, q_1)=\inf_\gamma \int_0^tL(\gamma (s), \dot\gamma (s))ds=\min_\gamma \int_0^tL(\gamma (s), \dot\gamma (s))ds$$ where the infimum is taken on the set of $C^2$ curves $\gamma~: [0, t]\rightarrow M$ such that $\gamma (0)=q_0$ and $\gamma (t)=q_1$.
1. A function $v~: V\rightarrow {\mathbb {R}}$ defined on a subset $V$ of ${\mathbb {R}}^d$ is [*$K$-semi-concave*]{} if for every $x\in V$, there exists a linear form $p_x$ defined on ${\mathbb {R}}^d$ so that: $$\forall y\in V, v(y)\leq v(x)+p_x(y-x)+K\| y-x\|^2.$$ Then we say that $p_x$ is a [*K-super-differential*]{} of $v$ at $x$.
2. Let us fix a finite atlas ${\cal A}$ of the manifold $M$; a function $u~: M\rightarrow {\mathbb {R}}$ is [*$K$-semi-concave*]{} if for every chart $(U, \phi)$ belonging to ${\cal A}$, $u\circ\phi^{-1}$ is $K$-semi-concave. Then a [*$K$-super-differential*]{} of $u$ at $q$ is $p_x\circ | G_+(x)=E^u(x)\oplus R(x) $
Weak K.A.M. solutions and Green bundles { # sec3 }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
In this incision, we echo the weak KAM theory and give a kinship between some tangent cones to the pseudographs of the weak KAM solutions and the Green package. These results imply theorem \[th1\ ]. The proof are given in section \[sec4\ ].
decrepit KAM theory
---------------
We do n’t give any validation in this section, but all the results that we give are proved in [ @Fa1 ] or [ @Be1 ].
If $ t>0 $, the function $ A_t~: M\times M\rightarrow { \mathbb { R}}$ is defined by: $ $ A_t(q_0, q_1)=\inf_\gamma \int_0^tL(\gamma (s), \dot\gamma (s))ds=\min_\gamma \int_0^tL(\gamma (s), \dot\gamma (s))ds$$ where the infimum is taken on the stage set of $ C^2 $ curves $ \gamma~: [ 0, t]\rightarrow M$ such that $ \gamma (0)=q_0 $ and $ \gamma (t)=q_1$.
1. A function $ v~: V\rightarrow { \mathbb { R}}$ defined on a subset $ V$ of $ { \mathbb { R}}^d$ is [ * $ K$-semi - concave * ] { } if for every $ x\in V$, there exist a linear form $ p_x$ defined on $ { \mathbb { R}}^d$ so that: $ $ \forall y\in V, v(y)\leq v(x)+p_x(y - x)+K\| y - x\|^2.$$ Then we say that $ p_x$ is a [ * K - ace - differential * ] { } of $ v$ at $ x$.
2. Let us cook a finite atlas $ { \cal A}$ of the manifold $ M$; a function $ u~: M\rightarrow { \mathbb { R}}$ is [ * $ K$-semi - concave * ] { } if for every chart $ (U, \phi)$ belonging to $ { \cal A}$, $ u\circ\phi^{-1}$ is $ K$-semi - concave. Then a [ * $ K$-super - derived function * ] { } of $ u$ at $ q$ is $ p_x\circ | G_+(x)=E^k(x)\oplus R(x) $
Weak K.A.M. solmtions and Green bundlev {#sec3}
=======================================
Ih this sdction, we recall the weak KAL rheort and give a relationsfip betwevn some tqngeit cones to the 'aeudographs or the xeak KAM solutipns and tha Green bundlev. Ghzse results imply theorem \[th1\]. The projfs are glven in sectiog \[seb4\].
Wqak IAM theory
---------------
We don’t give any proof ih this vection, but akl the results that we givf arf proved in [@Fa1] or [@He1].
If $t>0$, the dunceuon $A_t~: M\timer M\rightargmw {\mathbb {T}}$ is defined by: $$A_t(q_0, q_1)=\inf_\gamma \ing_0^tL(\gakma (s), \dot\gqmna (d))gs=\min_\gamma \mnt_0^tL(\gwmma (s), \dot\gamma (s))ds$$ fhere tne infimum is baken on the set of $C^2$ curves $\jamma~: [0, t]\rightarrow M$ such thad $\yamma (0)=q_0$ and $\gamma (t)=q_1$.
1. A funcjion $e~: V\rkthtxrrkw {\mzthbb {G}}$ dxfined on a subset $V$ od ${\mathbb {R}}^d$ is [*$K$-semo-cjbcave*]{} if for svery $v\ig V$, there exists a linear form $p_x$ defintd on ${\mathbb {R}}^d$ so that: $$\foraol y\in V, v(y)\leq v(x)+p_x(y-x)+N\| y-x\|^2.$$ Then we say that $p_x$ is a [*K-super-differential*]{} of $v$ at $x$.
2. Let js ylw a wunlte atlas ${\cal A}$ of the manifold $M$; a function $t~: M\tinhtarrow {\mathbb {V}}$ is [*$K$-semi-concave*]{} of fpt every chart $(O, \phi)$ bzmohging to ${\cal A}$, $u\cigc\phi^{-1}$ if $K$-seni-concave. Them a [*$K$-super-differential*]{} of $u$ at $q$ is $p_x\cprc | G_+(x)=E^u(x)\oplus R(x) $ Weak K.A.M. solutions and {#sec3} In this we recall the a between some tangent to the pseudographs the weak KAM solutions and the bundles. These results imply theorem \[th1\]. The proofs are given in section \[sec4\]. KAM theory --------------- We don’t give any proof in this section, but all results we are in [@Fa1] or [@Be1]. If $t>0$, the function $A_t~: M\times M\rightarrow {\mathbb {R}}$ is defined by: q_1)=\inf_\gamma \int_0^tL(\gamma (s), \dot\gamma (s))ds=\min_\gamma \int_0^tL(\gamma (s), \dot\gamma where the infimum is on the set of $C^2$ $\gamma~: t]\rightarrow M$ that (0)=q_0$ $\gamma (t)=q_1$. 1. function $v~: V\rightarrow {\mathbb {R}}$ defined on a subset $V$ of ${\mathbb {R}}^d$ is [*$K$-semi-concave*]{} if for $x\in V$, a linear $p_x$ on {R}}^d$ so that: V, v(y)\leq v(x)+p_x(y-x)+K\| y-x\|^2.$$ Then we is a [*K-super-differential*]{} of $v$ at $x$. 2. us fix finite atlas ${\cal A}$ of the $M$; a function $u~: M\rightarrow {\mathbb {R}}$ is if for every chart $(U, \phi)$ belonging to ${\cal A}$, $u\circ\phi^{-1}$ is $K$-semi-concave. Then a $u$ at $q$ is | G_+(x)=E^u(x)\oplus R(x) $
Weak K.A.M. solutiOns and GreeN bundLes {#Sec3}
=======================================
in This SectIon, we recall the WEak KaM theory and give a relatiOnshiP bETweeN SoMe tanGent conES tO THe pSeUdOgrApHS oF the wEak kAM soluTions and thE GrEeN bundles. ThesE ReSults imply TheOrem \[th1\]. The proOfs Are givEn In sECtion \[Sec4\].
weak KaM theoRY
---------------
We don’T give any pRoOF in thiS Section, BUT aLl thE results that we givE ArE Proved in [@Fa1] or [@Be1].
if $t>0$, the FuNCtION $A_t~: m\tiMes M\rightaRrOw {\matHBb {R}}$ is deFInED BY: $$A_t(Q_0, Q_1)=\inf_\gamma \int_0^tl(\gamma (s), \dot\gAMma (S))ds=\min_\GaMma \INt_0^tL(\gaMma (s), \dOt\GAmmA (s))ds$$ where thE infImum is takEn on thE Set of $C^2$ cURves $\gamMa~: [0, t]\rigHtaRroW M$ suCH tHaT $\gaMmA (0)=Q_0$ anD $\GaMma (T)=Q_1$.
1. A fUnction $v~: v\rIgHtarrOw {\maTHBB {r}}$ defIneD on a SubseT $V$ of ${\mathbb {R}}^d$ iS [*$K$-sEmi-cONcaVe*]{} if fOr eveRy $x\iN V$, There Exists A lineAr Form $p_x$ defined on ${\MathBb {R}}^d$ so thaT: $$\foRaLl y\In v, v(y)\leQ V(x)+p_x(y-x)+k\| y-x\|^2.$$ theN we say tHat $p_x$ is A [*k-suPeR-DIFfErential*]{} of $v$ at $x$.
2. Let uS fIX A fInite atlAs ${\cal A}$ OF tHe MAnifold $M$; A fUncTion $U~: m\RightArroW {\MaThbb {R}}$ is [*$K$-Semi-coNCaVe*]{} If for evErY chart $(u, \pHi)$ bEloNging TO ${\cal a}$, $u\circ\Phi^{-1}$ is $K$-seMi-conCAve. Then a [*$K$-super-DIfferential*]{} of $U$ At $Q$ IS $p_X\Circ | G_+(x)=E^u(x)\oplus R(x) $
Weak K. A.M.sol uti on s an d Gr een bundles {# s ec3}
===================== ===== == = ==== = == ===
In this se c t ion ,we re ca l lthe w eak KAM th eory and g ive a relationshi p b etween som e t angent cones to the p se udo g raphs of theweak K A M solu tions and t h e Gree n bundle s . T hese results imply th e or e m \[th1\]. The proof sa re g ive n i n section\[ sec4\ ] .
Weak KA M t heo r y
----------- ----
We do n ’tgive a ny pr o of inthisse c tio n, but alltheresults t hat we give ar e proved in [@ Fa1 ] o r [@ B e1 ].
I f$ t>0 $ ,the fun ction $A _t ~: M\ti mesM \ r i ghta rro w {\ mathb b {R}}$ is de fin ed b y : $ $A_t( q_0,q_1) =\ inf_\ gamma\int_ 0^ tL(\gamma (s),\dot \gamma (s ))d s= \mi n_ \gamm a \int_ 0^t L(\ gamma ( s), \do t \ga mm a ( s) )ds$$ where the in fi m u mis taken on th e s et of $C^2$ c urv es $ \ g amma~ : [0 , t ]\righta rrow M $ s uc h that$\ gamma(0 )=q _0$ and$ \gam ma (t) =q_1$.
1. A function $v~:V \rightarrow { \ ma t h bb {R}} $ d efined on a sub s et $ V$ o f $ {\m a thbb{R}}^ d$ is [*$K$-semi-concave* ]{ } if f or ev ery $x\in V$, there exi s t s a linea r fo r m$ p_x$ defined o n ${\ mathbb {R} } ^d$ so t hat:$$\foral l y\in V, v (y)\leqv(x )+p _x( y-x ) + K\ | y-x\|^2.$$T h en w esay tha t $ p_x$ is a[*K -su per -d ifferenti al*]{} o f$v $at $x $.
2 . Let us f ixafin ite a t las ${ \calA}$of t h e m anifold $M $ ; a f un ct ion$u~ :M\rig htar r ow{\mathb b {R}}$ i s [ * $K$- se mi -concav e*]{} if forev ery chart$( U,\phi)$ b elonging to ${\cal A}$, $u\circ \ phi^{-1 }$is $K $-se mi-concav e.Then a [* $ K$-sup er-dif feren ti al* ] { } of$ u $at$q $ is $p_x\ c i rc | G_+(x)=E^u(x)\oplus R(x)_ $
Weak_K.A.M. solutions and Green_bundles {#sec3}
=======================================
In_this_section, we_recall_the weak KAM_theory and give_a relationship between some_tangent cones to_the_pseudographs of the weak KAM solutions and the Green bundles. These results imply theorem_\[th1\]._The proofs_are_given_in section \[sec4\].
Weak KAM theory
---------------
We_don’t give any proof in_this section,_but all the results that we give are_proved_in [@Fa1] or_[@Be1].
If $t>0$, the function $A_t~: M\times M\rightarrow {\mathbb {R}}$_is defined by: $$A_t(q_0, q_1)=\inf_\gamma \int_0^tL(\gamma_(s), \dot\gamma (s))ds=\min_\gamma_\int_0^tL(\gamma_(s),_\dot\gamma (s))ds$$ where the_infimum is taken on the set_of $C^2$ curves $\gamma~: [0, t]\rightarrow_M$ such that $\gamma (0)=q_0$ and $\gamma_(t)=q_1$.
1. A function $v~: V\rightarrow_{\mathbb {R}}$ defined on a_subset $V$_of ${\mathbb {R}}^d$ is [*$K$-semi-concave*]{}_if for every_$x\in V$,_there exists a_linear form $p_x$ defined on ${\mathbb_{R}}^d$ so that:_$$\forall y\in V, v(y)\leq v(x)+p_x(y-x)+K\| y-x\|^2.$$_Then_we say that_$p_x$_is_a [*K-super-differential*]{}_of $v$ at_$x$.
2._ Let_us_fix a finite atlas ${\cal A}$_of_the manifold $M$; a function $u~: M\rightarrow_{\mathbb {R}}$ is [*$K$-semi-concave*]{}_if_for every chart $(U,_\phi)$ belonging to ${\cal A}$,_$u\circ\phi^{-1}$ is $K$-semi-concave. Then a [*$K$-super-differential*]{}_of $u$_at $q$_is $p_x\circ |
,0) – (.2,0); (0,-4) \[fill = black\] circle (.07); at (0,1.4) [$u_1$]{}; at (0,-.7) [$u_2$]{}; at (1.6,0) [$\Pi$]{}; at (0,-5) [Fig. 2.2]{}; at (5,-3) [$\partial \Pi = u_1u_2$]{}; at (0,-2.2) [$\Delta_2$]{};
The inequality $|{\Delta}(2)| \le | {\partial}{\Delta}|$ and its stronger version $\sum_{\Pi \in {\Delta}(2)} |{\partial}\Pi | \le | {\partial}{\Delta}|$ are immediate from property (A).
\(b) Suppose that a disk diagram ${\Delta}$ over [pr1]{} does not have property (A).
First assume that ${\Delta}$ contains a face $\Pi$ whose boundary path ${\partial}\Pi$ is not simple. Then, as in the proof of part (a), there is a factorization ${\partial}\Pi = u_1 u_2$, where $u_1, u_2$ are closed subpaths of ${\partial}\Pi$ and $0 < |u_1|, |u_2| < |{\partial}\Pi|$. Renaming $u_1$ and $u_2$ if necessary, we may assume that $u_2$ bounds a disk subdiagram ${\Delta}_2$ of ${\Delta}$ such that ${\Delta}_2$ does not contain $\Pi$, see Fig. 2.2. If $u_2$ is not simple, then we can replace $u_2$ by its closed subpath $u'_2$ such that $0 < |u'_2| < |u_2|$ and $u'_2$ bounds a disk subdiagram ${\Delta}'_2$ that contains no $\Pi$. Hence, choosing a shortest path $u_2$ as above, we may assume that $u_2$ is simple. By Lemma \[vk\] applied to ${\Delta}_2$ with ${\partial}{\Delta}_2 = u_2$, we have $$\label{ud2}
{\varphi}(u_2) \equiv {\varphi}({\partial}{\Delta}_2) \overset{{\mathcal{G}}_2} = 1.$$
Denote ${\varphi}({\partial}\Pi | , 0) – (.2,0); (0,-4) \[fill = black\ ] circle (.07); at (0,1.4) [ $ u_1 $ ] { }; at (0,-.7) [ $ u_2 $ ] { }; at (1.6,0) [ $ \Pi$ ] { }; at (0,-5) [ Fig. 2.2 ] { }; at (5,-3) [ $ \partial \Pi = u_1u_2 $ ] { }; at (0,-2.2) [ $ \Delta_2 $ ] { };
The inequality $ |{\Delta}(2)| \le | { \partial}{\Delta}|$ and its stronger version $ \sum_{\Pi \in { \Delta}(2) } |{\partial}\Pi | \le | { \partial}{\Delta}|$ are immediate from property (deoxyadenosine monophosphate).
\(b) presuppose that a disk diagram $ { \Delta}$ over [ pr1 ] { } does not have property (vitamin a).
First assume that $ { \Delta}$ contain a font $ \Pi$ whose boundary path $ { \partial}\Pi$ is not elementary. Then, as in the proof of part (a), there be a factorization $ { \partial}\Pi = u_1 u_2 $, where $ u_1, u_2 $ are closed subpaths of $ { \partial}\Pi$ and $ 0 < |u_1|, |u_2| < |{\partial}\Pi|$. Renaming $ u_1 $ and $ u_2 $ if necessary, we may simulate that $ u_2 $ restrict a disk subdiagram $ { \Delta}_2 $ of $ { \Delta}$ such that $ { \Delta}_2 $ does not contain $ \Pi$, see Fig. 2.2. If $ u_2 $ is not simple, then we can replace $ u_2 $ by its shut subpath $ u'_2 $ such that $ 0 < |u'_2| < |u_2|$ and $ u'_2 $ bounds a disk subdiagram $ { \Delta}'_2 $ that contains no $ \Pi$. therefore, choosing a shortest path $ u_2 $ as above, we may assume that $ u_2 $ is elementary. By Lemma \[vk\ ] applied to $ { \Delta}_2 $ with $ { \partial}{\Delta}_2 = u_2 $, we have $ $ \label{ud2 }
{ \varphi}(u_2) \equiv { \varphi}({\partial}{\Delta}_2) \overset{{\mathcal{G}}_2 } = 1.$$
Denote $ { \varphi}({\partial}\Pi | ,0) – (.2,0); (0,-4) \[fill = black\] circle (.07); at (0,1.4) [$u_1$]{}; at (0,-.7) [$u_2$]{}; at (1.6,0) [$\Pi$]{}; at (0,-5) [Fij. 2.2]{}; at (5,-3) [$\lartial \Oi = u_1u_2$]{}; at (0,-2.2) [$\Delta_2$]{};
The inequalivy $|{\Dwlta}(2)| \oe | {\partial}{\Delta}|$ and kts stronher versuon $\wym_{\Pi \in {\Delta}(2)} |{\partlcl}\Pi | \pe | {\'actial}{\Delta}|$ are lmmediate fsom property (A).
\(t) Ru'pose that a disk diagram ${\Delta}$ over [pr1]{} doex jot have propetty (A).
Gyrst assume that ${\Delta}$ contains a face $\Pi$ whove boundary psth ${\partial}\Pi$ is not simplf. Thfn, as in the proof of part (a), jgerq is a factorkzation ${\pagcial}\Pi = u_1 u_2$, where $u_1, u_2$ are closed subpaths uf ${\paxtial}\Pi$ and $0 < |u_1|, |k_2| < |{\partial}\Pi|$. Renaiing $u_1$ and $u_2$ if necevsary, wr may assume tmat $u_2$ boynds a disk subdiagrak ${\Delta}_2$ of ${\Delta}$ soch that ${\Dalca}_2$ does not contain $\Pu$, wee Fhg. 2.2. Ix $u_2$ kw nut aikpme, thej wx can replade $u_2$ by its closed subpath $u'_2$ sicr that $0 < |u'_2| < |u_2|$ znd $u'_2$ fotnds a disk subdiagram ${\Delta}'_2$ that contapns ho $\Pi$. Hence, choosing a whortest path $u_2$ as ablve, we mar assume that $u_2$ is simple. By Lemma \[vk\] applied to ${\Dalta}_2$ xigh ${\kavbial}{\Awlha}_2 = u_2$, we have $$\label{ud2}
{\varphi}(u_2) \equiv {\varphi}({\pwdtoak}{\Delta}_2) \overseb{{\mathcal{G}}_2} = 1.$$
Denote ${\fagpny}({\partial}\Pi | ,0) – (.2,0); (0,-4) \[fill = black\] at [$u_1$]{}; at [$u_2$]{}; at (1.6,0) at [$\partial \Pi = at (0,-2.2) [$\Delta_2$]{}; inequality $|{\Delta}(2)| \le | {\partial}{\Delta}|$ and stronger version $\sum_{\Pi \in {\Delta}(2)} |{\partial}\Pi | \le | {\partial}{\Delta}|$ are immediate from (A). \(b) Suppose that a disk diagram ${\Delta}$ over [pr1]{} does not have (A). assume ${\Delta}$ a face $\Pi$ whose boundary path ${\partial}\Pi$ is not simple. Then, as in the proof of (a), there is a factorization ${\partial}\Pi = u_1 where $u_1, u_2$ are subpaths of ${\partial}\Pi$ and $0 |u_1|, < |{\partial}\Pi|$. $u_1$ $u_2$ necessary, we may that $u_2$ bounds a disk subdiagram ${\Delta}_2$ of ${\Delta}$ such that ${\Delta}_2$ does not contain $\Pi$, see 2.2. If not simple, we replace by its closed such that $0 < |u'_2| < bounds a disk subdiagram ${\Delta}'_2$ that contains no Hence, choosing shortest path $u_2$ as above, we assume that $u_2$ is simple. By Lemma \[vk\] to ${\Delta}_2$ with ${\partial}{\Delta}_2 = u_2$, we have $$\label{ud2} {\varphi}(u_2) \equiv {\varphi}({\partial}{\Delta}_2) \overset{{\mathcal{G}}_2} = 1.$$ | ,0) – (.2,0); (0,-4) \[fill = black\] circle (.07); at (0,1.4) [$u_1$]{}; at (0,-.7) [$u_2$]{}; at (1.6,0) [$\Pi$]{}; At (0,-5) [Fig. 2.2]{}; at (5,-3) [$\parTial \PI = u_1u_2$]{}; At (0,-2.2) [$\DElTa_2$]{};
ThE ineQuality $|{\Delta}(2)| \le | {\PArtiAl}{\Delta}|$ and its stronger vErsioN $\sUM_{\Pi \iN {\deLta}(2)} |{\paRtial}\Pi | \LE | {\pARTiaL}{\DElTa}|$ aRe IMmEdiatE frOm propeRty (A).
\(b) SuppoSe tHaT a disk diagraM ${\deLta}$ over [pr1]{} dOes Not have propeRty (a).
First AsSumE That ${\DEltA}$ contAins a fACe $\Pi$ whOse boundaRy PAth ${\parTIal}\Pi$ is NOT sImplE. Then, as in the proof OF pARt (a), there is a facTorizaTiON ${\pARTiaL}\Pi = U_1 u_2$, where $u_1, u_2$ aRe CloseD SubpathS Of ${\PARTiaL}\pi$ and $0 < |u_1|, |u_2| < |{\partiaL}\Pi|$. Renaming $U_1$ And $U_2$ if necEsSarY, We may aSsume ThAT $u_2$ bOunds a disk sUbdiAgram ${\DeltA}_2$ of ${\DelTA}$ such thAT ${\Delta}_2$ dOes not ConTaiN $\Pi$, sEE FIg. 2.2. if $u_2$ Is NOt sIMpLe, tHEn wE can replAcE $u_2$ By its ClosED SUBpatH $u'_2$ sUch tHat $0 < |u'_2| < |u_2|$ And $u'_2$ bounds a diSk sUbdiAGraM ${\DeltA}'_2$ that ContAiNs no $\PI$. Hence, ChoosInG a shortest path $u_2$ As abOve, we may aSsuMe ThaT $u_2$ Is simPLe. By LeMma \[Vk\] aPplied tO ${\Delta}_2$ wITh ${\pArTIAL}{\DElta}_2 = u_2$, we have $$\label{ud2}
{\VaRPHi}(U_2) \equiv {\vaRphi}({\paRTiAl}{\dElta}_2) \overSeT{{\maThcaL{g}}_2} = 1.$$
denotE ${\varPHi}({\Partial}\PI | ,0) – (.2,0); (0,-4) \[fil l = black\ ] cir cle (. 07 ); a t (0 ,1.4) [$u_1$]{ } ; at (0,-.7) [$u_2$]{}; at (1.6 ,0 ) [$\ P i$ ]{};at (0,- 5 )[ F ig. 2 .2 ]{} ;a t(5,-3 ) [ $\parti al \Pi = u _1u _2 $]{}; at (0, - 2. 2) [$\Delt a_2 $]{};
The i neq uality $ |{\ D elta} (2) | \le | {\ p artial }{\Delta} |$ and it s strong e r v ersi on $\sum_{\Pi \in {\ D elta}(2)} |{\p artial }\ P i| \le |{\partial} {\ Delta } |$ arei mm e d i ate from property (A).
\(b) Sup pose t ha t a disk d iagra m$ {\D elta}$ over [pr 1]{} does not h a ve prop e rty (A) .
Fir stass umet ha t${\ De l ta} $ c ont a ins a face$\ Pi $ who se b o u n d arypat h ${ \part ial}\Pi$ is n otsimp l e.Then, as i n th eproof of pa rt (a ), there is a fac tori zation ${ \pa rt ial }\ Pi =u _1 u_2 $,whe re $u_1 , u_2$a recl o s e dsubpaths of ${\par ti a l }\ Pi$ and$0 < | u _1 |, |u_2| <|{ \pa rtia l } \Pi|$ . Re n am ing $u_1 $ and$ u_ 2$ if nec es sary,we ma y a ssume that $u_2$ boundsa dis k subdiagram ${ \ Delta}_2$ of$ {\ D e lt a }$ s uch that ${\De lta} _ 2$ d oesn ot co n tain$\Pi$ ,s ee Fig. 2.2. If $u_2$is not s imple , then we can replace $ u _ 2 $ by its clo s ed subpath $u'_2$ such that $0 < |u'_2| < |u_2 |$ and $ u'_2$ bou n d s a disk su bdi agr am$ { \D elta}'_2$ tha t cont ai ns no $ \Pi $. Henc e,cho osi ngashortestpath $u_ 2$ a sab ove , wem ay assum etha t$u_ 2$ is simple . ByLemm a\[ v k\] applie d t o ${\D el ta }_2$ wi th ${\p arti a l}{ \Delta} _2 = u_2$ , w e hav e$$ \label{ ud2}
{\var ph i}(u_2) \e qu iv{\varp h i }({\part ial}{\Delta}_2) \overs e t{{\mat hca l{G}} _2}= 1.$$
D eno te ${\ var p hi}({\ partia l}\Pi | ,0) –_(.2,0); (0,-4)_\[fill = black\] circle_(.07); at_(0,1.4)_[$u_1$]{}; at_(0,-.7)_[$u_2$]{}; at (1.6,0)_[$\Pi$]{}; at (0,-5)_[Fig. 2.2]{}; at (5,-3)_[$\partial \Pi =_u_1u_2$]{};_at (0,-2.2) [$\Delta_2$]{};
The inequality $|{\Delta}(2)| \le | {\partial}{\Delta}|$ and its stronger version $\sum_{\Pi_\in_{\Delta}(2)} |{\partial}\Pi_|_\le_| {\partial}{\Delta}|$ are immediate from_property (A).
\(b) Suppose that a_disk diagram_${\Delta}$ over [pr1]{} does not have property (A).
First_assume_that ${\Delta}$ contains_a face $\Pi$ whose boundary path ${\partial}\Pi$ is not_simple. Then, as in the proof_of part (a),_there_is_a factorization ${\partial}\Pi =_u_1 u_2$, where $u_1, u_2$ are_closed subpaths of ${\partial}\Pi$ and $0_< |u_1|, |u_2| < |{\partial}\Pi|$. Renaming $u_1$_and $u_2$ if necessary, we may_assume that $u_2$ bounds a_disk subdiagram_${\Delta}_2$ of ${\Delta}$ such that_${\Delta}_2$ does not_contain $\Pi$,_see Fig. 2.2. If_$u_2$ is not simple, then we_can replace $u_2$_by its closed subpath $u'_2$ such_that_$0 < |u'_2|_<_|u_2|$_and $u'_2$_bounds a disk_subdiagram_${\Delta}'_2$ that_contains_no $\Pi$. Hence, choosing a shortest_path_$u_2$ as above, we may assume that_$u_2$ is simple. By_Lemma \[vk\]_applied to ${\Delta}_2$ with_${\partial}{\Delta}_2 = u_2$, we have_$$\label{ud2}
{\varphi}(u_2) \equiv {\varphi}({\partial}{\Delta}_2)_ \overset{{\mathcal{G}}_2}_= 1.$$
Denote_${\varphi}({\partial}\Pi |
sigma x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\
&\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.
\end{split}$$ The total consumer surplus is then: $$\label{eq:csmonoproof}
\textnormal{CS}^m\sum_{i=1}^m\sum_{j=1}^m\theta_{ij}\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.$$
Proof of Proposition \[prop:unique\] {#app:unique}
------------------------------------
In order to prove that the firms are in an equilibrium, we first follow similar steps as the proof of Proposition \[prop:monomarginalprices\] and determine the optimal prices. Suppose that ${\ell_{ij}^1},\;{\ell_{ij}^2}\in\left[\frac{1-\sigma}{2}{\ell_{\max}},(1-\sigma){\ell_{\max}}\right]$. In that region: $$\label{eq:demandduo}
D({\ell_{ij}^1},{\ell_{ij}^2})=\frac{4\sigma(1-\sigma-\frac{{\ell_{ij}^1}-{\ell_{ij}^2}}{{\ell_{\max}}})-(\frac{{\ell_{ij}^1}+{\ell_{ij}^2}}{{\ell_{\max}}}+\sigma-1)^2}{8\sigma(1-\sigma)}$$ $$\label{eq:demandderivduo}
\frac{\partial D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial{\ell_{ij}^1}}=\frac{1}{{\ell_{\max}}}\frac{-6\sigma-\frac{2{\ell_{ij}^1}}{{\ell_{\max}}}-\frac{2{\ell_{ij}^2}}{{\ell_{\max}}}+2}{8\sigma(1-\sigma)}$$ $$\label{eq:demandsecderivduo}
\frac{\partial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial({\ell_{ij}^1})^2}=\frac{1}{{\ell_{\max}}^2}\frac{-2}{8\sigma(1-\sigma | sigma x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\
& \frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma }.
\end{split}$$ The total consumer surplus is then: $ $ \label{eq: csmonoproof }
\textnormal{CS}^m\sum_{i=1}^m\sum_{j=1}^m\theta_{ij}\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.$$
Proof of Proposition \[prop: unique\ ] { # app: unique }
------------------------------------
In order to rise that the firm are in an equilibrium, we first adopt like steps as the proof of Proposition \[prop: monomarginalprices\ ] and specify the optimal prices. think that $ { \ell_{ij}^1},\;{\ell_{ij}^2}\in\left[\frac{1-\sigma}{2}{\ell_{\max}},(1-\sigma){\ell_{\max}}\right]$. In that region: $ $ \label{eq: demandduo }
D({\ell_{ij}^1},{\ell_{ij}^2})=\frac{4\sigma(1-\sigma-\frac{{\ell_{ij}^1}-{\ell_{ij}^2}}{{\ell_{\max}}})-(\frac{{\ell_{ij}^1}+{\ell_{ij}^2}}{{\ell_{\max}}}+\sigma-1)^2}{8\sigma(1-\sigma)}$$ $ $ \label{eq: demandderivduo }
\frac{\partial D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial{\ell_{ij}^1}}=\frac{1}{{\ell_{\max}}}\frac{-6\sigma-\frac{2{\ell_{ij}^1}}{{\ell_{\max}}}-\frac{2{\ell_{ij}^2}}{{\ell_{\max}}}+2}{8\sigma(1-\sigma)}$$ $ $ \label{eq: demandsecderivduo }
\frac{\partial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial({\ell_{ij}^1})^2}=\frac{1}{{\ell_{\max}}^2}\frac{-2}{8\sigma(1-\sigma | sigla x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\
&\frac{{\ell_{\max}}(\sntma^2+\sigka+1)-3{\ell_{in}^m}(1+\sigma-\ffac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.
\end{s'lit}$$ The ujtal consumer surpuus is thvn: $$\label{ew:csmiboproof}
\textnormal{CS}^m\amm_{i=1}^m\sbm_{o=1}^m\theta_{ij}\frac{{\elk_{\max}}(\sigma^2+\shgma+1)-3{\ell_{ij}^m}(1+\sigmd-\ffae{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.$$
Proof of Propositijn \[prop:umieue\] {#app:unique}
------------------------------------
Ig orcqr tk prove that the firms are in an esuilibrpum, we first follpw similar steps as the prlof lf Proposition \[proo:monomarginqlpryxes\] and detefmine the optimal pricgs. Suppose that ${\ell_{ij}^1},\;{\ell_{ij}^2}\in\left[\wrac{1-\sngma}{2}{\ell_{\max}},(1-\sugna){\epn_{\max}}\right]$. Ii that region: $$\label{eq:demangduo}
D({\ell_{ij}^1},{\ell_{ij}^2})=\frsc{4\smgma(1-\wigma-\frac{{\ell_{ij}^1}-{\ell_{ij}^2}}{{\eln_{\max}}})-(\frac{{\ell_{ij}^1}+{\ell_{ij}^2}}{{\gll_{\max}}}+\sigmd-1)^2}{8\sngma(1-\sigma)}$$ $$\label{eq:demabdeerivguo}
\ffqc{\pxrtjak S({\ell_{ij}^1},{\fll_{mj}^2})}{\partial{\elm_{ij}^1}}=\frac{1}{{\ell_{\mqx}}}\frac{-6\sigma-\frac{2{\ell_{ik}^1}}{{\ejo_{\max}}}-\frac{2{\ell_{ij}^2}}{{\sll_{\max}}}+2}{8\fidma(1-\sigma)}$$ $$\label{eq:demandsecderivduo}
\frdc{\pzrtial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\partiao({\ell_{ij}^1})^2}=\frac{1}{{\ell_{\max}}^2}\frac{-2}{8\slgma(1-\sigma | sigma x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\ &\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}. \end{split}$$ The total consumer then: \textnormal{CS}^m\sum_{i=1}^m\sum_{j=1}^m\theta_{ij}\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.$$ Proof Proposition \[prop:unique\] {#app:unique} that firms are in equilibrium, we first similar steps as the proof of \[prop:monomarginalprices\] and determine the optimal prices. Suppose that ${\ell_{ij}^1},\;{\ell_{ij}^2}\in\left[\frac{1-\sigma}{2}{\ell_{\max}},(1-\sigma){\ell_{\max}}\right]$. In that region: $$\label{eq:demandduo} $$\label{eq:demandderivduo} \frac{\partial D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial{\ell_{ij}^1}}=\frac{1}{{\ell_{\max}}}\frac{-6\sigma-\frac{2{\ell_{ij}^1}}{{\ell_{\max}}}-\frac{2{\ell_{ij}^2}}{{\ell_{\max}}}+2}{8\sigma(1-\sigma)}$$ $$\label{eq:demandsecderivduo} \frac{\partial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial({\ell_{ij}^1})^2}=\frac{1}{{\ell_{\max}}^2}\frac{-2}{8\sigma(1-\sigma | sigma x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\
&\fraC{{\ell_{\max}}(\sigMa^2+\sigMa+1)-3{\eLl_{iJ}^m}(1+\SigmA-\fraC{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sIGma}.
\eNd{split}$$ The total consumeR surpLuS Is thEN: $$\lAbel{eQ:csmonoPRoOF}
\TexTnOrMal{cS}^M\SuM_{i=1}^m\suM_{j=1}^m\Theta_{ij}\Frac{{\ell_{\max}}(\SigMa^2+\Sigma+1)-3{\ell_{ij}^m}(1+\sIGmA-\frac{{\ell_{ij}^M}}{{\elL_{\max}}})}{6\sigma}.$$
ProOf oF PropoSiTioN \[Prop:uNiqUe\] {#app:Unique}
------------------------------------
iN order To prove thAt THe firmS Are in an EQUiLibrIum, we first follow sIMiLAr steps as the prOof of PRoPOsITIon \[ProP:monomargiNaLpricES\] and detERmINE The OPtimal prices. SUppose that ${\eLL_{ij}^1},\;{\Ell_{ij}^2}\iN\lEft[\FRac{1-\sigMa}{2}{\ell_{\MaX}},(1-\SigMa){\ell_{\max}}\rigHt]$. In That regioN: $$\label{EQ:demandDUo}
D({\ell_{iJ}^1},{\ell_{ij}^2})=\FraC{4\siGma(1-\sIGmA-\fRac{{\ElL_{Ij}^1}-{\eLL_{iJ}^2}}{{\elL_{\Max}}})-(\Frac{{\ell_{iJ}^1}+{\eLl_{Ij}^2}}{{\ell_{\Max}}}+\sIGMA-1)^2}{8\SigmA(1-\siGma)}$$ $$\lAbel{eQ:demandderivdUo}
\fRac{\pARtiAl D({\elL_{ij}^1},{\elL_{ij}^2})}{\pArTial{\eLl_{ij}^1}}=\frAc{1}{{\ell_{\MaX}}}\frac{-6\sigma-\frac{2{\eLl_{ij}^1}}{{\Ell_{\max}}}-\fraC{2{\elL_{iJ}^2}}{{\elL_{\mAx}}}+2}{8\sigMA(1-\sigma)}$$ $$\LabEl{eQ:demandSecderiVDuo}
\FrAC{\PArTial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\parTiAL({\ElL_{ij}^1})^2}=\frac{1}{{\eLl_{\max}}^2}\fRAc{-2}{8\SiGMa(1-\sigma | sigma x+(1-\sigma)y-{\ell_ {ij}^m})dx dy=\\
&\f rac{ {\ell_{\max}}( \ sigm a^2+\sigma+1)-3{\ell_{ ij}^m }( 1 +\si g ma -\fra c{{\ell _ {i j } ^m} }{ {\ ell _{ \ ma x}}}) }{6 \sigma} .
\end {sp li t}$$ The tot a lconsumer s urp lus is then: $$ \label {e q:c s monop roo f}
\tex t normal {CS}^m\su m_ { i=1}^m \ sum_{j= 1 } ^m \the ta_{ij}\frac{{\el l _{ \ max}}(\sigma^2 +\sigm a+ 1 )- 3 { \el l_{ ij}^m}(1+\ si gma-\ f rac{{\e l l_ { i j }^m } }{{\ell_{\max }}})}{6\sig m a}. $$
Pr oo f o f Propo sitio n\ [pr op:unique\] {#a pp:unique }
---- - ------- - ------- ------ --- --- ----
I nord er top ro vet hat the fir ms a re in ane q u i libr ium , we firs t follow simi lar ste p s a s the proo f of P ropos ition\[pro p: monomarginalpri ces\ ] and det erm in e t he opti m al pri ces . S upposethat ${ \ ell _{ i j } ^1 },\;{\ell_{ij}^2}\ in \ l ef t[\frac{ 1-\sig m a} {2 } {\ell_{\ ma x}} ,(1- \ s igma) {\el l _{ \max}}\r ight]$ . I nthat re gi on: $$ \l abe l{e q:dem a nddu o}
D({\ell _{ij} ^ 1},{\ell_{ij}^ 2 })=\frac{4\si g ma ( 1 -\ s igma -\f rac{{\ell_{ ij}^ 1 }-{\ ell_ { ij }^2 } }{{\e ll_{\ ma x }} } )-(\frac{{\ell_{ij} ^1 }+{\el l_{ij }^2}}{{\ell_{ \max}}}+\s i g m a-1)^2}{ 8\si g ma ( 1-\sigma)}$$ $ $\lab el{eq:dema n dderivdu o}
\frac{ \partialD ( {\ell_{i j}^ 1}, {\e ll_ { i j} ^2})}{\partia l { \ell _{ ij}^1}} =\f rac{1}{ {\e ll_ {\m ax} }} \frac{-6\ sigma-\f ra c{ 2{ \e ll_ {ij}^ 1 }}{{\ell _{ \ma x} }}- \frac { 2{\ell _{ij} ^2}} {{ \e l l_{ \max}}} + 2} { 8 \sig ma (1 -\si gma )} $$ $$ \lab e l{e q:deman dsecderiv duo }
\ fr ac{\par tial^2D({\ell _{ ij}^1},{\e ll _{i j}^2}) } { \partial ({\ell_{ij}^1})^2}=\fra c {1}{{\e ll_ {\max }}^2 }\frac{-2 }{8 \sigma (1- \ sigma | sigma x+(1-\sigma)y-{\ell_{ij}^m})dxdy=\\
_ _ _ &\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.
__ _\end{split}$$_The total consumer_surplus is then:_$$\label{eq:csmonoproof}
_\textnormal{CS}^m\sum_{i=1}^m\sum_{j=1}^m\theta_{ij}\frac{{\ell_{\max}}(\sigma^2+\sigma+1)-3{\ell_{ij}^m}(1+\sigma-\frac{{\ell_{ij}^m}}{{\ell_{\max}}})}{6\sigma}.$$
Proof of Proposition \[prop:unique\]_{#app:unique}
------------------------------------
In_order to prove that the firms are in an equilibrium, we first follow similar_steps_as the_proof_of_Proposition \[prop:monomarginalprices\] and determine the_optimal prices. Suppose that ${\ell_{ij}^1},\;{\ell_{ij}^2}\in\left[\frac{1-\sigma}{2}{\ell_{\max}},(1-\sigma){\ell_{\max}}\right]$._In that_region: $$\label{eq:demandduo}
D({\ell_{ij}^1},{\ell_{ij}^2})=\frac{4\sigma(1-\sigma-\frac{{\ell_{ij}^1}-{\ell_{ij}^2}}{{\ell_{\max}}})-(\frac{{\ell_{ij}^1}+{\ell_{ij}^2}}{{\ell_{\max}}}+\sigma-1)^2}{8\sigma(1-\sigma)}$$ $$\label{eq:demandderivduo}
__ \frac{\partial D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial{\ell_{ij}^1}}=\frac{1}{{\ell_{\max}}}\frac{-6\sigma-\frac{2{\ell_{ij}^1}}{{\ell_{\max}}}-\frac{2{\ell_{ij}^2}}{{\ell_{\max}}}+2}{8\sigma(1-\sigma)}$$_$$\label{eq:demandsecderivduo}
\frac{\partial^2D({\ell_{ij}^1},{\ell_{ij}^2})}{\partial({\ell_{ij}^1})^2}=\frac{1}{{\ell_{\max}}^2}\frac{-2}{8\sigma(1-\sigma |
rho_{N+1} = 0$. This corresponds to the linear system $$\label{e:upmat}
\begin{pmatrix}
-1 & 2 & 0 &\cdots & 0 \\
2 & -1 & 2 & \cdots &0 \\
\vdots & \vdots & \ddots &\ddots & \vdots\\
0 & 0 & \cdots & 2 & -1
\end{pmatrix} \boldsymbol{\rho} = \omega \boldsymbol{1}$$ Though the matrix is tri-diagonal, it is not diagonally dominant, so its solvability is not immediately clear. However,
The matrix in has no kernel for any $N$.
Letting $A_N$ denote this matrix, proving it has no kernel is equivalent to showing $\det A_N \neq 0$ for any $N$. Indeed, $$\label{e:matrecursion}
\begin{split}
\det A_{N} &= -1 \cdot \det A_{N-1} -2 \cdot
\begin{vmatrix}
2 & 2 &0 & \cdots & 0\\
0 & -1 &2 & \cdots & 0\\
0 & 2 & -1 &\cdots & 0\\
\vdots & \ddots & \ddots & \ddots&\vdots\\
0 & 0 & \cdots & 2 & -1
\end{vmatrix}\\
& = - \det A_{N-1} - 4 \det A_{N-2},
\end{split}$$ giving us a recursion relation for the determinant. By inspection, $$\det A_1= -1, \quad \det A_2 = -3.$$ By induction, recursion relation demands that for all $N$, the determinant of $A_N$ must be odd. Hence, it is never zero.
A consequence of this is the following Corollary.
Any nontrivial phase locked solution has $\omega \neq 0$. Moreover, given any real valued function $\omega(t) \neq 0$ for all $t \in
\mathbb{R}$, there exists a nontrivial phase locked solution $\{
(\phi_j, \rho_j) \}_{j=1}^N$ to.
For a given $N$, the linear system can always be solved, and | rho_{N+1 } = 0$. This corresponds to the linear system $ $ \label{e: upmat }
\begin{pmatrix }
-1 & 2 & 0 & \cdots & 0 \\
2 & -1 & 2 & \cdots & 0 \\
\vdots & \vdots & \ddots & \ddots & \vdots\\
0 & 0 & \cdots & 2 & -1
\end{pmatrix } \boldsymbol{\rho } = \omega \boldsymbol{1}$$ Though the matrix is tri - diagonal, it is not diagonally prevailing, so its solvability is not immediately clean. However,
The matrix in has no kernel for any $ N$.
Letting $ A_N$ announce this matrix, proving it have no kernel is equivalent to showing $ \det A_N \neq 0 $ for any $ N$. Indeed, $ $ \label{e: matrecursion }
\begin{split }
\det A_{N } & = -1 \cdot \det A_{N-1 } -2 \cdot
\begin{vmatrix }
2 & 2 & 0 & \cdots & 0\\
0 & -1 & 2 & \cdots & 0\\
0 & 2 & -1 & \cdots & 0\\
\vdots & \ddots & \ddots & \ddots&\vdots\\
0 & 0 & \cdots & 2 & -1
\end{vmatrix}\\
& = - \det A_{N-1 } - 4 \det A_{N-2 },
\end{split}$$ give us a recursion relation for the determinant. By inspection, $ $ \det A_1= -1, \quad \det A_2 = -3.$$ By initiation, recursion relation demands that for all $ N$, the antigenic determinant of $ A_N$ must be odd. Hence, it is never zero.
A consequence of this is the following Corollary.
Any nontrivial phase locked solution has $ \omega \neq 0$. furthermore, given any real respect function $ \omega(t) \neq 0 $ for all $ t \in
\mathbb{R}$, there exists a nontrivial phase locked solution $ \ {
(\phi_j, \rho_j) \}_{j=1}^N$ to.
For a contribute $ N$, the linear system can always be clear, and | rho_{J+1} = 0$. This corresponds to uhe linear system $$\label{e:npmat}
\gegin{pmagrix}
-1 & 2 & 0 &\cdots & 0 \\
2 & -1 & 2 & \cdits &0 \\
\vdots & \vdots & \ddots &\ddlts & \vdors\\
0 & 0 & \cdots & 2 & -1
\end{pmatrjw} \bolbsbmbol{\rho} = \omega \boldsymbon{1}$$ Though the mdtfir is tri-diagonal, it is not diagonallr dominsnh, so its solvafiliuy ys nkn lmmediately clear. However,
The matdix in ias no kernel fpr any $N$.
Letting $A_N$ denote hhis matrix, proving it has no kerbel yw equivalent to showing $\det A_N \neq 0$ for any $N$. Indeed, $$\label{e:matrecjrsiou}
\begin{skljt}
\det A_{N} &= -1 \cdjt \det A_{N-1} -2 \cdot
\begin{vkatrix}
2 & 2 &0 & \xdots & 0\\
0 & -1 &2 & \cvots & 0\\
0 & 2 & -1 &\cqots & 0\\
\vdots & \ddots & \edits & \gdotv&\vdogw\\
0 & 0 & \cdohs & 2 & -1
\ehd{vmatrix}\\
& = - \det A_{N-1} - 4 \det A_{G-2},
\end{split}$$ fiving uf a recursion relation for the determindnt. By inspection, $$\det A_1= -1, \qyad \det A_2 = -3.$$ By inductlon, recurfion relation demands that for all $N$, the determindnt oh $X_N$ nufg bf odd. Hence, it is never zero.
A consequence of egix ps the following Gorollary.
Any nontrovlak phase locked solutnkn has $\omega \neq 0$. Mogeover, diven any real valied function $\omega(t) \neq 0$ foe all $t \in
\iqthbb{R}$, there existd a nontrivnal phsse lpcked solution $\{
(\phi_j, \rko_j) \}_{j=1}^N$ to.
For a gigen $N$, the uinear system cav akwdys be solved, and | rho_{N+1} = 0$. This corresponds to the $$\label{e:upmat} -1 & & 0 &\cdots -1 2 & \cdots \\ \vdots & & \ddots &\ddots & \vdots\\ 0 0 & \cdots & 2 & -1 \end{pmatrix} \boldsymbol{\rho} = \omega \boldsymbol{1}$$ Though matrix is tri-diagonal, it is not diagonally dominant, so its solvability is not clear. The in no kernel for any $N$. Letting $A_N$ denote this matrix, proving it has no kernel is to showing $\det A_N \neq 0$ for any Indeed, $$\label{e:matrecursion} \begin{split} \det &= -1 \cdot \det A_{N-1} \cdot 2 & &0 \cdots 0\\ 0 & &2 & \cdots & 0\\ 0 & 2 & -1 &\cdots & 0\\ \vdots & \ddots & & \ddots&\vdots\\ 0 & & & \end{vmatrix}\\ & = A_{N-1} - 4 \det A_{N-2}, \end{split}$$ recursion relation for the determinant. By inspection, $$\det -1, \quad A_2 = -3.$$ By induction, recursion demands that for all $N$, the determinant of must be odd. Hence, it is never zero. A consequence of this is the following nontrivial phase locked solution $\omega \neq 0$. given real function \neq 0$ all $t \in \mathbb{R}$, there exists a nontrivial phase locked solution (\phi_j, \rho_j) \}_{j=1}^N$ to. For a given $N$, the linear always solved, and | rho_{N+1} = 0$. This corresponds to the lInear systeM $$\labeL{e:uPmaT}
\bEgin{PmatRix}
-1 & 2 & 0 &\cdots & 0 \\
2 & -1 & 2 & \cdots &0 \\
\vDOts & \vDots & \ddots &\ddots & \vdots\\
0 & 0 & \cdoTs & 2 & -1
\end{PmATrix} \BOlDsymbOl{\rho} = \omEGa \BOLdsYmBoL{1}$$ ThOuGH tHe matRix Is tri-diAgonal, it is Not DiAgonally domiNAnT, so its solvAbiLity is not immEdiAtely cLeAr. HOWever,
the MatriX in has NO kerneL for any $N$.
LEtTIng $A_N$ dENote thiS MAtRix, pRoving it has no kernEL iS Equivalent to shOwing $\dEt a_n \nEQ 0$ For Any $n$. Indeed, $$\labEl{E:matrECursion}
\BEgIN{SPliT}
\Det A_{N} &= -1 \cdot \det A_{n-1} -2 \cdot
\begin{vMAtrIx}
2 & 2 &0 & \cdotS & 0\\
0 & -1 &2 & \cDotS & 0\\
0 & 2 & -1 &\Cdots & 0\\
\vDots & \dDoTS & \ddOts & \ddots&\vdoTs\\
0 & 0 & \cdOts & 2 & -1
\end{vmaTrix}\\
& = - \deT a_{N-1} - 4 \det A_{N-2},
\ENd{split}$$ Giving Us a RecUrsiON rElAtiOn FOr tHE dEteRMinAnt. By insPeCtIon, $$\deT A_1= -1, \quAD \DET A_2 = -3.$$ By IndUctiOn, recUrsion relatioN deMandS ThaT for aLl $N$, thE detErMinanT of $A_N$ mUst be OdD. Hence, it is never Zero.
a consequeNce Of ThiS iS the fOLlowinG CoRolLary.
Any NontrivIAl pHaSE LOcKed solution has $\omegA \nEQ 0$. moReover, giVen any REaL vALued funcTiOn $\oMega(T) \NEq 0$ for All $t \IN
\mAthbb{R}$, thEre exiSTs A nOntriviAl Phase lOcKed SolUtion $\{
(\PHi_j, \rHo_j) \}_{j=1}^N$ tO.
For a givEn $N$, thE Linear system caN Always be solveD, AnD | rho_{N+1} = 0$. This corre sponds tothe l ine arsy stem $$\ label{e:upmat} \b egin{pmatrix}
-1 & 2&0 &\c d ot s & 0 \\
2& -1&2& \ cd o ts &0 \ \
\vdo ts & \vdot s & \ ddots &\ddot s & \vdots\\ 0 & 0 & \cd ots & 2 & - 1
\end{ pma trix} \bold s ymbol{ \rho} = \ om e ga \bo l dsymbol { 1 }$ $ Th ough the matrix i s t r i-diagonal, it is no td ia g o nal lydominant,so itss olvabil i ty i s no t immediatelyclear. Howe v er,
Thema tri x in ha s noke r nel for any $N $.
Letting $ A_N$ d e note th i s matri x, pro vin g i t ha s n oker ne l is eq uiv a len t to sho wi ng $\de t A_ N \ n eq 0 $ f or a ny $N $. Indeed, $$ \la bel{ e :ma trecu rsion }
\begi n{spli t}
\det A_{N} &= -1\cdot \de t A _{ N-1 }-2 \c d ot
\be gin{vma trix}
2 & 2 &0 & \cdots & 0 \\ 0 & -1 &2 &\c d ots & 0\ \ 0 & 2& -1 &\ cdots &0\\
\vdots & \ddot s& \ ddo ts &\ ddot s&\vdo ts\\
0 & 0 & \cdots& 2 & -1
\e n d {v m atri x}\ \
& = - \ d et A _{N- 1 }- 4 \detA_{N- 2} ,
\end{split}$$ giv in g us a recu rsion relatio n for thed e t erminant . By in s pection, $$\de t A_1 = -1, \qua d \det A_ 2 = - 3.$$ Byinduction , recursio n r ela tio n d e m an ds that for a l l $N$ ,the det erm inant o f $ A_N $ m ust b e odd. He nce, itis n ev er ze ro.
A consequ en ceof th is is the fo llowi ng C or ol l ary .
Anyn on t r ivia lph aseloc ke d sol utio n ha s $\ome ga \neq 0 $.M oreo ve r, givenany real valu ed function$\ ome ga(t)\ n eq 0$ fo r all $t \in
\mathbb{ R }$, the reexist s anontrivia l p hase l ock e d solu tion $ \{
(\ phi _ j , \rh o _ j) \} _{ j=1}^N$ to .
Fo r a g iv en $ N$, the linear system can alw ays be solved , a nd | rho_{N+1} =_0$. This_corresponds to the linear_system $$\label{e:upmat}
__\begin{pmatrix} _
_ _-1 & 2_& 0 &\cdots &_0 \\
__ 2 & -1 & 2 & \cdots &0 \\
\vdots_&_\vdots &_\ddots_&\ddots_& \vdots\\
_0 & 0 & \cdots_& 2_& -1
\end{pmatrix} \boldsymbol{\rho} = \omega \boldsymbol{1}$$_Though_the matrix is_tri-diagonal, it is not diagonally dominant, so its solvability_is not immediately clear. However,
The matrix_in has no_kernel_for_any $N$.
Letting $A_N$ denote_this matrix, proving it has no_kernel is equivalent to showing $\det_A_N \neq 0$ for any $N$. Indeed,_$$\label{e:matrecursion}
\begin{split}
_ \det_A_{N} &=_-1 \cdot \det A_{N-1} -2_\cdot
_ _ \begin{vmatrix}
_ _2 &_2 &0 & \cdots & 0\\
__ ___ 0_& -1 &2_&_\cdots &_0\\
_ __0 & 2 & -1 &\cdots &_0\\
__ \vdots_& \ddots & \ddots &_\ddots&\vdots\\
_ _0 &_0 & \cdots & 2 & -1
_ \end{vmatrix}\\
_ & = -_\det A_{N-1}_-_4 \det A_{N-2},
__ _\end{split}$$ giving us a recursion relation for_the determinant._By inspection, $$\det A_1= -1, \quad_\det A_2 = -3.$$_By_induction, recursion relation demands that for_all $N$, the determinant of $A_N$_must be odd. Hence, it_is_never_zero.
A consequence of this is_the following Corollary.
Any nontrivial phase locked_solution has $\omega_\neq 0$. Moreover, given any real valued_function_$\omega(t) \neq 0$ for all $t_\in
_ \mathbb{R}$, there exists a nontrivial_phase_locked_solution $\{
(\phi_j, \rho_j)_\}_{j=1}^N$ to.
For a given $N$, the_linear system can always be solved, and |
) –(-2.8,1.6) node \[left\] [ ]{}; (-1.5,1.7) –(-2.8,1.7); (-1.8,1.8) –(-2.8,1.8) node \[left\] [ ]{};
--------- ----------------------- -- -----------------
$\circ$ Elements in $S_{i,j}$ Elements in $S$
--------- ----------------------- -- -----------------
We are now ready to prove Conjecture \[conjD\].
We have $g(T)=g(\cX)$ and in particular $H(P)=T.$
Proposition \[prop:contained\] implies that $g(T) \ge g(\cX)$. Hence the theorem follows once we show that $g(T) \le g(\cX)$. However, using the first two items of Lemma \[lem:sij\], we see that $$g(T) \le g(S)-\sum_{i=0}^{q-1}\sum_{j=1}^{q-1}|S_{i,j}|.$$ Using Lemma \[genusS\] and item three of Lemma \[lem:sij\] we obtain $$\begin{aligned}
g(T) &\leq \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \sum_{j=1}^{q-1} q(q^3-q+1-jq^2+i)\\
&= \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \sum_{j=1}^{q-1} q(q^3-q+1) +\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}jq^3-\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}iq\\
&= \frac{q^6-q^4}{2}-q^2(q-1)(q^3-q+1)+\frac{q^5(q-1)}{2}-\frac{q^2(q-1)^2}{2}=\frac{q^5-2q^3+q^2}{2}=g(\cX).\end{aligned}$$
A direct consequence of the above theorem is that $H | ) – (-2.8,1.6) node \[left\ ] [ ] { }; (-1.5,1.7) – (-2.8,1.7); (-1.8,1.8) – (-2.8,1.8) node \[left\ ] [ ] { };
--------- ----------------------- -- -----------------
$ \circ$ Elements in $ S_{i, j}$ Elements in $ S$
--------- ----------------------- -- -----------------
We are now ready to rise Conjecture \[conjD\ ].
We take $ g(T)=g(\cX)$ and in particular $ H(P)=T.$
Proposition \[prop: contained\ ] implies that $ g(T) \ge g(\cX)$. Hence the theorem postdate once we show that $ g(T) \le g(\cX)$. However, using the beginning two detail of Lemma \[lem: sij\ ], we see that $ $ g(T) \le g(S)-\sum_{i=0}^{q-1}\sum_{j=1}^{q-1}|S_{i, j}|.$$ use Lemma \[genusS\ ] and item three of Lemma \[lem: sij\ ] we obtain $ $ \begin{aligned }
g(T) & \leq \frac{q^6 - q^4}{2 } -\sum_{i=0}^{q-1 } \sum_{j=1}^{q-1 } q(q^3 - q+1 - jq^2+i)\\
& = \frac{q^6 - q^4}{2 } -\sum_{i=0}^{q-1 } \sum_{j=1}^{q-1 } q(q^3 - q+1) + \sum_{i=0}^{q-1 } \sum_{j=1}^{q-1}jq^3-\sum_{i=0}^{q-1 } \sum_{j=1}^{q-1}iq\\
& = \frac{q^6 - q^4}{2}-q^2(q-1)(q^3 - q+1)+\frac{q^5(q-1)}{2}-\frac{q^2(q-1)^2}{2}=\frac{q^5 - 2q^3+q^2}{2}=g(\cX).\end{aligned}$$
A lineal consequence of the above theorem is that $ H | ) –(-2.8,1.6) nlde \[left\] [ ]{}; (-1.5,1.7) –(-2.8,1.7); (-1.8,1.8) –(-2.8,1.8) node \[left\] [ ]{};
--------- ----------------------- -- -----------------
$\circ$ Elemenvs in $S_{j,j}$ Dlements in $S$
--------- ----------------------- -- -----------------
We are now rwady uj prove Conjecture \[conjD\].
We jave $g(T)=g(\xX)$ aid in particular $H(P)=T.$
Propositioh \[pro':cintained\] impligs that $g(T) \ga g(\cX)$. Hence tha ghzorem follows once we show that $g(T) \lq g(\cX)$. Hpwfver, using the firxe twk items of Lemma \[lem:sij\], we see thaf $$g(T) \le g(S)-\sum_{i=0}^{q-1}\sum_{j=1}^{q-1}|X_{i,j}|.$$ Using Lemma \[genusS\] and itel three of Lemma \[lfm:sij\] we objzin $$\vegin{aligned}
e(T) &\leq \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \som_{j=1}^{q-1} q(q^3-q+1-jq^2+i)\\
&= \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \sum_{j=1}^{d-1} q(q^3-q+1) +\sum_{i=0}^{q-1} \sum_{h=1}^{q-1}hq^3-\skk_{i=0}^{q-1} \sum_{j=1}^{q-1}iq\\
&= \frwc{q^6-q^4}{2}-q^2(q-1)(q^3-q+1)+\frac{q^5(q-1)}{2}-\frac{q^2(q-1)^2}{2}=\xrac{q^5-2q^3+q^2}{2}=b(\cX).\end{aligned}$$
A dicect consequence of the auove theorem is that $H | ) –(-2.8,1.6) node \[left\] [ ]{}; (-1.5,1.7) –(-2.8,1.8) \[left\] [ --------- ----------------------- -- Elements $S$ --------- ----------------------- ----------------- We are ready to prove Conjecture \[conjD\]. We $g(T)=g(\cX)$ and in particular $H(P)=T.$ Proposition \[prop:contained\] implies that $g(T) \ge g(\cX)$. Hence theorem follows once we show that $g(T) \le g(\cX)$. However, using the first items Lemma we that $$g(T) \le g(S)-\sum_{i=0}^{q-1}\sum_{j=1}^{q-1}|S_{i,j}|.$$ Using Lemma \[genusS\] and item three of Lemma \[lem:sij\] we obtain $$\begin{aligned} &\leq \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \sum_{j=1}^{q-1} q(q^3-q+1-jq^2+i)\\ &= \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} q(q^3-q+1) +\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}jq^3-\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}iq\\ \frac{q^6-q^4}{2}-q^2(q-1)(q^3-q+1)+\frac{q^5(q-1)}{2}-\frac{q^2(q-1)^2}{2}=\frac{q^5-2q^3+q^2}{2}=g(\cX).\end{aligned}$$ A direct consequence of above is that | ) –(-2.8,1.6) node \[left\] [ ]{}; (-1.5,1.7) –(-2.8,1.7); (-1.8,1.8) –(-2.8,1.8) node \[left\] [ ]{};
--------- ----------------------- -- -----------------
$\circ$ ElemeNts in $S_{i,j}$ ElEmentS in $s$
--------- ----------------------- -- -----------------
We ArE now ReadY to prove ConjecTUre \[cOnjD\].
We have $g(T)=g(\cX)$ and in paRticuLaR $h(P)=T.$
PROpOsitiOn \[prop:cONtAINed\] ImPlIes ThAT $g(t) \ge g(\cx)$. HeNce the tHeorem follOws OnCe we show that $G(t) \lE g(\cX)$. HoweveR, usIng the first tWo iTems of leMma \[LEm:sij\], We sEe thaT $$g(T) \le g(s)-\Sum_{i=0}^{q-1}\sUm_{j=1}^{q-1}|S_{i,j}|.$$ UsInG lemma \[gENusS\] and ITEm ThreE of Lemma \[lem:sij\] we oBTaIN $$\begin{aligned}
g(t) &\leq \frAc{Q^6-Q^4}{2} -\sUM_{I=0}^{q-1} \sUm_{j=1}^{Q-1} q(q^3-q+1-jq^2+i)\\
&= \fraC{q^6-Q^4}{2} -\sum_{i=0}^{Q-1} \Sum_{j=1}^{q-1} q(q^3-Q+1) +\SuM_{I=0}^{Q-1} \Sum_{J=1}^{Q-1}jq^3-\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}Iq\\
&= \frac{q^6-q^4}{2}-q^2(q-1)(q^3-Q+1)+\FraC{q^5(q-1)}{2}-\fraC{q^2(Q-1)^2}{2}=\frAC{q^5-2q^3+q^2}{2}=g(\cx).\end{aLiGNed}$$
a direct consEqueNce of the aBove thEOrem is tHAt $H | ) –(-2.8,1.6) node \[left\ ] [ ]{}; ( -1.5, 1.7 ) – (- 2.8, 1.7) ; (-1.8,1.8) – ( -2.8 ,1.8) node \[left\] []{};
---- - -- -- -- ------- - -- - - --- -- -- ---- -- ----- --- -------
$\cir c$El ements in $S _ {i ,j}$ Ele ments in $S$
------ -- - - - ----- --- ----- ------ - - -- - --------- -- - ----
W e are n o w r eady to prove Conject u re \[conjD\].
We have$g ( T) = g (\c X)$ and in pa rt icula r $H(P)= T .$ P rop o sition \[prop :contained\ ] im pliesth at$ g(T) \ ge g( \c X )$. Hence thetheo rem follo ws onc e we sho w that $ g(T) \ leg(\ cX)$ . H ow eve r, usi n gthe fir st two i te ms of L emma \ [ l em:s ij\ ], w e see that $$g(T)\le g(S ) -\s um_{i =0}^{ q-1} \s um_{j =1}^{q -1}|S _{ i,j}|.$$ UsingLemm a \[genus S\] a ndit em th r ee ofLem ma\[lem:s ij\] we obt ai n $ $\ begin{aligned}
g(T )& \ le q \frac {q^6-q ^ 4} {2 } -\sum_{ i= 0}^ {q-1 } \sum_ {j=1 } ^{ q-1} q(q ^3-q+1 - jq ^2 +i)\\
& = \fr ac {q^ 6-q ^4}{2 } -\s um_{i= 0}^{q-1} \sum _ {j=1}^{q-1} q( q ^3-q+1) +\sum _ {i = 0 }^ { q-1} \s um_{j=1}^{q -1}j q ^3-\ sum_ { i= 0}^ { q-1}\sum_ {j = 1} ^ {q-1}iq\\
&= \fra c{ q^6-q^ 4}{2} -q^2(q-1)(q^3 -q+1)+\fra c { q ^5(q-1)} {2}- \ fr a c{q^2(q-1)^2}{ 2}=\f rac{q^5-2q ^ 3+q^2}{2 }=g(\ cX).\end {aligned} $ $
A dire ctcon seq uen c e o f the above t h e orem i s that$H | ) –(-2.8,1.6)_node \[left\]_[ ]{}; (-1.5,1.7) –(-2.8,1.7);_(-1.8,1.8) –(-2.8,1.8)_node_\[left\] [_]{};
_ --------- -----------------------_-- -----------------
_ $\circ$ Elements_in $S_{i,j}$ __ Elements in $S$
--------- ----------------------- -- -----------------
We are now ready_to_prove Conjecture_\[conjD\].
We_have_$g(T)=g(\cX)$ and in particular $H(P)=T.$
Proposition_\[prop:contained\] implies that $g(T) \ge_g(\cX)$. Hence_the theorem follows once we show that $g(T)_\le_g(\cX)$. However, using_the first two items of Lemma \[lem:sij\], we see_that $$g(T) \le g(S)-\sum_{i=0}^{q-1}\sum_{j=1}^{q-1}|S_{i,j}|.$$ Using Lemma_\[genusS\] and item_three_of_Lemma \[lem:sij\] we obtain_$$\begin{aligned}
g(T) &\leq \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}_q(q^3-q+1-jq^2+i)\\
&= \frac{q^6-q^4}{2} -\sum_{i=0}^{q-1}_\sum_{j=1}^{q-1} q(q^3-q+1) +\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}jq^3-\sum_{i=0}^{q-1} \sum_{j=1}^{q-1}iq\\
&= _\frac{q^6-q^4}{2}-q^2(q-1)(q^3-q+1)+\frac{q^5(q-1)}{2}-\frac{q^2(q-1)^2}{2}=\frac{q^5-2q^3+q^2}{2}=g(\cX).\end{aligned}$$
A direct consequence of the above_theorem is that $H |
omega}^{1}\,.\end{aligned}$$ The largest corrections to the above approach occur for the PNC pion-exchange interaction which, due to its long range, produces some extra energy dependence and sizable $P\leftrightarrow D$ transition amplitudes. They can show up when the contribution of the $S\leftrightarrow P$ transition amplitude is suppressed, like in this work.
For $\omega_{\gamma}=2.235$ MeV, which is very close to the disintegration threshold, we get the asymmetry $$\begin{aligned}
A_{\gamma}^{(th)} & \approx & \left[-8.44\, h_{\rho}^{0}-17.65\, h_{\rho}^{2}+3.63\, h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\label{eq:A-threshold}\end{aligned}$$ Using the DDH “best” values as an estimate, we got $A_{\gamma}^{(th)}\approx2.53\times10^{-8}$. By detailed balancing, one expects that $A_{\gamma}^{(th)}$ equals the circular polarization $P_{\gamma}^{(th)}$ observed in the radiative thermal neutron capture by proton, given the same kinematics. Though our result does not exactly correspond to the same kinematics as the inverse process usually considered (the kinetic energy of thermal neutrons $\sim$0.025 eV), it agrees both in sign and order of magnitude with existing calculations of $P_{\gamma}^{(th)}$ [@Lassey:1975; @Desplanques:1975; @Craver:1976am]. We also performed a similar calculation for the latter case with A$v_{18}$, and the result is $$\begin{aligned}
P_{\gamma}^{(th)} & \approx & \left[-8.75\, h_{\rho}^{0}-17.47\, h_{\rho}^{2}+3.39\, h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\end{aligned}$$ This is very close to the result of $A_{\gamma}$ quoted above.
It is noticed that our expression of $A_{\gamma}^{(th)}$ at very low energy, and therefore that one for $P_{\gamma}^{(th)}$, contains a contribution from the | omega}^{1}\,.\end{aligned}$$ The largest corrections to the above approach occur for the PNC pion - substitution interaction which, due to its retentive range, produces some extra department of energy dependence and sizable $ P\leftrightarrow D$ transition amplitude. They can show up when the contribution of the $ S\leftrightarrow P$ transition amplitude is suppress, like in this work.
For $ \omega_{\gamma}=2.235 $ MeV, which is very close to the annihilation threshold, we get the asymmetry $ $ \begin{aligned }
A_{\gamma}^{(th) } & \approx & \left[-8.44\, h_{\rho}^{0}-17.65\, h_{\rho}^{2}+3.63\, h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\label{eq: A - threshold}\end{aligned}$$ use the DDH “ good ” values as an appraisal, we got $ A_{\gamma}^{(th)}\approx2.53\times10^{-8}$. By detailed reconciliation, one expect that $ A_{\gamma}^{(th)}$ equals the circular polarization $ P_{\gamma}^{(th)}$ observed in the radiative thermal neutron capture by proton, yield the same kinematics. Though our result does not precisely correspond to the like kinematics as the inverse process normally considered (the kinetic energy of thermal neutrons $ \sim$0.025 eV), it agree both in sign and order of magnitude with existing calculations of $ P_{\gamma}^{(th)}$ [ @Lassey:1975; @Desplanques:1975; @Craver:1976am ]. We besides perform a similar calculation for the latter case with A$ v_{18}$, and the result is $ $ \begin{aligned }
P_{\gamma}^{(th) } & \approx & \left[-8.75\, h_{\rho}^{0}-17.47\, h_{\rho}^{2}+3.39\, h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\end{aligned}$$ This is very close to the result of $ A_{\gamma}$ quoted above.
It is noticed that our expression of $ A_{\gamma}^{(th)}$ at very low department of energy, and consequently that one for $ P_{\gamma}^{(th)}$, contains a contribution from the | omeha}^{1}\,.\end{aligned}$$ The largest corrections to the abmve aplroach ozcur for the PNC pion-exchangx inreracupon which, due to its uong rangv, producew sonw extra enxdgy depcudencs and wizable $P\leftrlghtarrow D$ transition am[lktbdes. They can show up when the contrybution ov the $S\leftrigrtargor P$ fgaksition amplitude is suppressed, mike in this work.
For $\omega_{\gamma}=2.235$ MeV, which is vfry flose to the disinhegration tyresrild, we get tfe asymmetgv $$\begin{alighed}
A_{\gamma}^{(th)} & \approx & \left[-8.44\, h_{\rho}^{0}-17.65\, f_{\rho}^{2}+3.63\, k_{\omega}^{0}\right.\bobumhgr \\
& & \left.\hwpace{2bm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\label{aq:A-threxhold}\end{alignec}$$ Uvint the DDH “best” values as an estimate, we got $A_{\gamka}^{(ch)}\approx2.53\times10^{-8}$. By detaulwd bananchng, ube dxptctx fhat $A_{\hamja}^{(th)}$ equala the circuoar polarization $P_{\gsmiq}^{(th)}$ observed jn the rwdiative thermal neutron capture by prouon, gjven the same kinematicw. Though our result dles not evactly correspond to the same kinematics as the itversx oroecfr udually considered (the kinetic energy of thermwm meltrons $\sim$0.025 eV), it cgrees both in sogj sgd order of mxgnitubs sith existing calcklationf of $P_{\gamma}^{(th)}$ [@Jassry:1975; @Desplanques:1975; @Craver:1976am]. We qlso performvd a similar calculatiln for the uattgr casr with A$v_{18}$, and the resulc is $$\bsgin{aligned}
O_{\gamma}^{(th)} & \xpprox & \left[-8.75\, h_{\rhu}^{0}-17.47\, h_{\gho}^{2}+3.39\, v_{\omega}^{0}\right.\nonumber \\
& & \lest.\hspace{2cn}+O(c_{1},c_{3},e_{6})\right]\tioes10^{-3}\,.\emd{aligged}$$ This id vern close to the resupt of $A_{\camma}$ quotfd above.
It is noticed that our xepression of $S_{\gdmmd}^{(th)}$ at vzry loe energy, and eherefore that one fox $P_{\gamoa}^{(th)}$, contapns a convribution frjm the | omega}^{1}\,.\end{aligned}$$ The largest corrections to the above for PNC pion-exchange which, due to extra dependence and sizable D$ transition amplitudes. can show up when the contribution the $S\leftrightarrow P$ transition amplitude is suppressed, like in this work. For $\omega_{\gamma}=2.235$ which is very close to the disintegration threshold, we get the asymmetry $$\begin{aligned} & & h_{\rho}^{0}-17.65\, h_{\omega}^{0}\right.\nonumber \\ & & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\label{eq:A-threshold}\end{aligned}$$ Using the DDH “best” values as an estimate, we got $A_{\gamma}^{(th)}\approx2.53\times10^{-8}$. detailed balancing, one expects that $A_{\gamma}^{(th)}$ equals the polarization $P_{\gamma}^{(th)}$ observed in radiative thermal neutron capture by given same kinematics. our does exactly correspond to same kinematics as the inverse process usually considered (the kinetic energy of thermal neutrons $\sim$0.025 eV), it both in order of with calculations $P_{\gamma}^{(th)}$ [@Lassey:1975; @Desplanques:1975; also performed a similar calculation for with A$v_{18}$, and the result is $$\begin{aligned} P_{\gamma}^{(th)} \approx & h_{\rho}^{0}-17.47\, h_{\rho}^{2}+3.39\, h_{\omega}^{0}\right.\nonumber \\ & & This is very close to the result of quoted above. It is noticed that our expression of $A_{\gamma}^{(th)}$ at very low energy, and one for $P_{\gamma}^{(th)}$, contains contribution from the | omega}^{1}\,.\end{aligned}$$ The largest CorrectionS to thE abOve ApProaCh ocCur for the PNC piON-excHange interaction which, dUe to iTs LOng rANgE, prodUces somE ExTRA enErGy DepEnDEnCe and SizAble $P\leFtrightarrOw D$ TrAnsition amplITuDes. They can ShoW up when the coNtrIbutioN oF thE $s\leftRigHtarrOw P$ traNSition Amplitude Is SUppresSEd, like iN THiS worK.
For $\omega_{\gamma}=2.235$ MeV, WHiCH is very close to The disInTEgRATioN thReshold, we gEt The asYMmetry $$\bEGiN{ALIgnED}
A_{\gamma}^{(th)} & \apprOx & \left[-8.44\, h_{\rho}^{0}-17.65\, h_{\RHo}^{2}+3.63\, h_{\Omega}^{0}\rIgHt.\nONumber \\
& & \Left.\hSpACe{2cM}+O(c_{1},c_{3},c_{6})\right]\tImes10^{-3}\,.\Label{eq:A-tHreshoLD}\end{aliGNed}$$ UsinG the DDh “beSt” vAlueS As An EstImATe, wE GoT $A_{\gAMma}^{(Th)}\approx2.53\TiMeS10^{-8}$. By deTailED BALancIng, One eXpectS that $A_{\gamma}^{(th)}$ EquAls tHE ciRculaR polaRizaTiOn $P_{\gaMma}^{(th)}$ oBservEd In the radiative tHermAl neutron CapTuRe bY pRoton, GIven thE saMe kInematiCs. ThougH Our ReSULT dOes not exactly correSpOND tO the same KinemaTIcS aS The inverSe ProCess USUally ConsIDeRed (the kiNetic eNErGy Of thermAl NeutroNs $\Sim$0.025 EV), iT agreES botH in sigN and ordeR of maGNitude with exisTIng calculatioNS oF $p_{\GaMMa}^{(th)}$ [@lasSey:1975; @DesplanqUes:1975; @CRAver:1976Am]. We ALsO peRFormeD a simIlAR cALculation for the lattEr Case wiTh A$v_{18}$, aNd the result is $$\Begin{alignED}
p_{\Gamma}^{(th)} & \aPproX & \LeFT[-8.75\, h_{\rho}^{0}-17.47\, h_{\rho}^{2}+3.39\, h_{\omegA}^{0}\righT.\nonumber \\
& & \lEFt.\hspace{2Cm}+O(c_{1},c_{3},C_{6})\right]\tiMes10^{-3}\,.\end{aliGNEd}$$ This is VerY clOse To tHE ReSult of $A_{\gamma}$ qUOTed aBoVe.
It is nOtiCed that Our ExpResSioN oF $A_{\gamma}^{(th)}$ At very loW eNeRgY, aNd tHerefORe that onE fOr $P_{\GaMma}^{(Th)}$, conTAins a cOntriButiOn FrOM thE | omega}^{1}\,.\end{aligned} $$ The lar gestcor rec ti onsto t he above appro a ch o ccur for the PNC pion- excha ng e int e ra ction which, du e toit slon gr an ge, p rod uces so me extra e ner gy dependencea nd sizable $ P\l eftrightarro w D $ tran si tio n ampl itu des.They c a n show up whenth e contr i butiono f t he $ S\leftrightarrowP $t ransition ampl itudeis su p p res sed , like inth is wo r k.
For $\ o m e ga_ { \gamma}=2.235 $ MeV, whic h is verycl ose to the disi nt e gra tion thresh old, we get t he asy m metry $ $ \begin{ aligne d}A_{ \gam m a} ^{ (th )} & \ a pp rox & \ left[-8. 44 \, h_{\ rho} ^ { 0 } -17. 65\ , h_ {\rho }^{2}+3.63\,h_{ \ome g a}^ {0}\r ight. \non um ber \ \
& & \le ft .\hspace{2cm}+O (c_{ 1},c_{3}, c_{ 6} )\r ig ht]\t i mes10^ {-3 }\, .\label {eq:A-t h res ho l d } \e nd{aligned}$$ Usin gt h eDDH “bes t” val u es a s an esti ma te, weg o t $A_ {\ga m ma }^{(th)} \appro x 2. 53 \times1 0^ {-8}$. B y d eta iledb alan cing,one expe cts t h at $A_{\gamma} ^ {(th)}$ equal s t h e c i rcul arpolarizatio n $P _ {\ga mma} ^ {( th) } $ obs erved i n t h e radiative thermal n eutron capt ure by proton , given th e s ame kine mati c s. Though our res ult d oes not ex a ctly cor respo nd to th e same ki n e matics a s t heinv ers e pr ocess usually c onsi de red (th e k ineticene rgy of th er mal neutr ons $\si m$ 0. 02 5eV) , ita grees bo th in s ign ando rder o f mag nitu de w i thexistin g c a l cula ti on s of $P _{ \gamm a}^{ ( th) }$ [@La ssey:1975 ; @ D espl an qu es:1975 ; @Craver:197 6a m]. We als oper formed a similar calculation for the la t ter cas e w ith A $v_{ 18}$, and th e resu lti s $$\b egin{a ligne d}
P_ { \ gamma } ^ {( th) }& \approx& \le ft[-8 .7 5\,h_{\rho }^{0}-17.47\, h_{\ r ho} ^{2}+3.39\, h _{\ omeg a } ^{ 0}\ r ig h t.\ no n umb e r \\
& & \left .\hspace{2 cm } +O (c_{1},c_{ 3 },c _{ 6})\rig ht]\tim es10^ { -3}\,.\ end{align ed}$$ Thi sis v e r y c lose to th e result of $A_{\ g amma} $ q uoted ab ove.
It is noti ced th a t o ur ex pressi on of $A _{\ga mm a}^{(th) }$ at very low energy,and th erefo rethat onefor $P_ {\gamma}^ {(th )}$, conta ins acontr ibu t ion f romt he | omega}^{1}\,.\end{aligned}$$ The_largest corrections_to the above approach_occur for_the_PNC pion-exchange_interaction_which, due to_its long range,_produces some extra energy_dependence and sizable_$P\leftrightarrow_D$ transition amplitudes. They can show up when the contribution of the $S\leftrightarrow P$_transition_amplitude is_suppressed,_like_in this work.
For $\omega_{\gamma}=2.235$ MeV,_which is very close to_the disintegration_threshold, we get the asymmetry $$\begin{aligned}
A_{\gamma}^{(th)} & \approx_&_\left[-8.44\, h_{\rho}^{0}-17.65\, h_{\rho}^{2}+3.63\,_h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\label{eq:A-threshold}\end{aligned}$$ Using the DDH_“best” values as an estimate, we_got $A_{\gamma}^{(th)}\approx2.53\times10^{-8}$. By_detailed_balancing,_one expects that $A_{\gamma}^{(th)}$_equals the circular polarization $P_{\gamma}^{(th)}$ observed_in the radiative thermal neutron capture_by proton, given the same kinematics. Though_our result does not exactly correspond_to the same kinematics as_the inverse_process usually considered (the kinetic_energy of thermal_neutrons $\sim$0.025_eV), it agrees_both in sign and order of_magnitude with existing_calculations of $P_{\gamma}^{(th)}$ [@Lassey:1975; @Desplanques:1975; @Craver:1976am]._We_also performed a_similar_calculation_for the_latter case with_A$v_{18}$,_and the_result_is $$\begin{aligned}
P_{\gamma}^{(th)} & \approx & \left[-8.75\,_h_{\rho}^{0}-17.47\,_h_{\rho}^{2}+3.39\, h_{\omega}^{0}\right.\nonumber \\
& & \left.\hspace{2cm}+O(c_{1},c_{3},c_{6})\right]\times10^{-3}\,.\end{aligned}$$_This is very close_to_the result of $A_{\gamma}$_quoted above.
It is noticed that_our expression of $A_{\gamma}^{(th)}$ at very_low energy,_and therefore_that one for $P_{\gamma}^{(th)}$, contains a contribution from the |
_i {{\rm d}}x^i {{\rm d}}\eta
+ (\delta_{i j} + 2 C_{i j}) {{\rm d}}x^i {{\rm d}}x^j \Bigr] \,,
\end{gathered}$$ where $C_{i j}$ can be further split into 2 scalar, 2 vector and 2 tensor degrees of freedom. By a direct comparison of these two parametrisations, the relation between the two metric parametrisations is
$$\begin{aligned}
g_{0 0} &: 1 + 2 \alpha + \alpha^2 - \beta_i \beta^i = 1 + 2 A\,,\\
g_{0 i} &: \beta_i + 2 h_{i j} \beta^j = B_i\,, \\
g_{i j} &: \delta_{i j} + 2 h_{i j} = \delta_{i j} + 2 C_{i j}\,.\end{aligned}$$
At first order we obtain $$\begin{gathered}
A^{(1)} = \alpha^{(1)} \,, \qquad B_i^{(1)} = \beta^{(1)}_i \,, \qquad
C_{i j}^{(1)} = h_{i j}^{(1)} \,,\end{gathered}$$ and the two parametrisations are the same. However, at second order we get the relations $$\begin{gathered}
A^{(2)}= \alpha^{(2)} + \alpha^{(1) 2} - \beta^{(1)}_i \beta^{(1) i} \,, \qquad
B_i^{(2)} = \beta^{(2)}_i + 4 h_{i j}^{(1)} \beta^{(1) j} \,,\qquad
C_{i j}^{(2)} = h_{i j}^{(2)} \,.\end{gathered}$$
Construction of the distribution function for polarised light {#app:Dist}
=============================================================
We consider a two-dimensional polarisation plane defined by two unit complex vectors $\hat {{\boldsymbol{\epsilon}}}_{({{\mathrm{I}}})}$ and $\hat {{\boldsymbol{\epsilon}}}_{({{\mathrm{II}}})}$, which are mutually orthogonal, $\hat \epsilon_{({{\mathrm{A}}})}{}^{\star\mu} \hat \epsilon_{({{\mathrm{B}}})}{}^\nu g_{\mu\nu}
=\delta_{({{\mathrm{A}}})({{\mathrm{B}}})}$. | _ i { { \rm d}}x^i { { \rm d}}\eta
+ (\delta_{i j } + 2 C_{i j }) { { \rm d}}x^i { { \rm d}}x^j \Bigr ] \, ,
\end{gathered}$$ where $ C_{i j}$ can be further split into 2 scalar, 2 vector and 2 tensor degrees of freedom. By a lineal comparison of these two parametrisations, the relation back between the two metric parametrisations is
$ $ \begin{aligned }
g_{0 0 } &: 1 + 2 \alpha + \alpha^2 - \beta_i \beta^i = 1 + 2 A\,,\\
g_{0 i } &: \beta_i + 2 h_{i j } \beta^j = B_i\, , \\
g_{i j } &: \delta_{i j } + 2 h_{i j } = \delta_{i j } + 2 C_{i j}\,.\end{aligned}$$
At inaugural orderliness we obtain $ $ \begin{gathered }
A^{(1) } = \alpha^{(1) } \, , \qquad B_i^{(1) } = \beta^{(1)}_i \, , \qquad
C_{i j}^{(1) } = h_{i j}^{(1) } \,,\end{gathered}$$ and the two parametrisations are the same. However, at second club we drive the relations $ $ \begin{gathered }
A^{(2)}= \alpha^{(2) } + \alpha^{(1) 2 } - \beta^{(1)}_i \beta^{(1) i } \, , \qquad
B_i^{(2) } = \beta^{(2)}_i + 4 h_{i j}^{(1) } \beta^{(1) j } \,,\qquad
C_{i j}^{(2) } = h_{i j}^{(2) } \,.\end{gathered}$$
structure of the distribution affair for polarised lighter { # app: Dist }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
We consider a two - dimensional polarisation airplane defined by two unit complex vectors $ \hat { { \boldsymbol{\epsilon}}}_{({{\mathrm{I}}})}$ and $ \hat { { \boldsymbol{\epsilon}}}_{({{\mathrm{II}}})}$, which are mutually orthogonal, $ \hat \epsilon_{({{\mathrm{A}}})}{}^{\star\mu } \hat \epsilon_{({{\mathrm{B}}})}{}^\nu g_{\mu\nu }
= \delta_{({{\mathrm{A}}})({{\mathrm{B}}})}$. | _i {{\rl d}}x^i {{\rm d}}\eta
+ (\delta_{i j} + 2 C_{i j}) {{\rm d}}x^i {{\rm d}}x^j \Bmgr] \,,
\ens{gatherea}$$ where $C_{i j}$ can be further dpoit ibto 2 scalar, 2 vector ana 2 tensor degrees of hreedom. By a dirxdt comparison lf tkewe two paramettisations, tha relation betfedn the two metric parametrisations is
$$\fegin{alogjed}
g_{0 0} &: 1 + 2 \alpra + \sjpha^2 - \beta_i \beta^i = 1 + 2 A\,,\\
g_{0 i} &: \beta_i + 2 h_{j j} \bete^j = B_i\,, \\
g_{i j} &: \drlta_{i j} + 2 h_{i j} = \delta_{i j} + 2 C_{i u}\,.\end{aligned}$$
At firsh order we ibtayb $$\begin{gathefed}
A^{(1)} = \alpha^{(1)} \,, \qquad B_i^{(1)} = \beta^{(1)}_i \,, \qquad
C_{i j}^{(1)} = h_{i j}^{(1)} \,,\end{gaghereb}$$ and the tqo pagdmetrisatiois are the same. However, at second order we get bhe rxlatuons $$\begin{gathered}
A^{(2)}= \alpha^{(2)} + \alpha^{(1) 2} - \beja^{(1)}_i \beta^{(1) i} \,, \qquad
B_i^{(2)} = \beta^{(2)}_i + 4 y_{i j}^{(1)} \teta^{(1) j} \,,\dwuaa
C_{j o}^{(2)} = g_{i j}^{(2)} \,.\ejd{gethered}$$
Consfruction of the distribution finsnoon for poladised jidht {#app:Dist}
=============================================================
We consider a two-dimensional pomarisation plane definee by two unit complex vectors $\rat {{\boldsymbol{\epsilon}}}_{({{\mathrm{I}}})}$ and $\hat {{\boldsymbol{\epvilon}}}_{({{\jxthxn{II}}})}$, wfucj are mutually orthogonal, $\hat \epsilon_{({{\mathrm{A}}})}{}^{\sezr\ku} \hat \epsilon_{({{\mabhrm{B}}})}{}^\nu g_{\mu\nu}
=\delts_{({{\mwtntm{A}}})({{\mathrm{B}}})}$. | _i {{\rm d}}x^i {{\rm d}}\eta + (\delta_{i 2 j}) {{\rm {{\rm d}}x^j \Bigr] can further split into scalar, 2 vector 2 tensor degrees of freedom. By direct comparison of these two parametrisations, the relation between the two metric parametrisations $$\begin{aligned} g_{0 0} &: 1 + 2 \alpha + \alpha^2 - \beta_i \beta^i 1 2 g_{0 &: \beta_i + 2 h_{i j} \beta^j = B_i\,, \\ g_{i j} &: \delta_{i j} + h_{i j} = \delta_{i j} + 2 C_{i At first order we $$\begin{gathered} A^{(1)} = \alpha^{(1)} \,, B_i^{(1)} \beta^{(1)}_i \,, C_{i = j}^{(1)} \,,\end{gathered}$$ and two parametrisations are the same. However, at second order we get the relations $$\begin{gathered} A^{(2)}= \alpha^{(2)} + 2} - i} \,, B_i^{(2)} \beta^{(2)}_i 4 h_{i j}^{(1)} \,,\qquad C_{i j}^{(2)} = h_{i j}^{(2)} the distribution function for polarised light {#app:Dist} ============================================================= consider a polarisation plane defined by two unit vectors $\hat {{\boldsymbol{\epsilon}}}_{({{\mathrm{I}}})}$ and $\hat {{\boldsymbol{\epsilon}}}_{({{\mathrm{II}}})}$, which are orthogonal, $\hat \epsilon_{({{\mathrm{A}}})}{}^{\star\mu} \hat \epsilon_{({{\mathrm{B}}})}{}^\nu g_{\mu\nu} =\delta_{({{\mathrm{A}}})({{\mathrm{B}}})}$. | _i {{\rm d}}x^i {{\rm d}}\eta
+ (\delta_{i j} + 2 C_{i j}) {{\rm d}}X^i {{\rm d}}x^j \BigR] \,,
\end{gAthEreD}$$ wHere $c_{i j}$ cAn be further splIT intO 2 scalar, 2 vector and 2 tensor DegreEs OF freEDoM. By a dIrect coMPaRISon Of ThEse TwO PaRametRisAtions, tHe relation BetWeEn the two metrIC pArametrisaTioNs is
$$\begin{aliGneD}
g_{0 0} &: 1 + 2 \alphA + \aLphA^2 - \Beta_i \BetA^i = 1 + 2 A\,,\\
g_{0 i} &: \Beta_i + 2 h_{I J} \beta^j = b_i\,, \\
g_{i j} &: \deltA_{i J} + 2 H_{i j} = \delTA_{i j} + 2 C_{i j}\,.\eND{AlIgneD}$$
At first order we obTAiN $$\Begin{gathered}
A^{(1)} = \Alpha^{(1)} \,, \qQuAD B_I^{(1)} = \BEta^{(1)}_I \,, \qqUad
C_{i j}^{(1)} = h_{i j}^{(1)} \,,\eNd{GatheREd}$$ and thE TwO PARamETrisations are The same. HoweVEr, aT seconD oRdeR We get tHe relAtIOns $$\Begin{gatherEd}
A^{(2)}= \aLpha^{(2)} + \alpha^{(1) 2} - \Beta^{(1)}_i \bETa^{(1) i} \,, \qquaD
b_i^{(2)} = \beta^{(2)}_i + 4 H_{i j}^{(1)} \betA^{(1) j} \,,\qQuaD
C_{i j}^{(2)} = H_{I j}^{(2)} \,.\EnD{gaThERed}$$
cOnStrUCtiOn of the dIsTrIbutiOn fuNCTIOn foR poLariSed liGht {#app:Dist}
=============================================================
We cOnsIder A Two-DimenSionaL polArIsatiOn planE defiNeD by two unit complEx veCtors $\hat {{\bOldSyMboL{\ePsiloN}}}_{({{\Mathrm{i}}})}$ anD $\haT {{\boldsyMbol{\epsILon}}}_{({{\MaTHRM{Ii}}})}$, which are mutually oRtHOGoNal, $\hat \epSilon_{({{\mAThRm{a}}})}{}^{\Star\mu} \haT \ePsiLon_{({{\mATHrm{B}}})}{}^\nU g_{\mu\NU}
=\dElta_{({{\mathRm{A}}})({{\matHRm{b}}})}$. | _i {{\rm d}}x^i {{\rm d}}\ eta
+ (\d elta_ {ij}+2 C_ {i j }) {{\rm d}}x^ i {{\ rm d}}x^j \Bigr] \,,
\end{ ga t here d }$ $ whe re $C_{ i j } $ ca nbe fu rt h er spli t i nto 2 s calar, 2 v ect or and 2 tenso r d egrees offre edom. By a d ire ct com pa ris o n ofthe se tw o para m etrisa tions, th er elatio n betwee n th e tw o metric parametr i sa t ions is
$$\be gin{al ig n ed } g_ {00} &: 1 +2\alph a + \alp h a^ 2 - \b e ta_i \beta^i= 1 + 2 A\, , \\ g_{0i} &: \beta_ i + 2 h _ {ij} \beta^j= B_ i\,, \\
g_{i j } &: \de l ta_{i j } + 2h_{ i j } =\ de lt a_{ ij } + 2C_{ i j} \,.\end{ al ig ned}$ $
A t f i rstord er w e obt ain $$\begin{ gat here d }
A^{(1 )} =\alp ha ^{(1) } \,,\qqua dB_i^{(1)} = \be ta^{ (1)}_i \, , \ qq uad
C_{ij }^{(1) } = h_ {i j}^{ (1)} \, , \en d{ g a t he red}$$ and the two p a r am etrisati ons ar e t he same. Ho we ver , at s econd ord e rwe get t he rel a ti on s $$\be gi n{gath er ed}
A ^{(2) } = \ alpha^ {(2)} +\alph a ^{(1) 2} - \be t a^{(1)}_i \be t a^ { ( 1) i} \,, \qquad
B_ i^{( 2 )} = \be t a^ {(2 ) }_i + 4 h_ {i j} ^ {(1)} \beta^{(1) j} \,,\qq uad
C_{i j}^{(2)} = h_{i j} ^ { ( 2)} \,.\ end{ g at h ered}$$
Const ructi on of thed istribut ion f unctionfor polar i s ed light {# app :Di st} = == ============= = = ==== == ======= === ======= === === === === == ======
W e consid er a t wo -di mensi o nal pola ri sat io n p laned efined by t wo u ni tc omp lex vec t or s $\ha t{{ \bol dsy mb ol{\e psil o n}} }_{({{\ mathrm{I} }}) } $ an d$\ hat {{\ boldsymbol{\e ps ilon}}}_{( {{ \ma thrm{I I } }})}$, w hich are mutually ortho g onal, $ \ha t \ep silo n_{({{\ma thr m{A}}} )}{ } ^{\sta r\mu}\hat\e psi l o n_{({ { \ ma thr m{ B}}})}{}^\ n u g_ {\mu\ nu }
= \delta_ {({{\mathrm{A}}})( { {\m athrm{B}}})}$ . | _i {{\rm_d}}x^i {{\rm_d}}\eta
+ (\delta_{i j}_+ 2_C_{i_j}) {{\rm_d}}x^i_{{\rm d}}x^j \Bigr]_\,,
\end{gathered}$$ where_$C_{i j}$ can be_further split into_2_scalar, 2 vector and 2 tensor degrees of freedom. By a direct comparison of_these_two parametrisations,_the_relation_between the two metric parametrisations_is
$$\begin{aligned}
g_{0 0} &: 1_+ 2_\alpha + \alpha^2 - \beta_i \beta^i = 1_+_2 A\,,\\
g_{0_i} &: \beta_i + 2 h_{i j} \beta^j =_B_i\,, \\
g_{i j} &: \delta_{i_j} + 2_h_{i_j}_= \delta_{i j} +_2 C_{i j}\,.\end{aligned}$$
At first order we_obtain $$\begin{gathered}
A^{(1)} = \alpha^{(1)} \,,_\qquad B_i^{(1)} = \beta^{(1)}_i \,, \qquad
C_{i_j}^{(1)} = h_{i j}^{(1)} \,,\end{gathered}$$ and_the two parametrisations are the_same. However,_at second order we get_the relations $$\begin{gathered}
_A^{(2)}= _\alpha^{(2)} + \alpha^{(1)_2} - \beta^{(1)}_i \beta^{(1) i} _\,, \qquad
B_i^{(2)}_= \beta^{(2)}_i + 4 h_{i j}^{(1)}_\beta^{(1)_j} \,,\qquad
_C_{i_j}^{(2)}_= h_{i_j}^{(2)} \,.\end{gathered}$$
Construction of_the_distribution function_for_polarised light {#app:Dist}
=============================================================
We consider a two-dimensional_polarisation_plane defined by two unit complex vectors_$\hat {{\boldsymbol{\epsilon}}}_{({{\mathrm{I}}})}$ and $\hat_{{\boldsymbol{\epsilon}}}_{({{\mathrm{II}}})}$,_which are mutually orthogonal,_$\hat \epsilon_{({{\mathrm{A}}})}{}^{\star\mu} \hat \epsilon_{({{\mathrm{B}}})}{}^\nu g_{\mu\nu}
_=\delta_{({{\mathrm{A}}})({{\mathrm{B}}})}$. |
1^\circ$) in both foreground removal and relative calibration (i.e., $E_{\rm for} = E_{\rm
cal} = 10^{-2}$), CORE will be able to improve the recovery of the CIB spectrum amplitude of a factor of about 17, to achieve a marginal detection of the energy release associated with astrophysical reionization models, and to improve by a factor of approximately 4 the limits on early energy dissipations. On the other hand, an improvement of a factor of 20 for CIB amplitude, of 10 for $u$, and of around 25 for the chemical potential $\mu_0$, is found by improving the relative calibration error to $\simeq 0.1$%. Any further improvement in foreground mitigation and calibration will enable still more precise results to be achieved.
Partial support by ASI/INAF Agreement 2014-024-R.1 for the [*Planck*]{} LFI Activity of Phase E2 and by ASI through the contract I-022-11-0 LSPE is acknowledged. C.H.-M. acknowledges financial support of the Spanish Ministry of Economy and Competitiveness via I+D project AYA-2015-66211-C2-2-P. J.G.N. acknowledges financial support from the Spanish MINECO for a [*Ramon y Cajal*]{} fellowship (RYC-2013-13256) and the I+D 2015 project AYA2015-65887-P (MINECO/FEDER). C.J.M. is supported by an FCT Research Professorship, contract reference IF/00064/2012, funded by FCT/MCTES (Portugal) and POPH/FSE (EC). M.Q. is supported by the Brazilian research agencies CNPq and FAPERJ. We acknowledge the use of the ESA Planck Legacy Archive. Some of the results in this paper have been derived using the [HEALPix]{} package. The use of the computational cluster at INAF-IASF Bologna is acknowledged. It is a pleasure to thank Arpine Kozmanyan for useful discussions on the [CosmoMC]{} sampler.
Appendix – Likelihoods of CMB dipole parameters {#sec:Likelihoods}
===============================================
For the sake of completeness, we report here the likelihoods computed for CMB dipole parameters and the confidence levels in their estimation. We limit the presentation here to the lowest and highest resolutions among those | 1^\circ$) in both foreground removal and relative calibration (i.e., $ E_{\rm for } = E_{\rm
cal } = 10^{-2}$), CORE will be able to improve the recovery of the CIB spectrum amplitude of a factor of about 17, to achieve a bare signal detection of the energy release consociate with astrophysical reionization model, and to improve by a factor of approximately 4 the limit on early department of energy dissipations. On the other hand, an improvement of a factor of 20 for CIB amplitude, of 10 for $ u$, and of around 25 for the chemical electric potential $ \mu_0 $, is found by improving the relative calibration error to $ \simeq 0.1$%. Any further improvement in foreground moderation and calibration will enable still more precise results to be achieved.
Partial documentation by ASI / INAF Agreement 2014 - 024 - R.1 for the [ * Planck * ] { } LFI Activity of Phase E2 and by ASI through the contract I-022 - 11 - 0 LSPE is acknowledged. C.H.-M. acknowledges financial accompaniment of the Spanish Ministry of Economy and Competitiveness via I+D project AYA-2015 - 66211 - C2 - 2 - P. J.G.N. acknowledges financial support from the Spanish MINECO for a [ * Ramon y Cajal * ] { } family (RYC-2013 - 13256) and the I+D 2015 project AYA2015 - 65887 - P (MINECO / FEDER). C.J.M. is supported by an FCT Research Professorship, contract reference IF/00064/2012, funded by FCT / MCTES (Portugal) and POPH / FSE (EC). M.Q. is support by the Brazilian research agencies CNPq and FAPERJ. We acknowledge the use of the ESA Planck Legacy Archive. Some of the results in this newspaper have been derived using the [ HEALPix ] { } package. The use of the computational cluster at INAF - IASF Bologna is acknowledged. It is a pleasure to thank Arpine Kozmanyan for useful discussions on the [ CosmoMC ] { } sampler.
Appendix – Likelihoods of CMB dipole parameter { # sec: Likelihoods }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
For the sake of completeness, we report here the likelihoods calculate for CMB dipole argument and the assurance levels in their estimation. We limit the presentation here to the lowest and highest resolutions among those | 1^\cirf$) in both foreground remuval and relatirw caliuration (i.e., $E_{\rm wor} = E_{\rm
cal} = 10^{-2}$), CORE will be aule ro imkgove the recovery of ghe CIB siectrum anpliuude of a factor of about 17, to admieve e marginal detegtion of tha energy releave adsociated with astrophysical reionisation kofels, and to imkrove fy a factor of approximately 4 the limifs on eerly energy disxipations. On the other hanf, an improvement of a vactor of 20 dor SUB amplitude, of 10 for $u$, and of arouhd 25 for the chemical potential $\ou_0$, is found by umproghng the reletive balibration evgor to $\vimeq 0.1$%. Sny further imirovekenr in foreground mitigetion and calibratiog will endbme still more prexiwe revultv to ve xchjeted.
Lartiap snpport by AAI/INAF Agrewment 2014-024-R.1 for the [*Plamch*]{} LFI Activity of Phwsq E2 and by ASI through the contract I-022-11-0 LVPE is acknowledged. C.H.-M. acjnowledges financial dupport os the Spanish Ministry of Economy and Competitivetess tix I+B projdxt AYA-2015-66211-C2-2-P. J.G.N. acknowledges financial support froi tne Spanish MINECJ for a [*Rampn y Sajal*]{} fellowsfip (RYE-2013-13256) ahd the I+D 2015 project AYA2015-65887-P (MYNECO/DEDER). C.J.M. is xupported by an FCT Researcy Professorskip, contract referencz IF/00064/2012, funded yy FCT/KCTES (Portugal) and POPH/FSE (EE). M.Q. ia supported by the Bdxzilian research agvncias CNPq and FAPERJ. We acknjwledge tie usz of the ESA Planch Legacy Agchivc. Some of the resulhs in tvis paper jave been derived using the [HEALPix]{} package. Tne usv of the eomputstional clustqr at INAF-IASF Bolognc is azknowledges. It is a pleasure to thank Arphje Kozmanyan for usesul eiscyssions un the [CosmoMC]{} sampler.
Aiptndix – Likwlihoods of CMB diiole kadameters {#sec:Likzuuhoods}
===============================================
For the sske of clm'letegass, we repord hefe gne liyelihoods cioputrd for CMB dipole pasamefers and the conficekce levelw in theyr estimation. We limit the prestntatimn iere tp tre lowest and highest resolutikns among thjse | 1^\circ$) in both foreground removal and relative $E_{\rm = E_{\rm = 10^{-2}$), CORE the of the CIB amplitude of a of about 17, to achieve a detection of the energy release associated with astrophysical reionization models, and to improve a factor of approximately 4 the limits on early energy dissipations. On the hand, improvement a of 20 for CIB amplitude, of 10 for $u$, and of around 25 for the chemical $\mu_0$, is found by improving the relative calibration to $\simeq 0.1$%. Any improvement in foreground mitigation and will still more results be Partial support by Agreement 2014-024-R.1 for the [*Planck*]{} LFI Activity of Phase E2 and by ASI through the contract I-022-11-0 is acknowledged. financial support the Ministry Economy and Competitiveness project AYA-2015-66211-C2-2-P. J.G.N. acknowledges financial support MINECO for a [*Ramon y Cajal*]{} fellowship (RYC-2013-13256) the I+D project AYA2015-65887-P (MINECO/FEDER). C.J.M. is supported an FCT Research Professorship, contract reference IF/00064/2012, funded FCT/MCTES (Portugal) and POPH/FSE (EC). M.Q. is supported by the Brazilian research agencies CNPq and acknowledge the use of ESA Planck Legacy Some the in paper have derived using the [HEALPix]{} package. The use of the computational cluster INAF-IASF Bologna is acknowledged. It is a pleasure to thank for discussions on the sampler. Appendix – Likelihoods CMB parameters {#sec:Likelihoods} =============================================== For of we likelihoods for dipole parameters and the levels in their estimation. We the presentation here to among those | 1^\circ$) in both foreground removAl and relatIve caLibRatIoN (i.e., $E_{\Rm foR} = E_{\rm
cal} = 10^{-2}$), CORE wilL Be abLe to improve the recovery Of the cIb SpecTRuM amplItude of A FaCTOr oF aBoUt 17, tO aCHiEve a mArgInal detEction of thE enErGy release assOCiAted with asTroPhysical reioNizAtion mOdEls, ANd to iMprOve by A factoR Of apprOximately 4 ThE Limits ON early eNERgY disSipations. On the othER hANd, an improvemenT of a faCtOR oF 20 FOr CiB aMplitude, of 10 FoR $u$, and OF around 25 FOr THE CheMIcal potential $\Mu_0$, is found by IMprOving tHe RelATive caLibraTiON erRor to $\simeq 0.1$%. ANy fuRther imprOvemenT In foregROund mitIgatioN anD caLibrATiOn WilL eNAblE StIll MOre Precise rEsUlTs to bE achIEVED.
ParTiaL supPort bY ASI/INAF AgreeMenT 2014-024-R.1 foR The [*plancK*]{} LFI ACtivItY of PhAse E2 anD by ASi tHrough the contraCt I-022-11-0 LsPE is acknOwlEdGed. c.H.-m. acknOWledgeS fiNanCial supPort of tHE SpAnISH miNistry of Economy and coMPEtItivenesS via I+D PRoJeCT AYA-2015-66211-C2-2-P. J.G.n. aCknOwleDGEs finAnciAL sUpport frOm the SPAnIsH MINECO FoR a [*RamoN y cajAl*]{} fEllowSHip (RyC-2013-13256) and tHe I+D 2015 projEct AYa2015-65887-p (MINECO/FEDER). C.J.m. Is supported by AN Fct reSEarcH PrOfessorship, ContRAct rEferENcE IF/00064/2012, FUnded By FCT/mCteS (pOrtugal) and POPH/FSE (EC). m.Q. Is suppOrted By the BraziliaN research aGENCies CNPq And FapErj. We acknowledge The usE of the ESA PLAnck LegaCy ArcHive. Some Of the resuLTS in this pApeR haVe bEen DERiVed using the [HEalpix]{} pAcKage. The Use Of the coMpuTatIonAl cLuSter at INAf-IASF BolOgNa Is AcKnoWledgED. It is a plEaSurE tO thAnk ArPIne KozManyaN for UsEfUL diScussioNS oN THe [CoSmOMc]{} samPleR.
APpendIx – LiKEliHoods of cMB dipole ParAMeteRs {#SeC:LikeliHoods}
===============================================
For the saKe Of completeNeSs, wE reporT HEre the liKelihoods computed for CMB DIpole paRamEters And tHe confideNce Levels In tHEir estImatioN. We liMiT thE PResenTATiOn hErE to the loweST And HigheSt ResoLutions Among those | 1^\circ$) in both foregrou nd removal andrel ati ve cal ibra tion (i.e., $E _ {\rm for} = E_{\rm
cal} =10^{- 2} $ ), C O RE will be abl e t o imp ro ve th er ec overy of the CI B spectrum am pl itude of a f a ct or of abou t 1 7, to achiev e a margi na l d e tecti onof th e ener g y rele ase assoc ia t ed wit h astrop h y si calreionization mode l s, and to improve by afa c to r ofapp roximately 4 thel imits o n e a r l y e n ergy dissipat ions. On th e ot her ha nd , a n impro vemen to f a factor of20 f or CIB am plitud e , of 10 for $u$ , andofaro und2 5fo r t he che m ic alp ote ntial $\ mu _0 $, is fou n d b y im pro ving therelative cali bra tion err or to $\si meq0. 1$%.Any fu rther i mprovement in f oreg round mit iga ti onan d cal i bratio n w ill enable stillm ore p r e c is e results to be ac hi e v ed .
Parti al sup p or tb y ASI/IN AF Ag reem e n t 201 4-02 4 -R .1 for t he [*P l an ck *]{} LF IActivi ty of Ph ase E 2 and by AS I throug h the contract I-022 - 11-0 LSPE isa ck n o wl e dged . C .H.-M. ackn owle d gesfina n ci als uppor t ofth e S p anish Ministry of E co nomy a nd Co mpetitiveness via I+D p r o j ect AYA- 2015 - 66 2 11-C2-2-P. J.G .N. a cknowledge s financi al su pport fr om the Sp a n ish MINE COfor a[*R a m on y Cajal*]{}f e llow sh ip (RYC -20 13-1325 6)and th e I +D 2015 pro ject AYA 20 15 -6 58 87- P (MI N ECO/FEDE R) . C .J .M. is s u pporte d byan F CT R e sea rch Pro f es s o rshi p, c ontr act r efere nceI F/0 0064/20 12, funde d b y FCT /M CT ES (Por tugal) and PO PH /FSE (EC). M .Q. is su p p orted by the Brazilian research agencie s C NPq a nd F APERJ. We ac knowle dge the us e of t he ES APla n c k Leg a c yArc hi ve. Some o f the resu lt s in this p aper have been der i ved using the [H EAL Pix] { } p ack a ge . Th eu seo f the computatio nal cluste ra tINAF-IASFB olo gn a is ac knowled ged.I t is apleasureto thankAr pine K ozm anyan foruseful d iscussion s on t h e[Cosm oMC ]{} sa mp ler .
Ap pendix – L ikeli hoodsof CMB d ipole p arameter s {#sec:Likelihoods}
== ====== ===== === ========= === = === ========= ==== ==
For th e s ake of c omp l etene ss,w erep o rt he re t h e likelih o od s c o m pu ted for CMB d i pol e par ame t ers an d th e confidence leve l s in their est imat i o n.Wel imit t he presentatio n h er e to the l ow est and hig hest res ol u tions among those | 1^\circ$) in_both foreground_removal and relative calibration_(i.e., $E_{\rm_for}_= E_{\rm
cal}_=_10^{-2}$), CORE will_be able to_improve the recovery of_the CIB spectrum_amplitude_of a factor of about 17, to achieve a marginal detection of the energy_release_associated with_astrophysical_reionization_models, and to improve by_a factor of approximately 4_the limits_on early energy dissipations. On the other hand,_an_improvement of a_factor of 20 for CIB amplitude, of 10 for_$u$, and of around 25 for_the chemical potential_$\mu_0$,_is_found by improving the_relative calibration error to $\simeq 0.1$%._Any further improvement in foreground mitigation_and calibration will enable still more precise_results to be achieved.
Partial support by_ASI/INAF Agreement 2014-024-R.1 for the_[*Planck*]{} LFI_Activity of Phase E2 and_by ASI through_the contract_I-022-11-0 LSPE is_acknowledged. C.H.-M. acknowledges financial support of_the Spanish Ministry_of Economy and Competitiveness via I+D_project_AYA-2015-66211-C2-2-P. J.G.N. acknowledges_financial_support_from the_Spanish MINECO for_a_[*Ramon y_Cajal*]{}_fellowship (RYC-2013-13256) and the I+D 2015_project_AYA2015-65887-P (MINECO/FEDER). C.J.M. is supported by an_FCT Research Professorship, contract_reference_IF/00064/2012, funded by FCT/MCTES_(Portugal) and POPH/FSE (EC). M.Q._is supported by the Brazilian research_agencies CNPq_and FAPERJ._We acknowledge the use of the ESA Planck Legacy Archive. Some_of the results in this paper_have been derived using_the [HEALPix]{}_package._The use of_the_computational cluster_at INAF-IASF Bologna is acknowledged. It is_a pleasure_to thank Arpine Kozmanyan for useful_discussions on the [CosmoMC]{}_sampler.
Appendix_– Likelihoods of CMB dipole parameters_{#sec:Likelihoods}
===============================================
For the sake of completeness, we_report here the likelihoods computed_for_CMB_dipole parameters and the confidence_levels in their estimation. We limit_the presentation here_to the lowest and highest resolutions among_those |
order $2(q-1)$, with orbits of size $2$ and $q- 1$ on $\Omega$ [@Suz Theorem 9].
For the final assertion, if $H\cap \Omega$ lies in two hyperplanes then it is in a plane, and hence is a conic, which is not the case [@Ti3 pp. 51-52].
---
abstract: 'We consider random walks (RWs) and self-avoiding walks (SAWs) on disordered lattices directly at the percolation threshold. Applying numerical simulations, we study the scaling behavior of the models on the incipient percolation cluster in space dimensions $d=2, 3, 4$. Our analysis yields estimates of universal exponents, governing the scaling laws for configurational properties of RWs and SAWs.'
address:
- '$^1$ Institut für Theoretische Physik and Centre for Theoretical Sciences (NTZ),Universität Leipzig, Postfach 100920, D-04009 Leipzig, Germany'
- '$^2$ Institute for Condensed Matter Physics, National Academy of Sciences of Ukraine, UA–79011 Lviv, Ukraine'
author:
- 'V Blavatska$^{1,2}$ and W Janke$^1$'
title: 'Walking on fractals: diffusion and self-avoiding walks on percolation clusters'
---
Introduction
============
The model of a random walk (RW) provides a good description of diffusion processes, such as for example encountered for electrons in metals or colloidal particles in solution [@RWbook]. The averaged mean square displacement of the diffusive particle at time $t$ (or, equivalently, after $t$ steps on a lattice) scales as $$\langle R^2 \rangle \sim t^{2\nu_{{\rm RW}}}, \label{diff}$$ where in a non-disordered medium $\nu_{{\rm RW}}=1/2$, independently of the space dimension $d$. A RW is a fractal object, with fractal dimension $d_{{\rm RW}}=1/\nu_{RW}$. The number of all possible trajectories $C_t$ for a randomly walking particle of $t$ steps can be found exactly: $C_t=z_0^t$, where $z_0$ is the coordination number of the corresponding lattice.
Forbidding the trajectory of a random walk to cross itself | order $ 2(q-1)$, with orbits of size $ 2 $ and $ q- 1 $ on $ \Omega$ [ @Suz Theorem 9 ].
For the final assertion, if $ H\cap \Omega$ dwell in two hyperplanes then it is in a airplane, and hence is a conic, which is not the subject [ @Ti3 pp. 51 - 52 ].
---
abstract:' We study random walks (RWs) and self - avoiding walk (SAWs) on disorder lattices immediately at the percolation brink. Applying numeral simulations, we study the scaling demeanor of the models on the incipient percolation cluster in space dimensions $ d=2, 3, 4$. Our analysis yields estimates of cosmopolitan exponents, governing the scaling laws for configurational property of RWs and SAWs.'
address:
-' $ ^1 $ Institut für Theoretische Physik and Centre for Theoretical Sciences (NTZ),Universität Leipzig, Postfach 100920, D-04009 Leipzig, Germany'
-' $ ^2 $ Institute for Condensed Matter Physics, National Academy of Sciences of Ukraine, UA–79011 Lviv, Ukraine'
writer:
-' V Blavatska$^{1,2}$ and W Janke$^1 $'
championship:' Walking on fractals: diffusion and self - avoiding walks on percolation bunch'
---
Introduction
= = = = = = = = = = = =
The model of a random walk (RW) provides a good description of diffusion process, such as for exercise encountered for electrons in metals or colloidal particles in solution [ @RWbook ]. The averaged mean square displacement of the diffusing atom at time $ t$ (or, equivalently, after $ t$ steps on a lattice) scale as $ $ \langle R^2 \rangle \sim t^{2\nu_{{\rm RW } } }, \label{diff}$$ where in a non - disordered medium $ \nu_{{\rm RW}}=1/2 $, independently of the space property $ d$. A RW is a fractal object, with fractal dimension $ d_{{\rm RW}}=1/\nu_{RW}$. The number of all possible trajectory $ C_t$ for a randomly walking particle of $ t$ steps can be found exactly: $ C_t = z_0^t$, where $ z_0 $ is the coordination number of the corresponding lattice.
Forbidding the trajectory of a random walk to cross itself | orfer $2(q-1)$, with orbits of sizt $2$ and $q- 1$ on $\Omegc$ [@Suz Tieorem 9].
Fkr the fknal assertion, if $H\cap \Omega$ lues ib two hyperplanes then it is in a plane, and yence is a conic, wmnch ia not vhe case [@Ti3 pp. 51-52].
---
sbstract: 'Wa consider rangoo calks (RWs) and self-avoiding walks (SAWf) on dixogdered latticef digestly at the percolation threshold. Applging nukerical simulstions, we study the scalinh bejavior of the modeps on the ibcipywnt percolatkon clusteg in space simensions $d=2, 3, 4$. Our analysis yieuds extimates od yniggrsal exponeits, gonerning the sgsling naws fot configuratiokal pcopeeties of RWs and SAWs.'
eddress:
- '$^1$ Institut für Theoretivcke Physik and Centre doe Themrethcal Wcidncts (ITZ),Hniverditäv Leipzig, Pkstfach 100920, D-04009 Oeipzig, Germany'
- '$^2$ Inxtynite for Condsnsed Iaeter Physics, National Academy of Sciencts of Ukraine, UA–79011 Lviv, Ukrainw'
author:
- 'V Blavatska$^{1,2}$ ajd W Jankq$^1$'
title: 'Walking on fractals: diffusion and self-avoiging xauks ok pefxopation clusters'
---
Introduction
============
The model of a ranqkm wslk (RW) providef a good dexcgiljion of diffuskon proceases, such as for edample gncounrered for elevtrons in metals or colloidql particles un solution [@RWbook]. The averagzd meam squsre displacement of the difrusive partlcle at tjoe $t$ (or, equivalevtlj, afder $t$ sttos on a lattice) ssales as $$\oangke R^2 \ravgle \sim t^{2\gu_{{\rm RW}}}, \lahel{dinx}$$ where in a non-didordeted megium $\nu_{{\rm GW}}=1/2$, independently of the space dmkension $d$. A TW hs d fractaj objcct, with fractaj dimension $d_{{\rk RW}}=1/\nu_{RC}$. The vumber of zll posvible trajestories $C_t$ fos a randomly xalking pwrtixle if $t$ stdos can be founc exactly: $C_t=z_0^t$, wherw $z_0$ is the coordinetiov number of the euteesponding latyicd.
Fowbpddmng tra trajectory of x rxmdom dalk to crowr itxelf | order $2(q-1)$, with orbits of size $2$ 1$ $\Omega$ [@Suz 9]. For the lies two hyperplanes then is in a and hence is a conic, which not the case [@Ti3 pp. 51-52]. --- abstract: 'We consider random walks (RWs) self-avoiding walks (SAWs) on disordered lattices directly at the percolation threshold. Applying numerical we the behavior the models on the incipient percolation cluster in space dimensions $d=2, 3, 4$. Our analysis yields of universal exponents, governing the scaling laws for properties of RWs and address: - '$^1$ Institut für Physik Centre for Sciences Leipzig, 100920, D-04009 Leipzig, - '$^2$ Institute for Condensed Matter Physics, National Academy of Sciences of Ukraine, UA–79011 Lviv, Ukraine' author: 'V Blavatska$^{1,2}$ Janke$^1$' title: on diffusion self-avoiding walks on --- Introduction ============ The model of (RW) provides a good description of diffusion processes, as for encountered for electrons in metals or particles in solution [@RWbook]. The averaged mean square of the diffusive particle at time $t$ (or, equivalently, after $t$ steps on a lattice) $$\langle R^2 \rangle \sim RW}}}, \label{diff}$$ where a medium RW}}=1/2$, of the dimension $d$. A RW is a fractal object, with fractal dimension RW}}=1/\nu_{RW}$. The number of all possible trajectories $C_t$ for a particle $t$ steps can found exactly: $C_t=z_0^t$, where is coordination number of the Forbidding trajectory walk cross | order $2(q-1)$, with orbits of size $2$ and $Q- 1$ on $\Omega$ [@SuZ TheoRem 9].
for ThE finAl asSertion, if $H\cap \OMEga$ lIes in two hyperplanes theN it is In A PlanE, AnD hencE is a conIC, wHICh iS nOt The CaSE [@TI3 pp. 51-52].
---
abStrAct: 'We coNsider randOm wAlKs (RWs) and self-AVoIding walks (sAWS) on disordereD laTtices DiRecTLy at tHe pErcolAtion tHResholD. Applying NuMErical SImulatiONS, wE stuDy the scaling behavIOr OF the models on thE incipIeNT pERColAtiOn cluster iN sPace dIMensionS $D=2, 3, 4$. OUR ANalYSis yields estiMates of univERsaL exponEnTs, gOVerninG the sCaLIng Laws for confIgurAtional prOpertiES of RWs aND SAWs.'
adDress:
- '$^1$ INstItuT für tHeOrEtiScHE PhYSiK anD cenTre for ThEoReTical scieNCES (nTZ),UNivErsiTät LeIpzig, Postfach 100920, d-04009 LeIpziG, gerMany'
- '$^2$ INstitUte fOr condeNsed MaTter PHySics, National AcaDemy Of ScienceS of ukRaiNe, uA–79011 LviV, ukrainE'
auThoR:
- 'V BlavaTska$^{1,2}$ and w janKe$^1$'
TITLe: 'walking on fractals: dIfFUSiOn and selF-avoidINg WaLKs on percOlAtiOn clUSTers'
---
INtroDUcTion
============
The mOdel of A RaNdOm walk (Rw) pRovideS a GooD deScripTIon oF diffuSion procEsses, SUch as for examplE Encountered foR ElECTrONs in MetAls or colloiDal pARticLes iN SoLutIOn [@RWbOok]. ThE aVErAGed mean square displaCeMent of The diFfusive particLe at time $t$ (oR, EQUivalentLy, afTEr $T$ Steps on a latticE) scalEs as $$\langle r^2 \Rangle \siM t^{2\nu_{{\rM RW}}}, \label{Diff}$$ where IN A non-disoRdeRed MedIum $\NU_{{\Rm rW}}=1/2$, independentLY Of thE sPace dimEnsIon $d$. A RW Is a FraCtaL obJeCt, with fraCtal dimeNsIoN $d_{{\Rm rW}}=1/\nU_{RW}$. ThE Number of AlL poSsIblE trajECtorieS $C_t$ foR a raNdOmLY waLking paRTiCLE of $t$ StEpS can Be fOuNd exaCtly: $c_T=z_0^t$, Where $z_0$ iS the coordInaTIon nUmBeR of the cOrresponding lAtTice.
ForbidDiNg tHe trajECTory of a rAndom walk to cross itself | order $2(q-1)$, with orbi ts of size $2$and $q -1$ o n $\ Omega$ [@Suz T h eore m 9].
For the final a ssert io n , if $H \cap\Omega$ li e s in t wo hy pe r pl anesthe n it is in a plan e,an d hence is a co nic, which is not the cas e [ @Ti3 p p. 51 - 52]. -- -
abs tract: 'We co nsider ra nd o m walk s (RWs)a n dself -avoiding walks ( S AW s ) on disordere d latt ic e sd i rec tly at the pe rc olati o n thres h ol d . App l ying numerica l simulatio n s,we stu dy th e scali ng be ha v ior of the mod elson the in cipien t percol a tion cl usterinspa ce d i me ns ion s$ d=2 , 3 , 4 $ . O ur analy si syield s es t i m a tesofuniv ersal exponents, g ove rnin g th e sca linglaws f or co nfigur ation al properties ofRWsand SAWs. '
a dd res s:
- '$ ^ 1$ Ins tit utfür The oretisc h e P hy s i k a nd Centre for Theo re t i ca l Scienc es (NT Z ), Un i versität L eip zig, P ostfa ch 1 0 09 20, D-04 009 Le i pz ig , Germa ny '
- '$ ^2 $ I nst itute forConden sed Matt er Ph y sics, National Academy of Sc i en c e so f Uk rai ne, UA–7901 1 Lv i v, U krai n e'
au t hor:- 'VBl a va t ska$^{1,2}$ and W J an ke$^1$ '
tit le: 'Walkingon fractal s : diffusio n an d s e lf-avoiding wa lks o n percolat i on clust ers'---
Int roduction = ======== ===
T hemod e l o f a random wa l k (RW )provide s a good d esc rip tio n o fdiffusion process es ,su ch as fore xample e nc oun te red fore lectro ns in met al so r c olloida l p a r ticl es i n so lut io n [@R Wboo k ].The ave raged mea n s q uare d is placeme nt of the dif fu sive parti cl e a t time $ t$ (or,equivalently, after $t$ steps o n a latt ice) scales a s $ $\lang leR ^2 \ra ngle \ sim t ^{ 2\n u _ {{\rm R W} }}, \ label{diff } $ $ w herein a n on-diso rdered medium $\nu _ {{\ rm RW}}=1/2$, in depe n d en tly of the s p ace d imension $d$. A RW is a f ra c ta l object,w ith f ractaldimensi on $d _ {{\rm R W}}=1/\nu _{RW}$. T he num b e r o f all poss ible tra jectories $C_t$ fo r a r and omly w al kin g par ticleo f $ t$ st eps ca nbe fou nd ex ac tly: $C_ t=z_0^t$, where $z_0$ i s thecoord ina tion numb ero f t he corres pond ing lattic e.
Fo rbidd ing the t raje c to ryo f a r ando m walk toc ro ssi t se lf | order_$2(q-1)$, with_orbits of size $2$_and $q-_1$_on $\Omega$_[@Suz_Theorem 9].
For the final_assertion, if $H\cap_\Omega$ lies in two_hyperplanes then it_is_in a plane, and hence is a conic, which is not the case [@Ti3_pp. 51-52].
_---
abstract: 'We_consider_random_walks (RWs) and self-avoiding walks_(SAWs) on disordered lattices directly_at the_percolation threshold. Applying numerical simulations, we study the_scaling_behavior of the_models on the incipient percolation cluster in space dimensions_$d=2, 3, 4$. Our analysis yields_estimates of universal_exponents,_governing_the scaling laws for_configurational properties of RWs and SAWs.'
address:
-_'$^1$ Institut für Theoretische Physik and_Centre for Theoretical Sciences (NTZ),Universität Leipzig, Postfach_100920, D-04009 Leipzig, Germany'
- '$^2$ Institute_for Condensed Matter Physics, National_Academy of_Sciences of Ukraine, UA–79011 Lviv,_Ukraine'
author:
- 'V Blavatska$^{1,2}$_and W_Janke$^1$'
title: 'Walking on_fractals: diffusion and self-avoiding walks on_percolation clusters'
---
Introduction
============
The model_of a random walk (RW) provides_a_good description of_diffusion_processes,_such as_for example encountered_for_electrons in_metals_or colloidal particles in solution [@RWbook]._The_averaged mean square displacement of the diffusive_particle at time $t$_(or,_equivalently, after $t$ steps_on a lattice) scales as_$$\langle R^2 \rangle \sim t^{2\nu_{{\rm RW}}},_\label{diff}$$ where_in a_non-disordered medium $\nu_{{\rm RW}}=1/2$, independently of the space dimension $d$. A_RW is a fractal object, with_fractal dimension $d_{{\rm RW}}=1/\nu_{RW}$._The number_of_all possible trajectories_$C_t$_for a_randomly walking particle of $t$ steps can_be found_exactly: $C_t=z_0^t$, where $z_0$ is the_coordination number of the_corresponding_lattice.
Forbidding the trajectory of a random_walk to cross itself |
rectangle spanned by the points $(1,1)$ and $(h(p),w(p))$, every $v$-edge points downwards, and every $h$-edge points rightwards; recall that our coordinates represent the rows and columns of a matrix. The derivation leading to the final configuration $(M,{\emptyset},g)$ simulates the assembly of tiles which form $\bar p$ as shown in Figure \[fig:simulationmethod\]. The north and west edges of the tile $t_{TL} = \bar p_{(1,1)}$ in the top left corner of $\bar p$ are labelled by $\#$, and therefore, the node $x_{TL} = {\vartheta}^{-1}(t_{TL})$ corresponding to $t_{TL}$ forms an initial graph $M_0$. The adjacent edges of two neighbouring tiles $s,t$ in $\bar p$ are labelled by the same colour. Suppose $s$ is the west neighbour of $t$, then $\sigma_E(s) = \sigma_{W}(t) \neq \#$ and both tiles belong to the same row, implying that $\sigma_N(s)=\# \iff \sigma_N(t) = \#$ and $\sigma_S(s) = \# \iff \sigma_S(t) = \#$. Therefore, their corresponding nodes in $G$ are connected by an $h$-edge $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_h$. Analogously, if $s$ is the north neighbour of $t$, then $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_v$. Next, we see that the hyperedges of type I and type II can be used in order to create the top row and left column of the graph $M$, respectively. Furthermore, the hyperedges of type III can be used in order to create all the remaining nodes of $M$. We conclude that $(M_0,\set{x_{TL}},id)\overset{*}{\underset{A}\rightarrow} (M,O,g)$ is a derivation in $A$ and we will prove that $(M,O,g)$ has to be a final configuration with $O={\emptyset}$. Observe, that hyperedges of types I and II leave all the nodes active while hyperedges of type III deactivate at least the top left node in the hyperedge. | rectangle spanned by the points $ (1,1)$ and $ (h(p),w(p))$, every $ v$-edge points downwards, and every $ h$-edge points rightwards; echo that our coordinate represent the rows and column of a matrix. The ancestry leading to the final shape $ (M,{\emptyset},g)$ simulates the assembly of tiles which shape $ \bar p$ as shown in Figure \[fig: simulationmethod\ ]. The north and west edge of the tile $ t_{TL } = \bar p_{(1,1)}$ in the top left corner of $ \bar p$ are pronounce by $ \#$, and therefore, the node $ x_{TL } = { \vartheta}^{-1}(t_{TL})$ corresponding to $ t_{TL}$ forms an initial graph $ M_0$. The adjacent edges of two neighbouring tiles $ second, t$ in $ \bar p$ are labelled by the same colour. presuppose $ s$ is the west neighbour of $ t$, then $ \sigma_E(s) = \sigma_{W}(t) \neq \#$ and both tiles belong to the same quarrel, implying that $ \sigma_N(s)=\ # \iff \sigma_N(t) = \#$ and $ \sigma_S(s) = \ # \iff \sigma_S(t) = \#$. Therefore, their corresponding nodes in $ G$ are connected by an $ h$-edge $ ({ \vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_h$. Analogously, if $ s$ is the union neighbour of $ t$, then $ ({ \vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_v$. Next, we see that the hyperedges of type I and type II can be used in order to make the top row and left column of the graph $ M$, respectively. Furthermore, the hyperedges of type III can be used in order to create all the remaining nodes of $ M$. We conclude that $ (M_0,\set{x_{TL}},id)\overset{*}{\underset{A}\rightarrow } (M, O, g)$ is a derivation in $ A$ and we will testify that $ (M, O, g)$ has to be a final shape with $ O={\emptyset}$. Observe, that hyperedges of type I and II bequeath all the nodes active while hyperedges of type III deactivate at least the top left node in the hyperedge. | reftangle spanned by the puints $(1,1)$ and $(h(p),w(p))$, every $t$-edge pkints dodnwards, and every $h$-edge poinvs rughtwqrds; recall that our cuordinated represwnt uhe rows and columns of a matris. The verivation leadlng to the xinal configurdtkou $(M,{\emptyset},g)$ simulates the assembly jf tilex ahich form $\bar p$ ax shosn in Figure \[fig:simulationmethod\]. Ths north and west edgrs of the tile $t_{TL} = \bar p_{(1,1)}$ ln tje top left corner of $\bar p$ ats lwvelled by $\#$, avd therefogz, the node $s_{TL} = {\vartheta}^{-1}(t_{TL})$ corresponding go $t_{TK}$ forms an ibitldl graph $M_0$. Vhe adtacent edges of two teighboiring tiles $s,t$ in $\bae p$ are labelled by tie same colour. Suppofe $s$ is tve west neighbour od $r$, thet $\sicma_E(r) = \skgmz_{W}(v) \nsq \#$ anf bkth tiles gelong to tye same row, implyinb eyat $\sigma_N(s)=\# \irf \sigia_G(t) = \#$ and $\sigma_S(s) = \# \iff \sigma_S(t) = \#$. Therexors, their corresponding nides in $G$ are connectgd by an $h$-qdge $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_h$. Analogously, if $s$ is dhe nkfth ncighciug of $t$, then $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_v$. Next, we sqs uhan the hyperedges jf type I anc hylg II can be usea in oxsed to create the too row agd ledt column of yhe graph $M$, respectively. Fuethermore, thv hyperedges of type IIN can be useb in otder tp create all the remainnng noses of $M$. We conclude ghat $(M_0,\set{x_{TL}},id)\ovdrsvt{*}{\ungerset{A}\rightarrow} (M,O,g)$ is a derivatiin iu $A$ and de woll prjve that $(M,L,g)$ has to be a final convigurction with $O={\emphyset}$. Observe, that hyperedges oh types I and IO neane all thz nodex active whilq hyperedges oy type IIN deacgivate at meast tie top left gode in the hfkeredge. | rectangle spanned by the points $(1,1)$ and $v$-edge downwards, and $h$-edge points rightwards; the and columns of matrix. The derivation to the final configuration $(M,{\emptyset},g)$ simulates assembly of tiles which form $\bar p$ as shown in Figure \[fig:simulationmethod\]. The and west edges of the tile $t_{TL} = \bar p_{(1,1)}$ in the top corner $\bar are by $\#$, and therefore, the node $x_{TL} = {\vartheta}^{-1}(t_{TL})$ corresponding to $t_{TL}$ forms an initial graph The adjacent edges of two neighbouring tiles $s,t$ $\bar p$ are labelled the same colour. Suppose $s$ the neighbour of then = \neq \#$ and tiles belong to the same row, implying that $\sigma_N(s)=\# \iff \sigma_N(t) = \#$ and $\sigma_S(s) = \# \sigma_S(t) = their corresponding in are by an $h$-edge Analogously, if $s$ is the north then $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_v$. Next, we see that the of type and type II can be used order to create the top row and left of the graph $M$, respectively. Furthermore, the hyperedges of type III can be used in create all the remaining of $M$. We that (M,O,g)$ a in $A$ we will prove that $(M,O,g)$ has to be a final configuration $O={\emptyset}$. Observe, that hyperedges of types I and II leave nodes while hyperedges of III deactivate at least top node in the hyperedge. | rectangle spanned by the poinTs $(1,1)$ and $(h(p),w(p))$, eVery $v$-EdgE poInTs doWnwaRds, and every $h$-edGE poiNts rightwards; recall thaT our cOoRDinaTEs RepreSent the ROwS ANd cOlUmNs oF a MAtRix. ThE deRivatioN leading to The FiNal configuraTIoN $(M,{\emptyset},G)$ siMulates the asSemBly of tIlEs wHIch foRm $\bAr p$ as Shown iN figure \[Fig:simulaTiONmethoD\]. the nortH ANd West Edges of the tile $t_{TL} = \BAr P_{(1,1)}$ In the top left coRner of $\BaR P$ aRE LabEllEd by $\#$, and theReFore, tHE node $x_{Tl} = {\VaRTHEta}^{-1}(T_{tL})$ correspondiNg to $t_{TL}$ formS An iNitial GrAph $m_0$. the adjAcent EdGEs oF two neighboUrinG tiles $s,t$ iN $\bar p$ aRE labellED by the sAme colOur. supPose $S$ Is ThE weSt NEigHBoUr oF $T$, thEn $\sigma_E(S) = \sIgMa_{W}(t) \nEq \#$ anD BOTH tilEs bElonG to thE same row, implyIng That $\SIgmA_N(s)=\# \ifF \sigmA_N(t) = \#$ aNd $\Sigma_s(s) = \# \iff \sIgma_S(T) = \#$. THerefore, their coRresPonding noDes In $g$ arE cOnnecTEd by an $H$-edGe $({\vArtheta}^{-1}(S),{\vartheTA}^{-1}(t))\iN E_H$. aNAlOgously, if $s$ is the norTh NEIgHbour of $t$, Then $({\vaRThEtA}^{-1}(S),{\varthetA}^{-1}(t))\In E_V$. NexT, WE see tHat tHE hYperedgeS of typE i aNd Type II cAn Be used In OrdEr tO creaTE the Top row And left cOlumn OF the graph $M$, respECtively. FurtheRMoRE, ThE HypeRedGes of type IIi can BE useD in oRDeR to CReate All thE rEMaINing nodes of $M$. We conclUdE that $(M_0,\Set{x_{Tl}},id)\overset{*}{\undErset{A}\righTARRow} (M,O,g)$ is A derIVaTIon in $A$ and we wilL provE that $(M,O,g)$ haS To be a finAl conFiguratiOn with $O={\emPTYset}$. ObseRve, ThaT hyPerEDGeS of types I and Ii LEave AlL the nodEs aCtive whIle HypEreDgeS oF type III dEactivatE aT lEaSt The Top leFT node in tHe HypErEdgE. | rectangle spanned by thepoints $(1 ,1)$and $( h( p),w (p)) $, every $v$-e d ge p oints downwards, and e very$h $ -edg e p oints rightw a rd s ; re ca ll th at ou r coo rdi nates r epresent t hero ws and colum n sof a matri x.The derivati onleadin gtot he fi nal conf igurat i on $(M ,{\emptys et } ,g)$ s i mulates t he ass embly of tiles wh i ch form $\bar p$as sho wn in F igu re\[fig:simu la tionm e thod\]. Th e n ort h and west edg es of the t i le$t_{TL }= \ b ar p_{ (1,1) }$ inthe top lef t co rner of $ \bar p $ are la b elled b y $\#$ , a ndther e fo re , t he nod e $ x_{ T L}= {\vart he ta }^{-1 }(t_ { T L } )$ c orr espo nding to $t_{TL}$for ms a n in itial grap h $M _0 $. Th e adja cented ges of two neig hbou ring tile s $ s, t$in $\ba r p$ ar e l abe lled by the sa m e c ol o u r .Suppose $s$ is the w e s tneighbou r of $ t $, t h en $\sig ma _E( s) = \ sigma _{W} ( t) \neq \# $ andb ot htiles b el ong to t hesam e row , imp lyingthat $\s igma_ N (s)=\# \iff \s i gma_N(t) = \# $ a n d $ \ sigm a_S (s) = \# \i ff \ sigm a_S( t )= \ # $. Th erefo re , t h eir corresponding n od es in$G$ a re connectedby an $h$- e d g e $({\va rthe t a} ^ {-1}(s),{\vart heta} ^{-1}(t))\ i n E_h$.Analo gously,if $s$ is t he north ne igh bou r o f $t $, then $({\v a r thet a} ^{-1}(s ),{ \varthe ta} ^{- 1}( t)) \i n E_v$. N ext, wese eth at th e hyp e redges o ftyp eI a nd ty p e II c an be use din ord er to c r ea t e the t op row an dleftcolu m n o f the g raph $M$, re s pect iv el y. Furt hermore, thehy peredges o ftyp e IIIc a n be use d in order to create al l the re mai ningnode s of $M$. We concl ude that $ (M_0,\ set{x _{ TL} } , id)\o v e rs et{ *} {\underset { A }\r ighta rr ow}(M,O,g) $ is a derivationi n $ A$ and we wil l p rove t ha t $ ( M, O ,g) $h ast o be a final con figuration w i th $O={\empt y set }$ . Obser ve, tha t hyp e redgesof typesI and IIle avea l l t he nodes a ctive wh ile hyper e dgeso ftypeIII deact iv ate at l east t h e t op le ft nod ein the hype re dge. | rectangle_spanned by_the points $(1,1)$ and_$(h(p),w(p))$, every_$v$-edge_points downwards,_and_every $h$-edge points_rightwards; recall that_our coordinates represent the_rows and columns_of_a matrix. The derivation leading to the final configuration $(M,{\emptyset},g)$ simulates the assembly of_tiles_which form_$\bar_p$_as shown in Figure \[fig:simulationmethod\]. The_north and west edges of_the tile_$t_{TL} = \bar p_{(1,1)}$ in the top left_corner_of $\bar p$_are labelled by $\#$, and therefore, the node $x_{TL}_= {\vartheta}^{-1}(t_{TL})$ corresponding to $t_{TL}$ forms_an initial graph_$M_0$._The_adjacent edges of two_neighbouring tiles $s,t$ in $\bar p$_are labelled by the same colour._Suppose $s$ is the west neighbour of_$t$, then $\sigma_E(s) = \sigma_{W}(t) \neq_\#$ and both tiles belong_to the_same row, implying that $\sigma_N(s)=\#_\iff \sigma_N(t) =_\#$ and_$\sigma_S(s) = \#_\iff \sigma_S(t) = \#$. Therefore,_their corresponding nodes_in $G$ are connected by an_$h$-edge_$({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in E_h$. Analogously,_if_$s$_is the_north neighbour of_$t$,_then $({\vartheta}^{-1}(s),{\vartheta}^{-1}(t))\in_E_v$._Next, we see that the hyperedges_of_type I and type II can be used in_order to create the_top_row and left column_of the graph $M$, respectively._Furthermore, the hyperedges of type III can_be used_in order_to create all the remaining nodes of $M$. We conclude that_$(M_0,\set{x_{TL}},id)\overset{*}{\underset{A}\rightarrow} (M,O,g)$ is a derivation in_$A$ and we will_prove that_$(M,O,g)$_has to be_a_final configuration_with $O={\emptyset}$. Observe, that hyperedges of types I_and II leave_all the nodes active while hyperedges_of type III deactivate at_least_the top left node in the_hyperedge. |
mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu }_{\alpha }(\xi )-\tilde{K}_{\alpha }(\xi )), \label{sx}$$ where $\tilde{K}_{C}^{I}(\xi )=K_{C}^{I}$, $$\tilde{K}_{\phi }^{i}(\xi )=K_{\phi }^{i}-\xi \bar{C}^{I\hspace{0.01in}}G^{Ii}({\mkern2mu\underline{\mkern-2mu\smash{\phi }\mkern-2mu}\mkern2mu },-\overleftarrow{\partial })-(1-\xi )\bar{C}^{I\hspace{0.01in}}G^{Ii}(0,-\overleftarrow{\partial }), \label{convex}$$ while the other $\xi $-dependent tilde sources have expressions that we do not need to report here. It suffices to say that they are $K_{\phi }^{i}$ independent, such that $\delta _{r}S_{\xi }/\delta {\mkern2mu\underline{\mkern-2mu\smash{K}\mkern-2mu}\mkern2mu }_{\alpha
}=-R^{\alpha }({\mkern2mu\underline{\mkern-2mu\smash{\Phi }\mkern-2mu}\mkern2mu })$, and linear in the quantum fields $\Phi $, apart from ${\mkern2mu\underline{\mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu }_{\phi }^{i}(\xi )$, which is quadratic. Thus the action $S_{\xi }$ and the transformation $F_{\xi }$ satisfy the assumptions that allow us to apply theorem \[thb\] and corollary \[cora\]. Actually, (\[ds\]) is of type (\[fa1\]); therefore we have the differential equations (\[eqw\]) and the renormalized canonical transformation (\[fra1\]).
We want to better characterize the renormalized version $F_{R}$ of $F_{\xi }$. We know that the derivative of the renormalized $\Gamma $ functional with respect to $\xi $ is governed by the renormalized version of the average $$\left\langle \widetilde{\frac{\partial F_{\xi }}{\partial \xi }}\right\rangle =\langle \Delta \Psi \rangle +\int \mathcal{R}_{\bar{C | mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu } _ { \alpha } (\xi) -\tilde{K}_{\alpha } (\xi) ), \label{sx}$$ where $ \tilde{K}_{C}^{I}(\xi) = K_{C}^{I}$, $ $ \tilde{K}_{\phi } ^{i}(\xi) = K_{\phi } ^{i}-\xi \bar{C}^{I\hspace{0.01in}}G^{Ii}({\mkern2mu\underline{\mkern-2mu\smash{\phi } \mkern-2mu}\mkern2mu }, -\overleftarrow{\partial }) -(1-\xi) \bar{C}^{I\hspace{0.01in}}G^{Ii}(0,-\overleftarrow{\partial }), \label{convex}$$ while the other $ \xi $ -dependent tilde sources have expressions that we do not necessitate to report here. It suffice to say that they are $ K_{\phi } ^{i}$ independent, such that $ \delta _ { r}S_{\xi } /\delta { \mkern2mu\underline{\mkern-2mu\smash{K}\mkern-2mu}\mkern2mu } _ { \alpha
} = -R^{\alpha } ({ \mkern2mu\underline{\mkern-2mu\smash{\Phi } \mkern-2mu}\mkern2mu }) $, and linear in the quantum field $ \Phi $, aside from $ { \mkern2mu\underline{\mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu } _ { \phi } ^{i}(\xi) $, which is quadratic. Thus the action $ S_{\xi } $ and the transformation $ F_{\xi } $ satisfy the assumption that allow us to apply theorem \[thb\ ] and corollary \[cora\ ]. Actually, (\[ds\ ]) is of character (\[fa1\ ]); consequently we have the differential equations (\[eqw\ ]) and the renormalized canonic transformation (\[fra1\ ]).
We want to better characterize the normalize version $ F_{R}$ of $ F_{\xi } $. We know that the derived function of the renormalized $ \Gamma $ functional with respect to $ \xi $ is regulate by the renormalized version of the average $ $ \left\langle \widetilde{\frac{\partial F_{\xi } } { \partial \xi } } \right\rangle = \langle \Delta \Psi \rangle + \int \mathcal{R}_{\bar{C | mkegn-2mu\smash{\tilde{K}}\mkern-2mu}\mktrn2mu }_{\alpha }(\xi )-\tilde{K}_{\alpia }(\xi )), \label{sx}$$ where $\tilde{K}_{C}^{I}(\xi )=K_{C}^{I}$, $$\tilde{K}_{\'hi }^{u}(\xi )=K_{\kki }^{i}-\xi \bar{C}^{I\hspace{0.01in}}E^{Ii}({\mkern2ml\underlinw{\mkecn-2mu\smash{\phi }\mkech-2mu}\mkerk2iu },-\obcrlefcacrow{\partial })-(1-\xi )\nar{C}^{I\hspace{0.01hn}}G^{Ii}(0,-\overleftasruw{\'artial }), \label{convex}$$ while the othew $\xi $-delejdent tilde soorces rave vxkressions that we do not need to deport iere. It sufficex to say that they are $K_{\phl }^{i}$ lndependent, such tjat $\delta _{r}W_{\xi }/\qwlta {\mkern2mu\jnderline{\mkern-2mu\smash{K}\jkern-2mu}\mkern2mu }_{\alpha
}=-R^{\alpha }({\mkerv2mu\unberline{\mkerb-2my\smwvh{\Phi }\mkern-2nu}\mkegn2mu })$, and lincsr in dhe quamtum fields $\Phl $, apert drom ${\mkern2mu\underline{\kkern-2mu\smash{\tilde{K}}\ikern-2mu}\mkaru2mu }_{\phi }^{i}(\xi )$, which is qyadrajic. Tvus gye xctjoi $S_{\si }$ anf tie transforjation $F_{\xi }$ satisfy the assumpuiogw that allow hs to wp[ly theorem \[thb\] and corollary \[cora\]. Actudllg, (\[ds\]) is of type (\[fa1\]); therwfore we have the difverential equations (\[eqw\]) and the renormalized canonical tratsforjxtiin (\[wea1\]).
Ae want to better characterize the renormalizqs fegsion $F_{R}$ of $F_{\xi }$. Ce know that the dfrodative of the renormaljzed $\Gamma $ functilnal wijh respect to $\xy $ ix governed by the renormalized version jd the average $$\left\pangle \widecilde{\ftac{\paryial F_{\xi }}{\partial \xi }}\rigkt\rangme =\langle \Dflta \Psi \dxngle +\int \mathcau{R}_{\bsr{W | mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu }_{\alpha }(\xi )-\tilde{K}_{\alpha }(\xi )), \label{sx}$$ )=K_{C}^{I}$, }^{i}(\xi )=K_{\phi \bar{C}^{I\hspace{0.01in}}G^{Ii}({\mkern2mu\underline{\mkern-2mu\smash{\phi }\mkern-2mu}\mkern2mu },-\overleftarrow{\partial the $\xi $-dependent tilde have expressions that do not need to report here. suffices to say that they are $K_{\phi }^{i}$ independent, such that $\delta _{r}S_{\xi {\mkern2mu\underline{\mkern-2mu\smash{K}\mkern-2mu}\mkern2mu }_{\alpha }=-R^{\alpha }({\mkern2mu\underline{\mkern-2mu\smash{\Phi }\mkern-2mu}\mkern2mu })$, and linear in the quantum fields $\Phi apart ${\mkern2mu\underline{\mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu }^{i}(\xi which is quadratic. Thus the action $S_{\xi }$ and the transformation $F_{\xi }$ satisfy the assumptions allow us to apply theorem \[thb\] and corollary Actually, (\[ds\]) is of (\[fa1\]); therefore we have the equations and the canonical (\[fra1\]). want to better the renormalized version $F_{R}$ of $F_{\xi }$. We know that the derivative of the renormalized $\Gamma $ with respect $ is by renormalized of the average F_{\xi }}{\partial \xi }}\right\rangle =\langle \Delta \mathcal{R}_{\bar{C | mkern-2mu\smash{\tilde{K}}\mkern-2mu}\Mkern2mu }_{\alpHa }(\xi )-\tIldE{K}_{\aLpHa }(\xi )), \LabeL{sx}$$ where $\tilde{K}_{c}^{i}(\xi )=K_{c}^{I}$, $$\tilde{K}_{\phi }^{i}(\xi )=K_{\phi }^{i}-\xi \bAr{C}^{I\hSpACe{0.01in}}g^{ii}({\Mkern2Mu\underLInE{\MKerN-2mU\sMasH{\pHI }\mKern-2mU}\mkErn2mu },-\ovErleftarroW{\paRtIal })-(1-\xi )\bar{C}^{I\hsPAcE{0.01in}}G^{Ii}(0,-\overLefTarrow{\partiaL }), \laBel{conVeX}$$ whILe the OthEr $\xi $-dEpendeNT tilde Sources haVe EXpressIOns that WE Do Not nEed to report here. It SUfFIces to say that tHey are $k_{\pHI }^{i}$ INDepEndEnt, such thaT $\dElta _{r}s_{\Xi }/\delta {\MKeRN2MU\unDErline{\mkern-2mu\Smash{K}\mkern-2MU}\mkErn2mu }_{\aLpHa
}=-R^{\ALpha }({\mkErn2mu\UnDErlIne{\mkern-2mu\sMash{\phi }\mkern-2mU}\mkern2MU })$, and linEAr in the QuantuM fiEldS $\Phi $, APaRt FroM ${\mKErn2MU\uNdeRLinE{\mkern-2mu\SmAsH{\tildE{K}}\mkERN-2MU}\mkeRn2mU }_{\phi }^{I}(\xi )$, whIch is quadratiC. ThUs thE ActIon $S_{\xI }$ and tHe trAnSformAtion $F_{\Xi }$ satIsFy the assumptionS thaT allow us tO apPlY thEoRem \[thB\] And corOllAry \[Cora\]. ActUally, (\[ds\]) IS of TyPE (\[FA1\]); tHerefore we have the dIfFEReNtial equAtions (\[EQw\]) AnD The renorMaLizEd caNONical TranSFoRmation (\[fRa1\]).
We waNT tO bEtter chArActeriZe The RenOrmalIZed vErsion $f_{R}$ of $F_{\xi }$. WE know THat the derivatiVE of the renormaLIzED $\gaMMa $ fuNctIonal with reSpecT To $\xi $ Is goVErNed BY the rEnormAlIZeD Version of the average $$\LeFt\langLe \widEtilde{\frac{\parTial F_{\xi }}{\parTIAL \xi }}\right\RangLE =\lANgle \Delta \Psi \raNgle +\iNt \mathcal{R}_{\BAr{C | mkern-2mu\smash{\tilde{K}} \mkern-2mu }\mke rn2 mu}_ {\al pha}(\xi )-\tilde { K}_{ \alpha }(\xi )), \lab el{sx }$ $ whe r e$\til de{K}_{ C }^ { I }(\ xi ) =K_ {C } ^{ I}$,$$\ tilde{K }_{\phi }^ {i} (\ xi )=K_{\phi }^ {i}-\xi \b ar{ C}^{I\hspace {0. 01in}} G^ {Ii } ({\mk ern 2mu\u nderli n e{\mke rn-2mu\sm as h {\phi} \mkern- 2 m u} \mke rn2mu },-\overlef t ar r ow{\partial }) -(1-\x i) \b a r {C} ^{I \hspace{0. 01 in}}G ^ {Ii}(0, - \o v e r lef t arrow{\partia l }), \lab e l{c onvex} $$ wh i le the othe r$ \xi $-dependen t ti lde sourc es hav e expres s ions th at wedonot nee d t orep or t he r e. It suf fices to s ay that the y a r e $K _{\ phi}^{i} $ independent , s ucht hat $\de lta _ {r}S _{ \xi } /\delt a {\m ke rn2mu\underline {\mk ern-2mu\s mas h{ K}\ mk ern-2 m u}\mke rn2 mu}_{\alp ha
}=-R ^ {\a lp h a }( {\mkern2mu\underli ne { \ mk ern-2mu\ smash{ \ Ph i} \mkern-2 mu }\m kern 2 m u })$ , an d l inear in the q u an tu m field s$\Phi$, ap art from ${\m kern2m u\underl ine{\ m kern-2mu\smash { \tilde{K}}\mk e rn - 2 mu } \mke rn2 mu }_{\phi}^{i } (\xi )$, wh ich is qu adrat ic . T h us the action $S_{\ xi }$ an d the transformati on $F_{\xi } $ satisfy the as s umptions thatallow us to app l y theore m \[t hb\] and corollar y \[cora\] . A ctu all y,( \ [d s\]) is of ty p e (\[ fa 1\]); t her efore w e h ave th e d if ferential equatio ns ( \[ eq w\] ) and the reno rm ali ze d c anoni c al tra nsfor mati on ( \ [fr a1\]).We w antto b ette r c ha racte rize the renorm alized ve rsi o n $F _{ R} $ of $F _{\xi }$. Wekn ow that th eder ivativ e of the r enormalized $\Gamma $ f u nctiona l w ith r espe ct to $\x i $ is go ver n ed bythe re norma li zed v ersio n of th eaverage $$ \ l eft \lang le \wi detilde {\frac{\partial F_ { \xi }}{\partial\xi }}\ r i gh t\r a ng l e = \l a ngl e \Delta \Psi \ra ngle +\int \ m at hcal{R}_{\ b ar{ C | mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu }_{\alpha_}(\xi )-\tilde{K}_{\alpha_}(\xi )), \label{sx}$$_where $\tilde{K}_{C}^{I}(\xi_)=K_{C}^{I}$,_$$\tilde{K}_{\phi }^{i}(\xi_)=K_{\phi_}^{i}-\xi \bar{C}^{I\hspace{0.01in}}G^{Ii}({\mkern2mu\underline{\mkern-2mu\smash{\phi }\mkern-2mu}\mkern2mu_},-\overleftarrow{\partial })-(1-\xi )\bar{C}^{I\hspace{0.01in}}G^{Ii}(0,-\overleftarrow{\partial_}), \label{convex}$$ while_the other $\xi_$-dependent_tilde sources have expressions that we do not need to report here. It suffices_to_say that_they_are_$K_{\phi }^{i}$ independent, such that_$\delta _{r}S_{\xi }/\delta {\mkern2mu\underline{\mkern-2mu\smash{K}\mkern-2mu}\mkern2mu }_{\alpha
}=-R^{\alpha_}({\mkern2mu\underline{\mkern-2mu\smash{\Phi }\mkern-2mu}\mkern2mu_})$, and linear in the quantum fields $\Phi_$,_apart from ${\mkern2mu\underline{\mkern-2mu\smash{\tilde{K}}\mkern-2mu}\mkern2mu_}_{\phi }^{i}(\xi )$, which is quadratic. Thus the action_$S_{\xi }$ and the transformation $F_{\xi_}$ satisfy the_assumptions_that_allow us to apply_theorem \[thb\] and corollary \[cora\]. Actually,_(\[ds\]) is of type (\[fa1\]); therefore_we have the differential equations (\[eqw\]) and_the renormalized canonical transformation (\[fra1\]).
We want_to better characterize the renormalized_version $F_{R}$_of $F_{\xi }$. We know_that the derivative_of the_renormalized $\Gamma $_functional with respect to $\xi $_is governed by_the renormalized version of the average_$$\left\langle_\widetilde{\frac{\partial F_{\xi }}{\partial_\xi_}}\right\rangle_=\langle \Delta_\Psi \rangle +\int_\mathcal{R}_{\bar{C |
Trapezium members (Component A and Component C, respectively; see Figure \[Trap1m\]). This explains why their basic observational parameters were (and some, still are) poorly known.
$\theta^1$OriE was discovered to be a a double-lined spectroscopic binary by Costero et al. (2006). Its nearly identical members are pre-MS stars with approximate spectral type G2IV; they are in circular orbit with $P=9.8952\,d$ and $K_1=K_2=84.4\,km\,s^{-1}$ [@Costero+08]. This binary is a variable radio [e.g. see @Felli+93] and X-ray source source ([*COUP*]{} 732); it also varies by several tenths of magnitud in the visual range [@Wolf94] and by hundredths of magnitud in the $3,5\,\mu$ and $4.5\,\mu$ bands [@Morales+12]. However, the scarcity of published absolute photometry data on this star is amazing, except for its K-magnitude for which five measures are found in the literature, averaging about 6.8$\,mag$. In a [*Spitzer*]{} survey dedicated to the search and characterization of variability of the ONC stars, [@Morales+12] found that Component E is a grazing eclipsing binary and derived the mass of its members to be $M_1=M_2=2.80\,M_{\odot}$ with 2% accuracy. Together with its measured high- precision proper motion [@Dzib+17] and parallax [@Menten+07; @Kounkel+17], these results make of $\theta^1$OriE the highest mass pre-MS star with well known physical parameters.
Component F has received very little attention and practically nothing was known about it. [@Herbig50], in the last sentence of his early paper on the spectroscopy of variable stars in the ONC, just before the acknowledgements and after discussing the spectral type of Component E, writes: “[*Star F can be classified, with considerably more confidence as of type B8*]{}”. This short entry was noticed and registered by [@Parenago54] in his catalogue of stars in the ONC, perpetuating the only spectral classification I know of, published for this star. We have obtained Échelle spectra of $\theta^1$OriF | Trapezium members (Component A and Component C, respectively; see Figure \[Trap1m\ ]). This explains why their basic experimental argument were (and some, still are) poorly know.
$ \theta^1$OriE was learn to be a a double - lined spectroscopic binary by Costero et al. (2006). Its closely identical member are pre - MS stars with approximate apparitional character G2IV; they are in circular orbit with $ P=9.8952\,d$ and $ K_1 = K_2=84.4\,km\,s^{-1}$ [ @Costero+08 ]. This binary star is a variable radio [ for example see @Felli+93 ] and X - ray source source ([ * COUP * ] { } 732); it also varies by respective tenths of magnitud in the visual range [ @Wolf94 ] and by one-hundredth of magnitud in the $ 3,5\,\mu$ and $ 4.5\,\mu$ bands [ @Morales+12 ]. However, the scarcity of published absolute photometry datum on this star is amazing, except for its kelvin - magnitude for which five measures are found in the literature, averaging about 6.8$\,mag$. In a [ * Spitzer * ] { } sketch dedicated to the search and characterization of variability of the ONC stars, [ @Morales+12 ] found that Component E is a grazing eclipsing binary and derived the mass of its members to be $ M_1 = M_2=2.80\,M_{\odot}$ with 2% accuracy. Together with its measured high- precision proper motion [ @Dzib+17 ] and parallax [ @Menten+07; @Kounkel+17 ], these results make of $ \theta^1$OriE the highest aggregate pre - MS ace with well known forcible argument.
Component F has receive very little attention and much nothing was known about it. [ @Herbig50 ], in the final sentence of his early paper on the spectroscopy of varying stars in the ONC, just before the acknowledgements and after discussing the spectral type of Component E, writes: “ [ * Star F can be classified, with considerably more confidence as of type B8 * ] { } ”. This short entrance was noticed and register by [ @Parenago54 ] in his catalog of stars in the ONC, perpetuating the only apparitional classification I know of, published for this star. We have obtained Échelle spectra of $ \theta^1$OriF | Trwpezium members (Componenu A and Component C, respxctivelg; see Fieure \[Trap1m\]). This explains why vheie basuc observational paramdters werv (and somw, stmll are) poorly kikwn.
$\theta^1$OriE sws dnsrovered to be a a double-lhned spectroscmpkc binary by Costero et al. (2006). Its nearlr identocwl members are pre-KF stzgs with approximate spectral tyle G2IV; uhey are in circulsr orbit with $P=9.8952\,d$ and $K_1=K_2=84.4\,km\,s^{-1}$ [@Coshero+08]. This binary id a variablg raquo [e.g. see @Feuli+93] and X-ray source soorce ([*COUP*]{} 732); it also varies by sevdral cenths of mqgbitkg in the viwual gange [@Wolf94] and by hungredths of magnitud ik the $3,5\,\mu$ and $4.5\,\mu$ bands [@Morales+12]. However, the scarcyty of putlnshed absolute photomwtey daja on thir stxr js ajazing, exrept for ita K-magnitudw for which five messtgrs are found in thq jiterature, averaging about 6.8$\,mag$. In a [*Spiuzer*]{} aurvey dedicated to the search and characterlzation os variability of the ONC stars, [@Morales+12] found that Compkvenu C kw w grazing eclipsing binary and derived the mafa pf its members tj be $M_1=M_2=2.80\,M_{\odoy}$ aiyr 2% accuracy. Tugethex wjth its measured hlgh- presisiob proper iotipn [@Dzib+17] and parallax [@Menten+07; @Kounkel+17], thefw results make of $\cheta^1$OriE thz highgst maxs pre-MS star with well knosn physical parametedr.
Component F has rebeivad very little attention agd practirally nothine wax knowg about it. [@Herblc50], in the last sentfnce lf his early paper on the spectroscopy of veciable stars on thv ONC, jusc befove the acknowleqgements and ayter diseussine the specnral type of Componegt E, writes: “[*Sdwr F can be rlassifieq, wirh cinsiderxcly more conficence as of type B8*]{}”. Rhis short entry wes nuficed and regisctree by [@Parenago54] im hks sanalpgta of stars it thd OVV, peroetuating tme unly spectral classificadion I know of, publishrd for thiw star. Wq have obtainrd Échelle spectra lf $\thxta^1$OriH | Trapezium members (Component A and Component C, Figure This explains their basic observational are) known. $\theta^1$OriE was to be a double-lined spectroscopic binary by Costero et (2006). Its nearly identical members are pre-MS stars with approximate spectral type G2IV; are in circular orbit with $P=9.8952\,d$ and $K_1=K_2=84.4\,km\,s^{-1}$ [@Costero+08]. This binary is a radio see and source source ([*COUP*]{} 732); it also varies by several tenths of magnitud in the visual range and by hundredths of magnitud in the $3,5\,\mu$ $4.5\,\mu$ bands [@Morales+12]. However, scarcity of published absolute photometry on star is except its for which five are found in the literature, averaging about 6.8$\,mag$. In a [*Spitzer*]{} survey dedicated to the search and of variability ONC stars, found Component is a grazing and derived the mass of its $M_1=M_2=2.80\,M_{\odot}$ with 2% accuracy. Together with its measured precision proper [@Dzib+17] and parallax [@Menten+07; @Kounkel+17], these make of $\theta^1$OriE the highest mass pre-MS star well known physical parameters. Component F has received very little attention and practically nothing was it. [@Herbig50], in the sentence of his paper the of stars in ONC, just before the acknowledgements and after discussing the spectral type Component E, writes: “[*Star F can be classified, with considerably as type B8*]{}”. This entry was noticed and by in his catalogue of the perpetuating classification know published for this star. have obtained Échelle spectra of | Trapezium members (Component a and ComponEnt C, rEspEctIvEly; sEe FiGure \[Trap1m\]). This eXPlaiNs why their basic observaTionaL pARameTErS were (And some, STiLL Are) PoOrLy kNoWN.
$\tHeta^1$ORiE Was discOvered to be A a dOuBle-lined specTRoScopic binaRy bY Costero et al. (2006). its Nearly IdEntICal meMbeRs are Pre-MS sTArs witH approximAtE SpectrAL type G2Iv; THeY are In circular orbit wiTH $P=9.8952\,D$ And $K_1=K_2=84.4\,km\,s^{-1}$ [@CosterO+08]. This bInARy IS A vaRiaBle radio [e.g. SeE @FellI+93] And X-ray SOuRCE SouRCe ([*COUP*]{} 732); it also vAries by seveRAl tEnths oF mAgnITud in tHe visUaL RanGe [@Wolf94] and by HundRedths of mAgnituD In the $3,5\,\mu$ ANd $4.5\,\mu$ banDs [@MoraLes+12]. howEver, THe ScArcItY Of pUBlIshED abSolute phOtOmEtry dAta oN THIS staR is AmazIng, exCept for its K-maGniTude FOr wHich fIve meAsurEs Are foUnd in tHe litErAture, averaging aBout 6.8$\,Mag$. In a [*SpiTzeR*]{} sUrvEy DedicATed to tHe sEarCh and chAracterIZatIoN OF VaRiability of the ONC sTaRS, [@moRales+12] fouNd that cOmPoNEnt E is a gRaZinG eclIPSing bInarY AnD derived The masS Of ItS memberS tO be $M_1=M_2=2.80\,M_{\OdOt}$ wIth 2% AccurACy. ToGether With its mEasurED high- precision PRoper motion [@DzIB+17] aND PaRAllaX [@MeNten+07; @Kounkel+17], ThesE ResuLts mAKe Of $\tHEta^1$OrIE the HiGHeST mass pre-MS star with wElL known PhysiCal parameters.
component F HAS Received Very LItTLe attention and PractIcally nothINg was knoWn aboUt it. [@HerbIg50], in the laST Sentence Of hIs eArlY paPER oN the spectroscOPY of vArIable stArs In the ONc, juSt bEfoRe tHe AcknowledGements aNd AfTeR dIscUssinG The spectRaL tyPe Of COmponENt E, wriTes: “[*StAr F cAn Be CLasSified, wITh CONsidErAbLy moRe cOnFidenCe as OF tyPe B8*]{}”. This Short entrY waS NotiCeD aNd regisTered by [@ParenaGo54] In his catalOgUe oF stars IN The ONC, peRpetuating the only spectrAL classiFicAtion i knoW of, publisHed For thiS stAR. We havE obtaiNed ÉcHeLle SPEctra OF $\ThEta^1$orIF | Trapezium members (Compon ent A andCompo nen t C ,resp ecti vely; see Figu r e \[ Trap1m\]). This explai ns wh yt heir ba sic o bservat i on a l pa ra me ter sw er e (an d s ome, st ill are) p oor ly known.
$\t h et a^1$OriE w asdiscovered t o b e a ado ubl e -line d s pectr oscopi c binar y by Cost er o et al . (2006) . It s ne arly identical me m be r s are pre-MS s tars w it h a p p rox ima te spectra ltypeG 2IV; th e ya r e in circular orbi t with $P=9 . 895 2\,d$an d $ K _1=K_2 =84.4 \, k m\, s^{-1}$ [@C oste ro+08]. T his bi n ary isa variab le rad io[e. g. s e e@F ell i+ 9 3]a nd X- r aysource s ou rc e ([* COUP * ] { } 732 );it a lso v aries by seve ral ten t hsof ma gnitu d in t he vi sual r ange[@ Wolf94] and byhund redths of ma gn itu din th e $3,5\ ,\m u$and $4. 5\,\mu$ ban ds [ @ Mo rales+12]. However ,t h escarcity of pu b li sh e d absolu te ph otom e t ry da ta o n t his star is am a zi ng , excep tfor it sK-m agn itude forwhichfive mea sures are found in t h e literature, av e r ag i ng a bou t 6.8$\,mag $. I n a [ *Spi t ze r*] { } sur vey d ed i ca t ed to the search an dcharac teriz ation of vari ability of t h e ONC st ars, [@ M orales+12] fou nd th at Compone n t E is a graz ing ecli psing bin a r y and de riv edthe ma s s o f its members t o be $ M_1=M_2 =2. 80\,M_{ \od ot} $ w ith 2 % accurac y. Toget he rwi th it s mea s ured hig h- pr ec isi on pr o per mo tion[@Dz ib +1 7 ] a nd para l la x [@Me nt en +07; @K ou nkel+ 17], the se resu lts makeof$ \the ta ^1 $OriE t he highest ma ss pre-MS st ar wi th wel l known ph ysical parameters.
Com p onent F ha s rec eive d very li ttl e atte nti o n andpracti cally n oth i n g was k no wnab out it. [@ H e rbi g50], i n th e lastsentence of his ea r lypaper on thespe ctro s c op y o f v a ria bl e st a r s in the ONC, j ust before t h eacknowledg e men ts and af ter dis cussi n g the s pectral t ype of Co mp onen t E,writes: “[ *Star Fcan be cl a ssifi e d, with co nsider ab lymoreconfid e nce as o f type B 8*]{}” . Thi sshort en try was noticed and reg istere d by[@P arenago54 ] i n hi s catalog ue o f stars in th e O NC, p erp e tuati ng t h eonl y spec tral classific a ti onI kn ow of, publ i s h edfor t his star.We h ave obtained Éche l le spectra of$\th e t a^1 $Or i F | Trapezium_members (Component_A and Component C,_respectively; see_Figure \[Trap1m\])._This explains_why_their basic observational_parameters were (and_some, still are) poorly_known.
$\theta^1$OriE was discovered_to_be a a double-lined spectroscopic binary by Costero et al. (2006). Its nearly identical_members_are pre-MS_stars_with_approximate spectral type G2IV; they_are in circular orbit with_$P=9.8952\,d$ and_$K_1=K_2=84.4\,km\,s^{-1}$ [@Costero+08]. This binary is a variable radio_[e.g._see @Felli+93] and_X-ray source source ([*COUP*]{} 732); it also varies by_several tenths of magnitud in the_visual range [@Wolf94]_and_by_hundredths of magnitud in_the $3,5\,\mu$ and $4.5\,\mu$ bands [@Morales+12]._However, the scarcity of published absolute_photometry data on this star is amazing,_except for its K-magnitude for which_five measures are found in_the literature,_averaging about 6.8$\,mag$. In a_[*Spitzer*]{} survey dedicated_to the_search and characterization_of variability of the ONC stars,_[@Morales+12] found that_Component E is a grazing eclipsing_binary_and derived the_mass_of_its members_to be $M_1=M_2=2.80\,M_{\odot}$_with_2% accuracy._Together_with its measured high- precision proper_motion_[@Dzib+17] and parallax [@Menten+07; @Kounkel+17], these results_make of $\theta^1$OriE the_highest_mass pre-MS star with_well known physical parameters.
Component F_has received very little attention and_practically nothing_was known_about it. [@Herbig50], in the last sentence of his early paper_on the spectroscopy of variable stars_in the ONC, just_before the_acknowledgements_and after discussing_the_spectral type_of Component E, writes: “[*Star F can_be classified,_with considerably more confidence as of_type B8*]{}”. This short_entry_was noticed and registered by [@Parenago54]_in his catalogue of stars in_the ONC, perpetuating the only_spectral_classification_I know of, published for_this star. We have obtained Échelle_spectra of $\theta^1$OriF |
studied in [@G; @Pat; @US]. It has been shown for generic $U(1)^2$ that in the process of the black hole evaporation the temperature increases, reaches the maximum and rapidly drops to zero when the mass of the black hole reaches the value of the central charge. Specific heat blows up when the temperature reaches the maximum and changes the sign, see Figs. 4, 6, 7 of [@US]. The change of the sign of the heat capacity happens at the non-vanishing temperature and means that the corresponding component of the Weinhold metric undergoes the same type of changes.
The second important point is that, as pointed out in [@GKK], when considering black holes with scalars we must augment the usual extensive thermodynamic variables such as the area $A$ and the conserved charges $(q,p)$ with the values of the moduli at infinity $\phi_\infty$.
This has the consequence that the thermodynamic configuration space is no longer flat ${ \bf R}^k \equiv (A, q,p)$ ( where $k$ is one plus two times the number of electric charges, but becomes its product with the scalar moduli space ${ \bf R}^k \times {\cal M}_\phi$.
It was shown in [@GKK] that the thermodynamic variables conjugate to to the moduli $\phi^a_\infty$, i.e. the analogue of the chemical potentials $\mu_a$ are minus the scalar charges, i.e. one has the relation $$dM=TdS+\psi^A dq_A + \chi_A dp^A - \Sigma _a d \phi ^a _\infty \.$$
The fact that the scalar moduli space ${\cal M}_\phi$ is no longer flat complicates, but of course does not invalidate, the usual thermodynamic formalism involving Legendre transformations (technically one should speak of Legendre submanifolds etc.) and in particular one is allowed to extend the definition of the Weinhold and Ruppeiner metrics to the scalar moduli space ${\cal M}_\phi$ provided we replace the ordinary derivative by the covariant derivative with respect to the metric $G_{ab}$ on the moduli space.
In the present paper we have been considering extreme black holes for which the temperature $T=0$ and it is the Weinhold metric $W_{ab}$ which seems to be the more appropriate one to consider. This may be defined by $$W_{ab}= | studied in [ @G; @Pat; @US ]. It has been shown for generic $ U(1)^2 $ that in the process of the black fix vaporization the temperature increases, reaches the maximum and quickly drops to zero when the mass of the black fix reaches the value of the central care. Specific heat blows up when the temperature reach the maximum and changes the sign, see Figs. 4, 6, 7 of [ @US ]. The change of the sign of the hotness capacity happens at the non - vanishing temperature and means that the comparable component of the Weinhold metric undergoes the like type of change.
The second important point is that, as pointed out in [ @GKK ], when considering black fix with scalars we must augment the usual extensive thermodynamic variables such as the area $ A$ and the conserved charges $ (q, p)$ with the values of the moduli at infinity $ \phi_\infty$.
This has the consequence that the thermodynamic configuration space is no longer flat $ { \bf R}^k \equiv (A, q, p)$ (where $ k$ is one plus two times the number of electric charges, but become its product with the scalar moduli quad $ { \bf R}^k \times { \cal M}_\phi$.
It was shown in [ @GKK ] that the thermodynamic variables conjugate to to the moduli $ \phi^a_\infty$, i.e. the analogue of the chemical potentials $ \mu_a$ are minus the scalar charge, i.e. one has the relation $ $ dM = TdS+\psi^A dq_A + \chi_A dp^A - \Sigma _ a d \phi ^a _ \infty \.$$
The fact that the scalar moduli quad $ { \cal M}_\phi$ is no longer flat complicates, but of course does not invalidate, the usual thermodynamic formalism involving Legendre transformations (technically one should talk of Legendre submanifolds etc .) and in particular one is allowed to extend the definition of the Weinhold and Ruppeiner metrics to the scalar modulus space $ { \cal M}_\phi$ leave we replace the ordinary derivative by the covariant derivative with respect to the measured $ G_{ab}$ on the moduli space.
In the present paper we have been considering extreme black holes for which the temperature $ T=0 $ and it is the Weinhold metric $ W_{ab}$ which seems to be the more appropriate one to consider. This may be define by $ $ W_{ab}= | stkdied in [@G; @Pat; @US]. It has been shown for generir $U(1)^2$ thaf in the process of the black hole etapoeatiob the temperature incrdases, reabhes the naxinym and rapmsly droif to dero chxn the mass of jhe black hone reaches the vxlbe of the central charge. Specific hewt blowx kp when the teiperseure gecches the maximum and changes tge sign, see Figs. 4, 6, 7 pf [@US]. The change of the sihn ov the heat capacitj happens aj thq non-vanishine temperatlxe and meana that the corresponding compondnt oy the Weinhile mfjric undergoxs the same type on changas.
The srcond importanb poiit iw that, as pointed out in [@GKK], when consiqering bldci holes with scalqrw we kust augownt tht uxuzl extfnsmve thermodgnamic variqbles such as the atew $A$ and the cohserveq sharges $(q,p)$ with the values of the modulp at infinity $\phi_\infty$.
This yas the consequence tjat the trermodynamic configuration space is no longer flad ${ \bf F}^k \tqmlv (A, w,p)$ ( where $k$ is one plus two times the number of sltctgic charges, but bccomes its product wltn the scalar muduli spade ${ \bf R}^k \times {\cap M}_\phi$.
Ij was whown in [@DKK] yhat the thermodynamic variqbles conjugcte to to the moduli $\'hi^a_\infty$, i.e. the analpgue of the chemical pocentiams $\mu_a$ are linus the rcalar charges, i.d. oke vas the relation $$dM=TdS+\psi^A dq_A + \chi_E dp^A - \Sigma _a d \phi ^a _\infty \.$$
The fact that the scalar mofuli dpdce ${\cal M}_\pji$ is no longer flat complicates, but of coursg dmes not invclidatc, the usual thewmodynamic forkalism nnvolvkng Legendge transfmrmations (tqchnically ona should spean of Legqndrw suvmanifouas etc.) and in larticulag jne us allowed to extekd thg sefinition of tkt Wtibhold and Rupprindr ienrirs to dhe scalar mmdulk sosce ${\cxl M}_\phi$ proyiddd wr replace the ordinasy dsrivative by the cpvwriant dgrivative with respect to the metric $G_{ab}$ on tie modnli spsce.
Yn the present paper we have bsen consiferlng extreme bjack holes for wkich the temperature $T=0$ and it is the Weiihold metric $W_{ab}$ which weems to be the morg akpropriate onx to cjnsider. Tvis may be defined bt $$W_{ab}= | studied in [@G; @Pat; @US]. It has for $U(1)^2$ that the process of temperature reaches the maximum rapidly drops to when the mass of the black reaches the value of the central charge. Specific heat blows up when the reaches the maximum and changes the sign, see Figs. 4, 6, 7 of The of sign the heat capacity happens at the non-vanishing temperature and means that the corresponding component of the metric undergoes the same type of changes. The important point is that, pointed out in [@GKK], when black with scalars must the extensive thermodynamic variables as the area $A$ and the conserved charges $(q,p)$ with the values of the moduli at infinity This has that the configuration is longer flat ${ \equiv (A, q,p)$ ( where $k$ two times the number of electric charges, but its product the scalar moduli space ${ \bf \times {\cal M}_\phi$. It was shown in [@GKK] the thermodynamic variables conjugate to to the moduli $\phi^a_\infty$, i.e. the analogue of the chemical are minus the scalar i.e. one has relation dq_A \chi_A - \Sigma d \phi ^a _\infty \.$$ The fact that the scalar moduli ${\cal M}_\phi$ is no longer flat complicates, but of course invalidate, usual thermodynamic formalism Legendre transformations (technically one speak Legendre submanifolds etc.) and one allowed definition the and Ruppeiner metrics to scalar moduli space ${\cal M}_\phi$ we replace the ordinary with respect to the metric $G_{ab}$ on the space. In the present paper we have considering extreme black holes for which the temperature $T=0$ and it is Weinhold metric seems to be the more appropriate one to This may be defined $$W_{ab}= | studied in [@G; @Pat; @US]. It has been sHown for genEric $U(1)^2$ ThaT in ThE proCess Of the black hole EVapoRation the temperature inCreasEs, REachES tHe maxImum and RApIDLy dRoPs To zErO WhEn the MasS of the bLack hole reAchEs The value of thE CeNtral chargE. SpEcific heat blOws Up when ThE teMPeratUre ReachEs the mAXimum aNd changes ThE Sign, seE figs. 4, 6, 7 of [@Us]. tHe ChanGe of the sign of the hEAt CApacity happens At the nOn-VAnISHinG teMperature aNd Means THat the cORrESPOndINg component of The Weinhold MEtrIc undeRgOes THe same Type oF cHAngEs.
The second ImpoRtant poinT is thaT, As pointED out in [@GkK], when ConSidErinG BlAcK hoLeS WitH ScAlaRS we Must augmEnT tHe usuAl exTENSIve tHerModyNamic Variables such As tHe arEA $A$ aNd the ConseRved ChArges $(Q,p)$ with The vaLuEs of the moduli at InfiNity $\phi_\inFty$.
thIs hAs The coNSequenCe tHat The therModynamIC coNfIGURaTion space is no longeR fLAT ${ \bF R}^k \equiv (a, q,p)$ ( wheRE $k$ Is ONe plus twO tImeS the NUMber oF eleCTrIc chargeS, but beCOmEs Its prodUcT with tHe ScaLar ModulI SpacE ${ \bf R}^k \tImes {\cal M}_\Phi$.
It WAs shown in [@GKK] thAT the thermodynAMiC VArIAbleS coNjugate to to The mODuli $\Phi^a_\INfTy$, i.E. The anAloguE oF ThE Chemical potentials $\mU_a$ Are minUs the Scalar charges, I.e. one has thE RELation $$dM=tdS+\pSI^A DQ_A + \chi_A dp^A - \Sigma _A d \phi ^A _\infty \.$$
The fACt that thE scalAr moduli Space ${\cal M}_\PHI$ is no lonGer FlaT coMplICAtEs, but of course DOEs noT iNvalidaTe, tHe usual TheRmoDynAmiC fOrmalism iNvolving leGeNdRe TraNsforMAtions (teChNicAlLy oNe shoULd speaK of LeGendRe SuBManIfolds eTC.) aND In paRtIcUlar One Is AllowEd to EXteNd the deFinition oF thE weinHoLd And RuppEiner metrics tO tHe scalar moDuLi sPace ${\caL m}_\Phi$ proviDed we replace the ordinary DErivatiVe bY the cOvarIant derivAtiVe with ResPEct to tHe metrIc $G_{ab}$ On The MODuli sPACe.
in tHe Present papER We hAve beEn ConsIdering Extreme black holes fOR whIch the temperaTurE $T=0$ anD IT iS thE weINhoLd MEtrIC $w_{ab}$ which seems to Be the more aPpROpRiate one to COnsIdEr. This mAy be defIned bY $$w_{ab}= | studied in [@G; @Pat; @US ]. It hasbeensho wnfo r ge neri c $U(1)^2$ tha t inthe process of the bla ck ho le evap o ra tionthe tem p er a t ure i nc rea se s ,reach esthe max imum and r api dl y drops to z e ro when themas s of the bla ckhole r ea che s theval ue of the c e ntralcharge. S pe c ific h e at blow s up whe n the temperature re a ches the maxim um and c h an g e s t hesign, seeFi gs. 4 , 6, 7 o f [ @ U S ].T he change ofthe sign of the heatca pac i ty hap pensat the non-vanish ingtemperatu re and means t h at thecorres pon din g co m po ne ntof the We inh o ldmetric u nd er goesthes a m e typ e o f ch anges .
The second im port a ntpoint is t hat, a s poi nted o ut in [ @GKK], when con side ring blac k h ol eswi th sc a lars w e m ust augmen t the u s ual e x t e ns ive thermodynamicva r i ab les such as th e a re a $A$ and t hecons e r ved c harg e s$(q,p)$with t h eva lues of t he mod ul i a t i nfini t y $\ phi_\i nfty$.
Thish as the consequ e nce that thet he r m od y nami c c onfiguratio n sp a ce i s no lo nge r flat ${ \ bf R} ^ k \equiv (A, q,p)$(where$k$ i s one plus tw o times th e n umber of ele c tr i c charges, but beco mes its pr o duct wit h the scalarmoduli sp a c e ${ \bf R} ^k\ti mes { \c al M}_\phi$.I t wa sshown i n [ @GKK] t hat th e t her mo dynamic v ariables c on ju ga teto to the modu li $\ ph i^a _\inf t y$, i. e. th e an al og u e o f the c h em i c al p ot en tial s $ \m u_a$arem inu s the s calar cha rge s , i. e. o ne hasthe relation$$ dM=TdS+\ps i^ A d q_A +\ c hi_A dp^ A - \Sigma _a d \phi ^a _\infty \. $$
T he f act thatthe scala r m o duli s pace $ {\cal M }_\ p h i$ is n olon ge r flat com p l ica tes,bu t of course does not invalida t e,the usual the rmo dyna m i cfor m al i smin v olv i n g Legendre tran sformation s( te chnicallyo nesh ould sp eak ofLegen d re subm anifoldsetc.) and i n pa r t icu lar one is allowed to exten d thed ef initi onof the W ein holdand Ru p pei ner m etrics t o thescala rmoduli s pace ${\cal M}_\phi$ pr ovided we r epl ace the o rdi n ary derivati ve b y the cova ria ntderiv ati v e wit h re s pe ctt o the met r ic $G_{ab } $ont h emoduli spac e .
In thepre s ent pa perwe have been cons i dering extreme bla c k ho les forwh ich the temper atu re $ T=0$ and i t is the We inhold m et r ic $W _{ab}$ which seemst o b e the m oreapp ropriateone t o consid er .T his ma y be d efined by $$ W _{ab } = | studied_in [@G;_@Pat; @US]. It has_been shown_for_generic $U(1)^2$_that_in the process_of the black_hole evaporation the temperature_increases, reaches the_maximum_and rapidly drops to zero when the mass of the black hole reaches the_value_of the_central_charge._Specific heat blows up when_the temperature reaches the maximum_and changes_the sign, see Figs. 4, 6, 7 of_[@US]._The change of_the sign of the heat capacity happens at the_non-vanishing temperature and means that the_corresponding component of_the_Weinhold_metric undergoes the same_type of changes.
The second important point_is that, as pointed out in_[@GKK], when considering black holes with scalars_we must augment the usual extensive_thermodynamic variables such as the_area $A$_and the conserved charges $(q,p)$_with the values_of the_moduli at infinity_$\phi_\infty$.
This has the consequence that the_thermodynamic configuration space_is no longer flat ${ \bf_R}^k_\equiv (A, q,p)$_(_where_$k$ is_one plus two_times_the number_of_electric charges, but becomes its product_with_the scalar moduli space ${ \bf R}^k_\times {\cal M}_\phi$.
It was_shown_in [@GKK] that the_thermodynamic variables conjugate to to_the moduli $\phi^a_\infty$, i.e. the analogue_of the_chemical potentials_$\mu_a$ are minus the scalar charges, i.e. one has the relation_$$dM=TdS+\psi^A dq_A + \chi_A dp^A -_\Sigma _a d \phi_^a _\infty_\.$$
The_fact that the_scalar_moduli space_${\cal M}_\phi$ is no longer flat complicates,_but of_course does not invalidate, the usual_thermodynamic formalism involving Legendre_transformations_(technically one should speak of Legendre_submanifolds etc.) and in particular one_is allowed to extend the_definition_of_the Weinhold and Ruppeiner metrics_to the scalar moduli space ${\cal_M}_\phi$ provided we_replace the ordinary derivative by the covariant_derivative_with respect to the metric $G_{ab}$_on_the moduli space.
In the present paper_we_have_been considering extreme black holes_for which the temperature $T=0$ and_it is the Weinhold metric $W_{ab}$ which seems to_be the more_appropriate one to consider. This_may_be_defined by $$W_{ab}= |
k(M).$$
We assume that those exponents $s_i$ having the same real part are ordered such that their imaginary parts are increasing. (The factor of $\rho^{-r}$ in is motived by the fact that $|u|_g = \rho^r |u|_{{\overline}g}$; thus the leading order behavior of $|u|_{ g}$ will be as $\rho^{{\mathrm{Re}}{(s_0)}}(\log\rho)^{p_0}$.)
If $u$ satisfies the definition above, we write $$u\sim \sum_{i=0}^{\infty} \sum_{p=0}^{p_i} \rho^{s_i-r} (\log\rho)^p {\overline}u_{ip}.$$ Let $\mathcal B_{\textup{\normalfont phg}}(M)$ be the collection of all tensor fields on $M$ which are polyhomogeneous as defined above. We furthermore denote by $\mathcal B^{\textup{\normalfont phg}}_\delta(M)$ those polyhomogeneous tensor fields that are of class $\mathcal B_\delta(M)$, and by $C^{k}_{\textup{\normalfont phg}}({\overline}M)$ those polyhomogeneous tensor fields extending to tensor fields of class $C^k$ on ${\overline}M$.
\[PhgSplitting\]
1. It follows from Lemma \[LinInd\] that if $u\in \mathcal B^{\textup{\normalfont phg}}_\delta(M)$ then we have ${\mathrm{Re}}(s_0)\geq \delta$.
2. Polyhomogeneous expansions are unique in the sense that if $$v\sim \sum_{i=0}^\infty \sum_{p=0}^{p_i} \rho^{s_i} (\log \rho)^p {\overline}v_{ip} \text{\ \ and \ \ } v\sim \sum_{i=0}^\infty \sum_{q=0}^{q_i} \rho^{t_i} (\log \rho)^p {\overline}w_{ip},$$ then $s_i=t_i$, $p_i=q_i$, and ${\overline}v_{ip}={\overline}w_{jp}$.
3. Tensor fields $u$ which are smooth on ${\overline}M$ are polyhomogeneous with a Taylor-series like expansion $$u\sim \sum_{n=0}^\ | k(M).$$
We assume that those exponents $ s_i$ having the same veridical character are ordered such that their imaginary parts are increase. (The factor of $ \rho^{-r}$ in is motived by the fact that $ |u|_g = \rho^r |u|_{{\overline}g}$; thus the precede club behavior of $ |u| _ { g}$ will be as $ \rho^{{\mathrm{Re}}{(s_0)}}(\log\rho)^{p_0}$.)
If $ u$ satisfies the definition above, we publish $ $ u\sim \sum_{i=0}^{\infty } \sum_{p=0}^{p_i } \rho^{s_i - r } (\log\rho)^p { \overline}u_{ip}.$$ Let $ \mathcal B_{\textup{\normalfont phg}}(M)$ be the solicitation of all tensor fields on $ M$ which are polyhomogeneous as defined above. We furthermore denote by $ \mathcal B^{\textup{\normalfont phg}}_\delta(M)$ those polyhomogeneous tensor fields that are of class $ \mathcal B_\delta(M)$, and by $ C^{k}_{\textup{\normalfont phg}}({\overline}M)$ those polyhomogeneous tensor field extending to tensor fields of class $ C^k$ on $ { \overline}M$.
\[PhgSplitting\ ]
1. It succeed from Lemma \[LinInd\ ] that if $ u\in \mathcal B^{\textup{\normalfont phg}}_\delta(M)$ then we have $ { \mathrm{Re}}(s_0)\geq \delta$.
2. Polyhomogeneous expansions are singular in the sense that if $ $ v\sim \sum_{i=0}^\infty \sum_{p=0}^{p_i } \rho^{s_i } (\log \rho)^p { \overline}v_{ip } \text{\ \ and \ \ } v\sim \sum_{i=0}^\infty \sum_{q=0}^{q_i } \rho^{t_i } (\log \rho)^p { \overline}w_{ip},$$ then $ s_i = t_i$, $ p_i = q_i$, and $ { \overline}v_{ip}={\overline}w_{jp}$.
3. Tensor fields $ u$ which are smooth on $ { \overline}M$ are polyhomogeneous with a Taylor - series like expansion $ $ u\sim \sum_{n=0}^\ | k(M).$$
Wf assume that those expokents $s_i$ having jhw same real lart are ordered such that their imajinaey paets are increasing. (The factor ov $\rho^{-r}$ ib is notived by the facb thaf $|u|_g = \cho^r |u|_{{\overline}g}$; thus the neading order tefarior of $|u|_{ g}$ will be as $\rho^{{\mathrm{Re}}{(s_0)}}(\ljg\rho)^{p_0}$.)
Ig $k$ satisfies thg defpnytioh above, we write $$u\sim \sum_{i=0}^{\infty} \suj_{p=0}^{p_i} \rhm^{s_i-r} (\log\rho)^p {\pverline}u_{ip}.$$ Let $\mathcal B_{\tfxtuo{\normalfont phg}}(M)$ bf the collextiog of all tensur fields on $M$ which ate polyhomogeneous as defined abuve. Wz furthermotz dejmte by $\mathral B^{\tvxtup{\normalfokn phg}}_\denta(M)$ thpse polyhomogekeous tebsor fields that are mf class $\mathcal B_\qelta(M)$, ang yy $C^{k}_{\textup{\normalfont pyg}}({\ovetline}K)$ thuwe oolghpmkgeneoks vensor fielss extendint to tensor fields pf bkass $C^k$ on ${\oberlinq}M$.
\[[hgSplitting\]
1. It follows from Lemma \[LinPnd\] fhat if $u\in \mathcal B^{\teztup{\normalfont phg}}_\delja(M)$ then wq have ${\mathrm{Re}}(s_0)\geq \delta$.
2. Polyhomogeneous expanshons erd uulzjw ln the sense that if $$v\sim \sum_{i=0}^\infty \sum_{p=0}^{p_i} \rhj^{a_i} (\lpg \rho)^p {\overlike}v_{ip} \text{\ \ and \ \ } v\dik \sum_{i=0}^\infty \suo_{q=0}^{q_i} \rkk^{t_j} (\log \rho)^p {\overlinf}w_{ip},$$ thgn $s_i=t_u$, $p_i=q_i$, anq ${\ovrrline}v_{ip}={\overline}w_{jp}$.
3. Tensoe fields $u$ wkicy are smooth on ${\ovzrline}M$ are 'olyhokogenrous with a Taylor-serier liie expansioj $$u\sim \suj_{v=0}^\ | k(M).$$ We assume that those exponents $s_i$ same part are such that their factor $\rho^{-r}$ in is by the fact $|u|_g = \rho^r |u|_{{\overline}g}$; thus the order behavior of $|u|_{ g}$ will be as $\rho^{{\mathrm{Re}}{(s_0)}}(\log\rho)^{p_0}$.) If $u$ satisfies the above, we write $$u\sim \sum_{i=0}^{\infty} \sum_{p=0}^{p_i} \rho^{s_i-r} (\log\rho)^p {\overline}u_{ip}.$$ Let $\mathcal B_{\textup{\normalfont phg}}(M)$ the of tensor on $M$ which are polyhomogeneous as defined above. We furthermore denote by $\mathcal B^{\textup{\normalfont phg}}_\delta(M)$ those tensor fields that are of class $\mathcal B_\delta(M)$, by $C^{k}_{\textup{\normalfont phg}}({\overline}M)$ those tensor fields extending to tensor of $C^k$ on \[PhgSplitting\] It from Lemma \[LinInd\] if $u\in \mathcal B^{\textup{\normalfont phg}}_\delta(M)$ then we have ${\mathrm{Re}}(s_0)\geq \delta$. 2. Polyhomogeneous expansions are unique in the that if \sum_{p=0}^{p_i} \rho^{s_i} \rho)^p \text{\ and \ \ \sum_{i=0}^\infty \sum_{q=0}^{q_i} \rho^{t_i} (\log \rho)^p {\overline}w_{ip},$$ and ${\overline}v_{ip}={\overline}w_{jp}$. 3. Tensor fields $u$ which are on ${\overline}M$ polyhomogeneous with a Taylor-series like expansion \sum_{n=0}^\ | k(M).$$
We assume that those exponeNts $s_i$ havinG the sAme ReaL pArt aRe orDered such that tHEir iMaginary parts are increaSing. (THe FActoR Of $\Rho^{-r}$ iN is motiVEd BY The FaCt ThaT $|u|_G = \RhO^r |u|_{{\ovErlIne}g}$; thuS the leadinG orDeR behavior of $|u|_{ G}$ WiLl be as $\rho^{{\mAthRm{Re}}{(s_0)}}(\log\rho)^{p_0}$.)
if $u$ SatisfIeS thE DefinItiOn aboVe, we wrITe $$u\sim \Sum_{i=0}^{\infty} \SuM_{P=0}^{p_i} \rho^{S_I-r} (\log\rhO)^P {\OvErliNe}u_{ip}.$$ Let $\mathcal B_{\tEXtUP{\normalfont phg}}(m)$ be the CoLLeCTIon Of aLl tensor fiElDs on $M$ WHich are POlYHOMogENeous as defineD above. We furTHerMore deNoTe bY $\MathcaL B^{\texTuP{\NorMalfont phg}}_\dElta(m)$ those polYhomogENeous teNSor fielDs that Are Of cLass $\MAtHcAl B_\DeLTa(M)$, ANd By $C^{K}_{\TexTup{\normaLfOnT phg}}({\oVerlINE}m)$ ThosE poLyhoMogenEous tensor fieLds ExteNDinG to teNsor fIeldS oF clasS $C^k$ on ${\oVerliNe}m$.
\[PhgSplitting\]
1. It FollOws from LeMma \[liNInD\] tHat if $U\In \mathCal b^{\teXtup{\norMalfont PHg}}_\dElTA(m)$ ThEn we have ${\mathrm{Re}}(s_0)\gEq \DELtA$.
2. PolyhomOgeneoUS eXpANsions arE uNiqUe in THE sensE thaT If $$V\sim \sum_{i=0}^\Infty \sUM_{p=0}^{P_i} \Rho^{s_i} (\loG \rHo)^p {\oveRlIne}V_{ip} \Text{\ \ aND \ \ } v\siM \sum_{i=0}^\iNfty \sum_{q=0}^{Q_i} \rho^{T_I} (\log \rho)^p {\overliNE}w_{ip},$$ then $s_i=t_i$, $p_I=Q_i$, AND ${\oVErliNe}v_{Ip}={\overline}w_{Jp}$.
3. TeNSor fIeldS $U$ wHicH Are smOoth oN ${\oVErLIne}M$ are polyhomogeneOuS with a tayloR-series like exPansion $$u\siM \SUM_{n=0}^\ | k(M).$$
We assume that th ose expone nts $ s_i $ h av ingthesame real part areordered such that thei r ima gi n aryp ar ts ar e incre a si n g . ( Th efac to r o f $\r ho^ {-r}$ i n is motiv edby the fact th a t$|u|_g = \ rho ^r |u|_{{\ov erl ine}g} $; th u s the le ading order behavi or of $|u |_ { g}$ w i ll be a s $\ rho^ {{\mathrm{Re}}{(s _ 0) } }(\log\rho)^{p _0}$.)
I f$ u $ s ati sfies thede finit i on abov e ,w e wri t e $$u\sim \su m_{i=0}^{\i n fty } \sum _{ p=0 } ^{p_i} \rho ^{ s _i- r} (\log\rh o)^p {\overli ne}u_{ i p}.$$ L e t $\mat hcal B _{\ tex tup{ \ no rm alf on t ph g }} (M) $ be the col le ct ion o f al l t e nsor fi elds on $ M$ which arepol yhom o gen eousas de fine dabove . We f urthe rm ore denote by $ \mat hcal B^{\ tex tu p{\ no rmalf o nt phg }}_ \de lta(M)$ thosep oly ho m o g en eous tensor fields t h a tare of c lass $ \ ma th c al B_\de lt a(M )$,a n d by$C^{ k }_ {\textup {\norm a lf on t phg}} ({ \overl in e}M )$those poly homoge neous te nsorf ields extendin g to tensor fi e ld s of clas s $ C^k$ on ${\ over l ine} M$.\[ Phg S plitt ing\]
1 .It follows from Lem ma \[Lin Ind\] that if $u\i n \mathcal B ^ {\textup {\no r ma l font phg}}_\de lta(M )$ then we have ${\ mathr m{Re}}(s _0)\geq \ d e lta$.
2 . Pol yho mog e n eo us expansions a re u ni que inthe sensetha t i f $ $v\ si m \sum_{i =0}^\inf ty \ su m_ {p= 0}^{p _ i} \rho^ {s _i} ( \lo g \rh o )^p {\ overl ine} v_ {i p } \ text{\\ a n d \ \ } v \sim \s um _{i=0 }^\i n fty \sum_{ q=0}^{q_i } \ r ho^{ t_ i} (\log\rho)^p {\ove rl ine}w_{ip} ,$ $ t hen $s _ i =t_i$, $ p_i=q_i$, and ${\overli n e}v_{ip }={ \over line }w_{jp}$.
3 . Ten sor fields $u$ w hichar e s m o oth o n ${ \ov er line}M$ ar e pol yhomo ge neou s witha Taylor-series li k e e xpansion $$u\ sim \su m _ {n =0} ^ \ | k(M).$$
We assume_that those_exponents $s_i$ having the_same real_part_are ordered_such_that their imaginary_parts are increasing._(The factor of $\rho^{-r}$_in is motived_by_the fact that $|u|_g = \rho^r |u|_{{\overline}g}$; thus the leading order behavior of $|u|_{_g}$_will be_as_$\rho^{{\mathrm{Re}}{(s_0)}}(\log\rho)^{p_0}$.)
If_$u$ satisfies the definition above,_we write $$u\sim \sum_{i=0}^{\infty} \sum_{p=0}^{p_i}_\rho^{s_i-r} (\log\rho)^p_{\overline}u_{ip}.$$ Let $\mathcal B_{\textup{\normalfont phg}}(M)$ be the collection_of_all tensor fields_on $M$ which are polyhomogeneous as defined above. We_furthermore denote by $\mathcal B^{\textup{\normalfont phg}}_\delta(M)$_those polyhomogeneous tensor_fields_that_are of class $\mathcal_B_\delta(M)$, and by $C^{k}_{\textup{\normalfont phg}}({\overline}M)$ those_polyhomogeneous tensor fields extending to tensor_fields of class $C^k$ on ${\overline}M$.
\[PhgSplitting\]
1. _It follows from Lemma \[LinInd\] that_if $u\in \mathcal B^{\textup{\normalfont phg}}_\delta(M)$_then we_have ${\mathrm{Re}}(s_0)\geq \delta$.
2. Polyhomogeneous_expansions are unique_in the_sense that if_$$v\sim \sum_{i=0}^\infty \sum_{p=0}^{p_i} \rho^{s_i} (\log \rho)^p_{\overline}v_{ip} \text{\ \_and \ \ } v\sim \sum_{i=0}^\infty_\sum_{q=0}^{q_i}_\rho^{t_i} (\log \rho)^p_{\overline}w_{ip},$$_then_$s_i=t_i$, $p_i=q_i$,_and ${\overline}v_{ip}={\overline}w_{jp}$.
3. _Tensor_fields $u$_which_are smooth on ${\overline}M$ are polyhomogeneous_with_a Taylor-series like expansion $$u\sim \sum_{n=0}^\ |
varieties].
The Grassmannian $\operatorname{Gr}(1,{\mathbb{P}}^3)$ of lines in ${\mathbb{P}}^3$ is a $4$-dimensional variety that embeds into ${\mathbb{P}}^5$ via the Plücker embedding. In particular, the line in $3$-space spanned by the distinct points $(x_0 : x_1 : x_2 : x_3), (y_0 : y_1 : y_2 : y_3) \in {\mathbb{P}}^3$ is identified with the point $(p_{0,1} : p_{0,2} : p_{0,3} : p_{1,2} : p_{1,3} : p_{2,3}) \in {\mathbb{P}}^5$, where $p_{i,j}$ is the minor formed of $i$th and $j$th columns of the matrix $\left[
\begin{smallmatrix}
x_0 & x_1 & x_2 & x_3 \\
y_0 & y_1 & y_2 & y_3
\end{smallmatrix}
\right]$. The Plücker coordinates $p_{i,j}$ satisfy the relation $p_{0,1} p_{2,3} - p_{0,2} p_{1,3} + p_{0,3} p_{1,2} = 0$. Moreover, every point in ${\mathbb{P}}^5$ satisfying this relation is the Plücker coordinates of some line. Dually, a line in ${\mathbb{P}}^3$ is the intersection of two distinct planes. If the planes are given by the equations $a_0 x_0 + a_1 x_1 + a_2 x_2 + a_3 x_3 = 0$ and $b_0 x_0 + b_1 x_1 + b_2 x_2 + b_3 x_3 = 0$, then the minors $q_{i,j}$ of the matrix $\left[
\begin{smallmatrix}
a_0 & a_1 & a_2 & a_3 \\
b_0 & b_1 & b_2 & b_3
\end{smallmatrix}
\right]$ are the dual Plücker coordinates and also satisfy $q_{0 | varieties ].
The Grassmannian $ \operatorname{Gr}(1,{\mathbb{P}}^3)$ of lines in $ { \mathbb{P}}^3 $ is a $ 4$-dimensional variety that embeds into $ { \mathbb{P}}^5 $ via the Plücker implant. In especial, the line in $ 3$-space spanned by the distinct point $ (x_0: x_1: x_2: x_3), (y_0: y_1: y_2: y_3) \in { \mathbb{P}}^3 $ is identified with the point $ (p_{0,1 }: p_{0,2 }: p_{0,3 }: p_{1,2 }: p_{1,3 }: p_{2,3 }) \in { \mathbb{P}}^5 $, where $ p_{i, j}$ is the minor shape of $ i$th and $ j$th column of the matrix $ \left [
\begin{smallmatrix }
x_0 & x_1 & x_2 & x_3 \\
y_0 & y_1 & y_2 & y_3
\end{smallmatrix }
\right]$. The Plücker organize $ p_{i, j}$ satisfy the relation $ p_{0,1 } p_{2,3 } - p_{0,2 } p_{1,3 } + p_{0,3 } p_{1,2 } = 0$. furthermore, every point in $ { \mathbb{P}}^5 $ satisfying this relation is the Plücker coordinates of some lineage. Dually, a line in $ { \mathbb{P}}^3 $ is the intersection of two distinct planes. If the planes are contribute by the equations $ a_0 x_0 + a_1 x_1 + a_2 x_2 + a_3 x_3 = 0 $ and $ b_0 x_0 + b_1 x_1 + b_2 x_2 + b_3 x_3 = 0 $, then the minors $ q_{i, j}$ of the matrix $ \left [
\begin{smallmatrix }
a_0 & a_1 & a_2 & a_3 \\
b_0 & b_1 & b_2 & b_3
\end{smallmatrix }
\right]$ are the dual Plücker coordinate and also satisfy $ q_{0 | varleties].
The Grassmannian $\okeratorname{Gr}(1,{\mathyv{P}}^3)$ of nines jn ${\mathbc{P}}^3$ is a $4$-dimensional variety vhat embees into ${\mathbb{P}}^5$ via thd Plücker vmbedding. In karticular, the liis in $3$-space spzkned yy the distinct koints $(x_0 : x_1 : x_2 : x_3), (y_0 : y_1 : y_2 : y_3) \iu {\mathbb{P}}^3$ is identified with the poigt $(p_{0,1} : p_{0,2} : o_{0,3} : p_{1,2} : p_{1,3} : p_{2,3}) \in {\iathnf{P}}^5$, wgvrt $p_{i,j}$ is the minor formed of $i$th znd $j$th columns of tne matrix $\left[
\begin{smalllatrlx}
x_0 & x_1 & x_2 & x_3 \\
y_0 & y_1 & i_2 & y_3
\end{smallmagrix}
\right]$. The Plücker doordinates $p_{i,j}$ satisfy the relxtion $p_{0,1} p_{2,3} - p_{0,2} p_{1,3} + p_{0,3} p_{1,2} = 0$. Moreover, etery pjint in ${\mathnn{P}}^5$ sathsfying this relation is thw Plücker coordinates mf some line. Dualli, a line it ${\jathbb{P}}^3$ is the inreesecthon mf tdi dkstjnrt llanes. If the planea are given by the equations $a_0 x_0 + a_1 x_1 + a_2 x_2 + a_3 s_3 = 0$ anq $f_0 x_0 + b_1 x_1 + b_2 x_2 + b_3 x_3 = 0$, then the minors $q_{i,b}$ or the matrix $\left[
\begib{smallmatrix}
a_0 & a_1 & a_2 & a_3 \\
b_0 & b_1 & b_2 & b_3
\end{smallmatrix}
\right]$ are the dual Plückxr coirqkbahes and also satisfy $q_{0 | varieties]. The Grassmannian $\operatorname{Gr}(1,{\mathbb{P}}^3)$ of lines in a variety that into ${\mathbb{P}}^5$ via the in $3$-space spanned the distinct points : x_1 : x_2 : x_3), : y_1 : y_2 : y_3) \in {\mathbb{P}}^3$ is identified with the point : p_{0,2} : p_{0,3} : p_{1,2} : p_{1,3} : p_{2,3}) \in {\mathbb{P}}^5$, where is minor of and $j$th columns of the matrix $\left[ \begin{smallmatrix} x_0 & x_1 & x_2 & x_3 \\ & y_1 & y_2 & y_3 \end{smallmatrix} \right]$. Plücker coordinates $p_{i,j}$ satisfy relation $p_{0,1} p_{2,3} - p_{0,2} + p_{1,2} = Moreover, point ${\mathbb{P}}^5$ satisfying this is the Plücker coordinates of some line. Dually, a line in ${\mathbb{P}}^3$ is the intersection of two planes. If are given the $a_0 + a_1 x_1 x_2 + a_3 x_3 = 0$ + b_1 x_1 + b_2 x_2 + b_3 = 0$, the minors $q_{i,j}$ of the matrix \begin{smallmatrix} a_0 & a_1 & a_2 & a_3 b_0 & b_1 & b_2 & b_3 \end{smallmatrix} \right]$ are the dual Plücker coordinates and $q_{0 | varieties].
The Grassmannian $\oPeratornamE{Gr}(1,{\maThbB{P}}^3)$ oF lInes In ${\maThbb{P}}^3$ is a $4$-dimensIOnal Variety that embeds into ${\mAthbb{p}}^5$ vIA the pLüCker eMbeddinG. in PARtiCuLaR, thE lINe In $3$-spaCe sPanned bY the distinCt pOiNts $(x_0 : x_1 : x_2 : x_3), (y_0 : y_1 : y_2 : y_3) \iN {\MaThbb{P}}^3$ is ideNtiFied with the pOinT $(p_{0,1} : p_{0,2} : p_{0,3} : p_{1,2} : p_{1,3} : P_{2,3}) \iN {\maTHbb{P}}^5$, wHerE $p_{i,j}$ iS the miNOr formEd of $i$th anD $j$TH columNS of the mATRiX $\lefT[
\begin{smallmatrix}
X_0 & X_1 & x_2 & X_3 \\
Y_0 & y_1 & y_2 & y_3
\end{smallmaTrix}
\riGhT]$. thE pLücKer CoordinateS $p_{I,j}$ satISfy the rELaTION $p_{0,1} p_{2,3} - P_{0,2} P_{1,3} + p_{0,3} p_{1,2} = 0$. Moreover, evEry point in ${\mAThbB{P}}^5$ satiSfYinG This reLatioN iS The plücker coorDinaTes of some Line. DuALly, a linE In ${\mathbB{P}}^3$ is thE inTerSectIOn Of Two DiSTinCT pLanES. If The planeS aRe Given By thE EQUAtioNs $a_0 X_0 + a_1 x_1 + a_2 X_2 + a_3 x_3 = 0$ anD $b_0 x_0 + b_1 x_1 + b_2 x_2 + b_3 x_3 = 0$, then The MinoRS $q_{i,J}$ of thE matrIx $\leFt[
\Begin{SmallmAtrix}
A_0 & a_1 & A_2 & a_3 \\
b_0 & b_1 & b_2 & b_3
\end{smallmAtriX}
\right]$ are The DuAl PLüCker cOOrdinaTes And Also satIsfy $q_{0 | varieties].
The Grassmann ian $\oper atorn ame {Gr }( 1,{\ math bb{P}}^3)$ ofl ines in ${\mathbb{P}}^3$ i s a $ 4$ - dime n si onalvariety th a t em be ds in to ${ \math bb{ P}}^5$via the Pl ück er embedding.I nparticular , t he line in $ 3$- spacesp ann e d bythe dist inct p o ints $ (x_0 : x_ 1: x_2 : x_3), ( y _ 0: y_ 1 : y_2 : y_3) \i n { \ mathbb{P}}^3$is ide nt i fi e d wi ththe point$( p_{0, 1 } : p_{ 0 ,2 } : p_ { 0,3} : p_{1,2 } : p_{1,3} : p _{2,3} )\in {\math bb{P} }^ 5 $,where $p_{i ,j}$ is the m inor f o rmed of $i$th a nd $j$ thcol umns of t hema t rix $\ lef t [
\begin{ sm al lmatr ix} x_0 &x_1& x_2 & x_3 \\
y _0 & y_1 & y_ 2 & y _3 \end{ smallm atrix }
\right]$. ThePlüc ker coord ina te s $ p_ {i,j} $ satis fythe relati on $p_{ 0 ,1} p _ { 2 ,3 } - p_{0,2} p_{1,3 }+ p_ {0,3} p_ {1,2}= 0 $. Moreover ,eve ry p o i nt in ${\ m at hbb{P}}^ 5$ sat i sf yi ng this r elatio nisthe Plüc k er c oordin ates ofsomel ine. Dually, a line in ${\ma t hb b { P} } ^3$isthe interse ctio n oftwod is tin c t pla nes.If th e planes are given b ythe eq uatio ns $a_0 x_0 + a_1 x_1 + a _ 2 x_2 +a_3x _3 = 0$ and $b_0x_0 + b_1 x_1 + b_2 x_2+ b_3 x_3 = 0 $, then t h e minors$q_ {i, j}$ of t he matrix $\lef t [
\ be gin{sma llm atrix}
a _0& a _1 & a_2 &a_3 \\ b_ 0& b _1 &b _2 & b_3 \ en d{s mallm a trix}
\rig ht]$ a re the dual P l üc k e r co or di nate s a nd also sat i sfy $q_{0 | varieties].
The Grassmannian_$\operatorname{Gr}(1,{\mathbb{P}}^3)$ of_lines in ${\mathbb{P}}^3$ is_a $4$-dimensional_variety_that embeds_into_${\mathbb{P}}^5$ via the_Plücker embedding. In_particular, the line in_$3$-space spanned by_the_distinct points $(x_0 : x_1 : x_2 : x_3), (y_0 : y_1 : y_2_:_y_3) \in_{\mathbb{P}}^3$_is_identified with the point $(p_{0,1}_: p_{0,2} : p_{0,3} :_p_{1,2} :_p_{1,3} : p_{2,3}) \in {\mathbb{P}}^5$, where $p_{i,j}$ is_the_minor formed of_$i$th and $j$th columns of the matrix $\left[
_\begin{smallmatrix}
x_0_& x_1 &_x_2_&_x_3 \\
_ y_0 & y_1 &_y_2 & y_3
\end{smallmatrix}_
\right]$. The Plücker coordinates $p_{i,j}$ satisfy the_relation $p_{0,1} p_{2,3} - p_{0,2} p_{1,3}_+ p_{0,3} p_{1,2} = 0$._Moreover, every_point in ${\mathbb{P}}^5$ satisfying this_relation is the_Plücker coordinates_of some line._Dually, a line in ${\mathbb{P}}^3$ is_the intersection of_two distinct planes. If the planes_are_given by the_equations_$a_0_x_0 +_a_1 x_1 +_a_2_x_2 +_a_3_x_3 = 0$ and $b_0 x_0_+_b_1 x_1 + b_2 x_2 + b_3_x_3 = 0$, then_the_minors $q_{i,j}$ of the_matrix $\left[
\begin{smallmatrix}
_ a_0 & a_1_& a_2_& a_3_\\
b_0 & b_1 & b_2 &_b_3
\end{smallmatrix}
\right]$ are_the dual Plücker coordinates_and also_satisfy_$q_{0 |
& = e^{i B_2(\zeta)} V e^{-i B_1(\zeta)}, \quad \zeta\in{{\mathbb{C}}}, {\label}{2.14o}\end{aligned}$$ where we have set $$\begin{aligned}
\begin{split}
B_1(\zeta) = \zeta (A_1^*-{\overline}z)^{-1} + {{\overline}\zeta} (A_1 - z I_{{{\mathcal H}}_1})^{-1} = B_1(\zeta)^*, \\
B_2(\zeta) = \zeta (A_2^*-{\overline}z)^{-1} + {{\overline}\zeta} (A_2 - z I_{{{\mathcal H}}_2})^{-1} = B_2(\zeta)^*. {\label}{2.14p}
\end{split}\end{aligned}$$ Consequently, $$\big\|e^{i \zeta (A_2^*-{\overline}z)^{-1}} V e^{-i \zeta (A_1^*-{\overline}z)^{-1}}\big\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2)}
= \|V\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2)},
\quad \zeta \in {{\mathbb{C}}}. {\label}{2.14q}$$ Since the left-hand side of is entire with respect to $\zeta\in{{\mathbb{C}}}$, the uniform boundedness in and Liouville’s theorem yield that the left-hand side of is actually constant with respect to $\zeta\in{{\mathbb{C}}}$. Thus, the left-hand side of equals its value at $\zeta=0$, allowing one to conclude that $$e^{i \zeta (A_2^*-{\overline}z)^{-1}} V e^{-i \zeta (A_1^*-{\overline}z)^{-1}} = V, \quad \zeta\in{{\mathbb{C}}}. {\label}{2.14r}$$ Differentiating with respect to $\zeta$ and subsequently taking $\zeta =0$, then yields $$V (A_1^*-{\overline}z)^{-1} = (A_2^*-{\overline}z)^{-1} V, {\label}{2.14s}$$ and consequently, $$V A_1^* = A_2^* V \, \text{ on } \, {\text{\rm{dom}}}(A_1^* | & = e^{i B_2(\zeta) } V e^{-i B_1(\zeta) }, \quad \zeta\in{{\mathbb{C } } }, { \label}{2.14o}\end{aligned}$$ where we have set $ $ \begin{aligned }
\begin{split }
B_1(\zeta) = \zeta (A_1^*-{\overline}z)^{-1 } + { { \overline}\zeta } (A_1 - z I_{{{\mathcal H}}_1})^{-1 } = B_1(\zeta)^ *, \\
B_2(\zeta) = \zeta (A_2^*-{\overline}z)^{-1 } + { { \overline}\zeta } (A_2 - z I_{{{\mathcal H}}_2})^{-1 } = B_2(\zeta)^ *. { \label}{2.14p }
\end{split}\end{aligned}$$ Consequently, $ $ \big\|e^{i \zeta (A_2^*-{\overline}z)^{-1 } } V e^{-i \zeta (A_1^*-{\overline}z)^{-1}}\big\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2) }
= \|V\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2) },
\quad \zeta \in { { \mathbb{C } } }. { \label}{2.14q}$$ Since the left - bridge player slope of is entire with respect to $ \zeta\in{{\mathbb{C}}}$, the uniform finiteness in and Liouville ’s theorem yield that the left - hand slope of is actually constant with respect to $ \zeta\in{{\mathbb{C}}}$. therefore, the leftover - hand side of equal its value at $ \zeta=0 $, allowing one to conclude that $ $ e^{i \zeta (A_2^*-{\overline}z)^{-1 } } five e^{-i \zeta (A_1^*-{\overline}z)^{-1 } } = V, \quad \zeta\in{{\mathbb{C } } }. { \label}{2.14r}$$ Differentiating with respect to $ \zeta$ and subsequently taking $ \zeta = 0 $, then yields $ $ V (A_1^*-{\overline}z)^{-1 } = (A_2^*-{\overline}z)^{-1 } V, { \label}{2.14s}$$ and consequently, $ $ volt A_1^ * = A_2^ * V \, \text { on } \, { \text{\rm{dom}}}(A_1^ * |
& = e^{l B_2(\zeta)} V e^{-i B_1(\zeta)}, \quaa \zeta\in{{\mathbb{C}}}, {\labxl}{2.14o}\end{amigned}$$ wfere we have set $$\begin{alignev}
\begun{splut}
B_1(\zeta) = \zeta (A_1^*-{\overlive}z)^{-1} + {{\overpine}\zeta} (A_1 - e I_{{{\mathcal H}}_1})^{-1} = B_1(\zevz)^*, \\
B_2(\zebc) = \zefw (A_2^*-{\orecline}z)^{-1} + {{\overling}\zeta} (A_2 - z I_{{{\kathcal H}}_2})^{-1} = B_2(\zeda)^*. {\label}{2.14p}
\end{split}\end{aligned}$$ Consequenely, $$\big\|r^{i \zeta (A_2^*-{\overling}z)^{-1}} V t^{-i \seta (A_1^*-{\overline}z)^{-1}}\big\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mafhcal H}}_2)}
= \|V\|_{{{\mathcal B}}({{\mayhcal H}}_1,{{\mathcal H}}_2)},
\quad \zeta \ln {{\mwthbb{C}}}. {\label}{2.14q}$$ Slnce the ledt-hage side of is entire wiuh respect tk $\zeta\in{{\mathbb{C}}}$, the uniform bouvdednzss in and Oiiuvlnle’s theoren yiejd that the left-hand side og is actually gonstent qith respect to $\zeta\ii{{\mathbb{C}}}$. Thus, the lest-hand sige of equals its vaouw at $\eeta=0$, dllodung ont tp donclufe vhat $$e^{i \zetz (A_2^*-{\overline}z)^{-1}} V e^{-i \zeta (A_1^*-{\overlime}s)^{-1}} = V, \quad \zetz\in{{\matrbf{C}}}. {\label}{2.14r}$$ Differentiating with respecu to $\aeta$ and subsequently tqking $\zeta =0$, then yielfs $$V (A_1^*-{\ovewline}z)^{-1} = (A_2^*-{\overline}z)^{-1} V, {\label}{2.14s}$$ and consequently, $$V D_1^* = A_2^* T \, \terb{ ub } \, {\text{\rm{dom}}}(A_1^* | & = e^{i B_2(\zeta)} V e^{-i B_1(\zeta)}, {\label}{2.14o}\end{aligned}$$ we have $$\begin{aligned} \begin{split} B_1(\zeta) (A_1 z I_{{{\mathcal H}}_1})^{-1} B_1(\zeta)^*, \\ B_2(\zeta) \zeta (A_2^*-{\overline}z)^{-1} + {{\overline}\zeta} (A_2 - I_{{{\mathcal H}}_2})^{-1} = B_2(\zeta)^*. {\label}{2.14p} \end{split}\end{aligned}$$ Consequently, $$\big\|e^{i \zeta (A_2^*-{\overline}z)^{-1}} V e^{-i \zeta B}}({{\mathcal H}}_1,{{\mathcal H}}_2)} = \|V\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2)}, \quad \zeta \in {{\mathbb{C}}}. {\label}{2.14q}$$ the side is with respect to $\zeta\in{{\mathbb{C}}}$, the uniform boundedness in and Liouville’s theorem yield that the left-hand side is actually constant with respect to $\zeta\in{{\mathbb{C}}}$. Thus, left-hand side of equals value at $\zeta=0$, allowing one conclude $$e^{i \zeta V \zeta = V, \quad {\label}{2.14r}$$ Differentiating with respect to $\zeta$ and subsequently taking $\zeta =0$, then yields $$V (A_1^*-{\overline}z)^{-1} = (A_2^*-{\overline}z)^{-1} {\label}{2.14s}$$ and A_1^* = V \text{ } \, {\text{\rm{dom}}}(A_1^* |
& = e^{i B_2(\zeta)} V e^{-i B_1(\zeta)}, \quad \zeta\in{{\Mathbb{C}}}, {\labEl}{2.14o}\enD{alIgnEd}$$ WherE we hAve set $$\begin{aliGNed}
\bEgin{split}
B_1(\zeta) = \zeta (A_1^*-{\oveRline}Z)^{-1} + {{\oVErliNE}\zEta} (A_1 - z i_{{{\mathcaL h}}_1})^{-1} = B_1(\ZETa)^*, \\
B_2(\ZeTa) = \ZetA (A_2^*-{\OVeRline}Z)^{-1} + {{\ovErline}\zEta} (A_2 - z I_{{{\mathCal h}}_2})^{-1} = B_2(\Zeta)^*. {\label}{2.14p}
\enD{SpLit}\end{aligNed}$$ consequently, $$\Big\|E^{i \zeta (a_2^*-{\oVerLIne}z)^{-1}} V E^{-i \zEta (A_1^*-{\oVerlinE}Z)^{-1}}\big\|_{{{\maThcal B}}({{\matHcAL H}}_1,{{\mathCAl H}}_2)}
= \|V\|_{{{\matHCAl b}}({{\matHcal H}}_1,{{\mathcal H}}_2)},
\quad \ZEtA \In {{\mathbb{C}}}. {\label}{2.14Q}$$ Since ThE LeFT-HanD siDe of is entiRe With rESpect to $\ZEtA\IN{{\MatHBb{C}}}$, the uniform Boundedness IN anD LiouvIlLe’s THeorem Yield ThAT thE left-hand siDe of Is actuallY constANt with rESpect to $\Zeta\in{{\MatHbb{c}}}$. ThuS, ThE lEft-HaND siDE oF eqUAls Its value At $\ZeTa=0$, allOwinG ONE To coNclUde tHat $$e^{i \Zeta (A_2^*-{\overline}Z)^{-1}} V e^{-I \zetA (a_1^*-{\ovErlinE}z)^{-1}} = V, \quAd \zeTa\In{{\matHbb{C}}}. {\laBel}{2.14r}$$ DIfFerentiating witH resPect to $\zetA$ anD sUbsEqUentlY Taking $\ZetA =0$, thEn yieldS $$V (A_1^*-{\overLIne}Z)^{-1} = (A_2^*-{\OVERlIne}z)^{-1} V, {\label}{2.14s}$$ and consEqUENtLy, $$V A_1^* = A_2^* V \, \teXt{ on } \, {\teXT{\rM{dOM}}}(A_1^* |
& = e^{i B_2(\zeta)} V e^ {-i B_1(\z eta)} , \q ua d \z eta\ in{{\mathbb{C} } }, {\label}{2.14o}\end{ align ed } $$ w h er e wehave se t $ $ \ beg in {a lig ne d }\begi n{s plit}
B _1(\zeta)= \z eta (A_1^*-{ \ ov erline}z)^ {-1 } + {{\overl ine }\zeta }(A_ 1 - zI_{ {{\ma thcalH }}_1}) ^{-1} = B _1 ( \zeta) ^ *, \\ B _2 (\ze ta) = \zeta (A_2^ * -{ \ overline}z)^{- 1} + { {\ o ve r l ine }\z eta} (A_2-z I_{ { {\mathc a lH } } _2} ) ^{-1} = B_2(\ zeta)^*. { \la bel}{2 .1 4p} \end{s plit} \e n d{a ligned}$$ C onse quently,$$\big \ |e^{i \ z eta (A_ 2^*-{\ ove rli ne}z ) ^{ -1 }}Ve ^{- i \ zet a (A _1^*-{\o ve rl ine}z )^{- 1 } } \ big\ |_{ {{\m athca l B}}({{\math cal H}} _ 1,{ {\mat hcalH}}_ 2) }
= \ |V\|_{ {{\ma th cal B}}({{\math calH}}_1,{{\ mat hc alH} }_2)} ,
\quad \z eta \in {{ \mathbb { C}} }. { \label}{2.14q}$$ S in c e t he left- hand s i de o f is enti re wi th r e s pectto $ \ ze ta\in{{\ mathbb { C} }} $, theun iformbo und edn ess i n and Liouv ille’s t heore m yield that th e left-hand si d eo f i s act ual ly constant wit h res pect to $\ z eta\i n{{\m at h bb { C}}}$. Thus, the le ft -handsideof equals its value at$ \ z eta=0$,allo w in g one to conclu de th at $$e^{i\ zeta (A_ 2^*-{ \overlin e}z)^{-1} } V e^{-i\ze ta(A_ 1^* - { \o verline}z)^{- 1 } } = V , \quad \z eta\in{ {\m ath bb{ C}} }. {\labe l}{2.14r }$ $Di ff ere ntiat i ng withre spe ct to $\ze t a$ and subs eque nt ly tak ing $\z e ta = 0$,th en yie lds $ $V (A _1^* - {\o verline }z)^{-1}= ( A _2^* -{ \o verline }z)^{-1} V, { \label}{2. 14 s}$ $ andc o nsequent ly, $$V A_1^* = A_2^* V \, \tex t{on }\, { \text{\rm {do m}}}(A _1^ * |
& =_e^{i B_2(\zeta)}_V e^{-i B_1(\zeta)}, _ \quad_\zeta\in{{\mathbb{C}}},_ __{\label}{2.14o}\end{aligned}$$ where we_have set $$\begin{aligned}
\begin{split}
B_1(\zeta)_= \zeta (A_1^*-{\overline}z)^{-1}_+ {{\overline}\zeta} (A_1_-_z I_{{{\mathcal H}}_1})^{-1} = B_1(\zeta)^*, \\
B_2(\zeta) = \zeta (A_2^*-{\overline}z)^{-1} + {{\overline}\zeta} (A_2_-_z I_{{{\mathcal_H}}_2})^{-1}_=_B_2(\zeta)^*. {\label}{2.14p}
\end{split}\end{aligned}$$ Consequently,_$$\big\|e^{i \zeta (A_2^*-{\overline}z)^{-1}} V e^{-i_\zeta (A_1^*-{\overline}z)^{-1}}\big\|_{{{\mathcal_B}}({{\mathcal H}}_1,{{\mathcal H}}_2)}
= \|V\|_{{{\mathcal B}}({{\mathcal H}}_1,{{\mathcal H}}_2)},
\quad \zeta_\in_{{\mathbb{C}}}. _ {\label}{2.14q}$$ Since the left-hand side of is entire_with respect to $\zeta\in{{\mathbb{C}}}$, the uniform_boundedness in and_Liouville’s_theorem_yield that the left-hand_side of is actually constant with_respect to $\zeta\in{{\mathbb{C}}}$. Thus, the left-hand_side of equals its value at $\zeta=0$,_allowing one to conclude that $$e^{i_\zeta (A_2^*-{\overline}z)^{-1}} V e^{-i \zeta_(A_1^*-{\overline}z)^{-1}} _= V, \quad \zeta\in{{\mathbb{C}}}. _ {\label}{2.14r}$$ Differentiating_with respect_to $\zeta$ and_subsequently taking $\zeta =0$, then yields_$$V (A_1^*-{\overline}z)^{-1} =_(A_2^*-{\overline}z)^{-1} V, {\label}{2.14s}$$ and_consequently,_$$V A_1^* =_A_2^*_V_\, \text{_on } \,_{\text{\rm{dom}}}(A_1^* |
**Masked Reconstruction Module:** We manually define the non-attribute area shown in Fig. \[fig:mask\_template\] on the reference UV texture map. Those $5$ attributes are divided into several different mask types or their combination,, lipsticks and smile share the same mask of Fig. \[fig:mask\_template\] (c). Together with the fully visible one (mask of all entries as $1$), we define mask $\Omega_{i}, i=1,2,..,5$ indicating all the categories. The reconstruction objective is as below. $$\begin{aligned}
\mathcal{L}_{ra} = \|(G_{a}(\hat{\mathbf{U}}_{t}, \mathbf{U}_{p}, \mathbf{p}) - \hat{\mathbf{U}}_{t})\odot \Omega_{i}\|_1,
\label{eq:3dagan_recon}
\vspace{-2mm}\end{aligned}$$ $i$ is determined by the target attribute code.
**Target Attribute Discriminator:** Separated from $Q$, we set an independent discriminator $A$ to evaluate whether the one-bit specific attribute is correctly generated or not. The positive sample set $\mathbf{R}_{a}$ consists of samples from the ground truth with the specific attribute. The negative sample set $\mathbf{F}_{a}$ are the samples generated from $G_{a}$. The target attribute discriminator is updated as: $$\begin{aligned}
\mathcal{L}_{A} = -\underset{\mathbf{U}^{*} \in \mathbf{R}_{a}}{\mathbb{E}} \log(A(\mathbf{U}^{*})) - \underset{\mathbf{U} \in \mathbf{F}_{a}}{\mathbb{E}}\log(1-A(\mathbf{U})))
\label{eq:3dagan_attrd}
\vspace{-2mm}\end{aligned}$$ Accordingly, the adversarial loss to update the generator is: $$\begin{aligned}
\mathcal{L}_{Av}=- \underset{\mathbf{U} \in \mathbf{F}_{a}}{\mathbb{E}}\log(A(\mathbf{U}))
\label{eq:3dagan_attradv}
\vspace{-4mm}\end{aligned}$$
In TC-GAN, we find that reconstruction loss other than recognition perception loss is sufficient to preserve identity. It also | * * Masked Reconstruction Module :* * We manually define the non - attribute area shown in Fig. \[fig: mask\_template\ ] on the reference ultraviolet texture function. Those $ 5 $ attribute are divided into respective different masquerade types or their combination, , lipsticks and smile partake the same mask of Fig. \[fig: mask\_template\ ] (c). in concert with the amply visible one (mask of all submission as $ 1 $), we define mask $ \Omega_{i }, i=1,2,.. ,5 $ indicate all the categories. The reconstruction objective is as below. $ $ \begin{aligned }
\mathcal{L}_{ra } = \|(G_{a}(\hat{\mathbf{U}}_{t }, \mathbf{U}_{p }, \mathbf{p }) - \hat{\mathbf{U}}_{t})\odot \Omega_{i}\|_1,
\label{eq:3dagan_recon }
\vspace{-2mm}\end{aligned}$$ $ i$ is determined by the target property code.
* * Target Attribute Discriminator :* * Separated from $ Q$, we set an independent differentiator $ A$ to evaluate whether the one - bit specific attribute is correctly generated or not. The positive sample distribution set $ \mathbf{R}_{a}$ consist of samples from the ground truth with the specific property. The negative sample set $ \mathbf{F}_{a}$ are the samples generated from $ G_{a}$. The target property differentiator is update as: $ $ \begin{aligned }
\mathcal{L}_{A } = -\underset{\mathbf{U}^ { * } \in \mathbf{R}_{a}}{\mathbb{E } } \log(A(\mathbf{U}^ { * }) ) - \underset{\mathbf{U } \in \mathbf{F}_{a}}{\mathbb{E}}\log(1 - A(\mathbf{U }) ))
\label{eq:3dagan_attrd }
\vspace{-2mm}\end{aligned}$$ Accordingly, the adversarial loss to update the generator is: $ $ \begin{aligned }
\mathcal{L}_{Av}=- \underset{\mathbf{U } \in \mathbf{F}_{a}}{\mathbb{E}}\log(A(\mathbf{U }) )
\label{eq:3dagan_attradv }
\vspace{-4mm}\end{aligned}$$
In TC - GAN, we find that reconstruction loss other than recognition percept personnel casualty is sufficient to preserve identity. It also | **Masned Reconstruction Modult:** We manually defnbe the non-atfribute xrea shown in Fig. \[fig:mask\_temppare\] on the reference UV textjre map. Tjose $5$ atrribntes are divided into seyzral slfferznv mask types or their comtination,, lipsthcys and smile share the same mask of Fyg. \[fig:maxk\_hemplate\] (c). Toggther rith nht fully visible one (mask of all ehtries es $1$), we define mssk $\Omega_{i}, i=1,2,..,5$ indicating alp thf categories. The rfconstructiin ofhective is ar below. $$\begin{aligned}
\majhcal{L}_{ra} = \|(G_{a}(\hat{\mathbf{U}}_{t}, \mathbf{U}_{p}, \mathyf{p}) - \hat{\matybd{U}}_{t})\lgot \Omega_{i}\|_1,
\oabel{vq:3dagan_recon}
\vspace{-2mm}\etd{alignrd}$$ $i$ is determlned uy tye target attribute cmde.
**Target Attributg Discrimitacor:** Separated from $Q$, qe set dn itdepdbdevt sixcdiminahor $A$ to evalhate whethee the one-bit specifoc qttribute is dorrecelr generated or not. The positive sample vet $\mathbf{R}_{a}$ consists of sqmples from the grounf truth wyth the specific attribute. The negative sample sed $\matibw{F}_{a}$ ave tfw damples generated from $G_{a}$. The target attributq dosbriminator is updcted as: $$\begin{alibnfd}
\kwthcal{L}_{A} = -\undgrset{\macgbr{U}^{*} \in \mathbf{R}_{a}}{\mathhb{E}} \log(W(\mathvf{U}^{*})) - \undewset{\kathbf{U} \in \mathbf{F}_{a}}{\mathbb{E}}\lig(1-A(\mathbf{U})))
\layel{wq:3dagan_attrd}
\vspace{-2lm}\end{aligneb}$$ Accotdinglu, the adversarial loss co updzte the genfrator is: $$\cegin{aligned}
\mathzal{K}_{Ae}=- \understg{\mathbf{U} \in \mathbs{F}_{a}}{\mathbb{X}}\log(A(\kathbf{U}))
\uabek{eq:3dagwn_attradv}
\vdpace{-4mm}\end{aligned}$$
In TC-GAJ, we yind dhat recondtruction loss other than recogimtion perceptoot lmss is sbfficicnt to preserve identity. It akso | **Masked Reconstruction Module:** We manually define the shown Fig. \[fig:mask\_template\] the reference UV are into several different types or their lipsticks and smile share the same of Fig. \[fig:mask\_template\] (c). Together with the fully visible one (mask of all as $1$), we define mask $\Omega_{i}, i=1,2,..,5$ indicating all the categories. The reconstruction is below. \mathcal{L}_{ra} \|(G_{a}(\hat{\mathbf{U}}_{t}, \mathbf{U}_{p}, \mathbf{p}) - \hat{\mathbf{U}}_{t})\odot \Omega_{i}\|_1, \label{eq:3dagan_recon} \vspace{-2mm}\end{aligned}$$ $i$ is determined by the target attribute code. Attribute Discriminator:** Separated from $Q$, we set an discriminator $A$ to evaluate the one-bit specific attribute is generated not. The sample $\mathbf{R}_{a}$ of samples from ground truth with the specific attribute. The negative sample set $\mathbf{F}_{a}$ are the samples generated from $G_{a}$. target attribute updated as: \mathcal{L}_{A} -\underset{\mathbf{U}^{*} \mathbf{R}_{a}}{\mathbb{E}} \log(A(\mathbf{U}^{*})) - \mathbf{F}_{a}}{\mathbb{E}}\log(1-A(\mathbf{U}))) \label{eq:3dagan_attrd} \vspace{-2mm}\end{aligned}$$ Accordingly, the adversarial the generator is: $$\begin{aligned} \mathcal{L}_{Av}=- \underset{\mathbf{U} \in \mathbf{F}_{a}}{\mathbb{E}}\log(A(\mathbf{U})) \vspace{-4mm}\end{aligned}$$ In we find that reconstruction loss other recognition perception loss is sufficient to preserve identity. also | **Masked Reconstruction ModulE:** We manuallY defiNe tHe nOn-AttrIbutE area shown in FiG. \[Fig:mAsk\_template\] on the refereNce UV TeXTure MAp. those $5$ AttribuTEs ARE diViDeD inTo SEvEral dIffErent maSk types or tHeiR cOmbination,, liPStIcks and smiLe sHare the same mAsk Of Fig. \[fIg:MasK\_TemplAte\] (C). TogeTher wiTH the fuLly visiblE oNE (mask oF All entrIES aS $1$), we dEfine mask $\Omega_{i}, i=1,2,..,5$ iNDiCAting all the catEgorieS. THE rECOnsTruCtion objecTiVe is aS Below. $$\beGIn{ALIGneD}
\Mathcal{L}_{ra} = \|(G_{a}(\hAt{\mathbf{U}}_{t}, \mAThbF{U}_{p}, \matHbF{p}) - \hAT{\mathbF{U}}_{t})\odOt \oMegA_{i}\|_1,
\label{eq:3daGan_rEcon}
\vspacE{-2mm}\end{ALigned}$$ $i$ IS determIned by The TarGet aTTrIbUte CoDE.
**TaRGeT AtTRibUte DiscrImInAtor:** SEparATED From $q$, we Set aN indePendent discriMinAtor $a$ To eValuaTe wheTher ThE one-bIt specIfic aTtRibute is correctLy geNerated or Not. thE poSiTive sAMple seT $\maThbF{R}_{a}$ consIsts of sAMplEs FROM tHe ground truth with tHe SPEcIfic attrIbute. THE nEgATive sampLe Set $\MathBF{f}_{a}$ are The sAMpLes generAted frOM $G_{A}$. THe targeT aTtribuTe DisCriMinatOR is uPdated As: $$\begin{aLigneD}
\Mathcal{L}_{A} = -\underSEt{\mathbf{U}^{*} \in \maTHbF{r}_{A}}{\mAThbb{e}} \loG(A(\mathbf{U}^{*})) - \unDersET{\matHbf{U} \IN \mAthBF{F}_{a}}{\maThbb{E}}\LoG(1-a(\mAThbf{U})))
\label{eq:3dagan_atTrD}
\vspacE{-2mm}\enD{aligned}$$ AccorDingly, the aDVERsarial lOss tO UpDAte the generatoR is: $$\beGin{aligned}
\MAthcal{L}_{AV}=- \undeRset{\mathBf{U} \in \mathBF{f}_{a}}{\mathbb{e}}\loG(A(\mAthBf{U}))
\LABeL{eq:3dagan_attraDV}
\VspaCe{-4Mm}\end{alIgnEd}$$
In TC-GaN, wE fiNd tHat ReConstructIon loss oThEr ThAn RecOgnitIOn percepTiOn lOsS is SuffiCIent to PreseRve iDeNtITy. IT also | **Masked Reconstruction Mo dule:** We manu all y d ef inethenon-attributea reashown in Fig. \[fig:ma sk\_t em p late \ ]on th e refer e nc e UVte xt ure m a p. Thos e $ 5$ attr ibutes are di vi ded into sev e ra l differen t m ask types or th eir co mb ina t ion,, li pstic ks and smileshare the s a me mas k of Fig . \[ fig: mask\_template\]( c) . Together with the f ul l yv i sib leone (maskof alle ntriesa s$ 1 $ ),w e define mask $\Omega_{i } , i =1,2,. ., 5$i ndicat ing a ll the categories . Th e reconst ructio n object i ve is a s belo w.$$\ begi n {a li gne d} \ma t hc al{ L }_{ ra} = \| (G _{ a}(\h at{\ m a t h bf{U }}_ {t}, \mat hbf{U}_{p}, \ mat hbf{ p })- \ha t{\ma thbf {U }}_{t })\odo t \Om eg a_{i}\|_1,
\la bel{ eq:3dagan _re co n}\v space { -2mm}\ end {al igned}$ $ $i$ i s de te r m i ne d by the target at tr i b ut e code.
**Tar g et A t tributeDi scr imin a t or:** Sep a ra ted from $Q$,w ese t an in de penden tdis cri minat o r $A $ to e valuatewheth e r the one-bits pecific attri b ut e is corr ect ly generate d or not. The po sit i ve sa mplese t $ \ mathbf{R}_{a}$ cons is ts ofsampl es from the g round trut h w ith thespec i fi c attribute. Th e neg ative samp l e set $\ mathb f{F}_{a} $ are the s amples g ene rat edfro m $G _{a}$. The ta r g et a tt ributedis crimina tor is up dat ed as: $$\b egin{ali gn ed }\m ath cal{L } _{A} = - \u nde rs et{ \math b f{U}^{ *} \i n \m at hb f {R} _{a}}{\ m at h b b{E} }\l og(A (\m at hbf{U }^{* } ))- \unde rset{\mat hbf { U} \ in \ mathbf{ F}_{a}}{\math bb {E}}\log(1 -A (\m athbf{ U } )))
\lab el{eq:3dagan_attrd}
\vs p ace{-2m m}\ end{a lign ed}$$ Acc ord ingly, th e adver sarial loss t o u p d ate t h e g ene ra tor is: $$ \ b egi n{ali gn ed}\mathca l{L}_{Av}=- \under s et{ \mathbf{U} \i n \ math b f {F }_{ a }} { \ma th b b{E } } \log(A(\mathbf{ U}))
\labe l{ e q: 3dagan_att r adv }\vspace {-4mm}\ end{a l igned}$ $
In TC- GAN, we f in d th a t re constructi on lossother tha n reco g ni tionper ceptio nlos s issuffic i ent to p reserv eidenti ty. I talso | **Masked Reconstruction_Module:** We_manually define the non-attribute_area shown_in_Fig. \[fig:mask\_template\] on_the_reference UV texture_map. Those $5$_attributes are divided into_several different mask_types_or their combination,, lipsticks and smile share the same mask of Fig. \[fig:mask\_template\] (c). Together_with_the fully_visible_one_(mask of all entries as_$1$), we define mask $\Omega_{i},_i=1,2,..,5$ indicating_all the categories. The reconstruction objective is as_below._$$\begin{aligned}
\mathcal{L}_{ra} = \|(G_{a}(\hat{\mathbf{U}}_{t},_\mathbf{U}_{p}, \mathbf{p}) - \hat{\mathbf{U}}_{t})\odot \Omega_{i}\|_1,
\label{eq:3dagan_recon}
\vspace{-2mm}\end{aligned}$$ $i$ is determined_by the target attribute code.
**Target Attribute_Discriminator:** Separated from_$Q$,_we_set an independent discriminator_$A$ to evaluate whether the one-bit_specific attribute is correctly generated or_not. The positive sample set $\mathbf{R}_{a}$ consists_of samples from the ground truth_with the specific attribute. The_negative sample_set $\mathbf{F}_{a}$ are the samples_generated from $G_{a}$._The target_attribute discriminator is_updated as: $$\begin{aligned}
\mathcal{L}_{A} = -\underset{\mathbf{U}^{*} \in_\mathbf{R}_{a}}{\mathbb{E}} \log(A(\mathbf{U}^{*})) -_\underset{\mathbf{U} \in \mathbf{F}_{a}}{\mathbb{E}}\log(1-A(\mathbf{U})))
\label{eq:3dagan_attrd}
\vspace{-2mm}\end{aligned}$$ Accordingly, the adversarial_loss_to update the_generator_is:_$$\begin{aligned}
\mathcal{L}_{Av}=- \underset{\mathbf{U}_\in \mathbf{F}_{a}}{\mathbb{E}}\log(A(\mathbf{U}))
\label{eq:3dagan_attradv}
\vspace{-4mm}\end{aligned}$$
In TC-GAN,_we_find that_reconstruction_loss other than recognition perception loss_is_sufficient to preserve identity. It also |
Wise
title: 'An Alignment Effect in FR I Radio Galaxies: U-band Polarimetry of the Abell 2597 Cluster Central Galaxy'
---
brmacro.tex psfig.sty
Introduction
============
The well known tendency for the blue optical continuum in distant, luminous radio galaxies to be aligned with their radio sources represents a potentially important evolutionary phase of radio galaxies. The blue optical continuum associated with the Fanaroff-Riley class two (FR II) radio sources in powerful radio galaxies often extends several tens of kiloparsecs into their halos along the axis of the radio source. The extended radio and optical continuum is usually associated with bright nebular emission. This so-called “alignment effect” becomes increasingly prominent in powerful radio galaxies at redshifts $z \gae 0.6$, but is rarely strong in radio galaxies at redshifts less than $z=0.1$ (see McCarthy 1993 for a review).
Distant, FR II radio sources are frequently bent and distorted in the vicinity of strong nebular emission. The nebular gas velocity fields are usually turbulent and shearing, with characteristic velocities of several hundreds of kilometers per second. Such properties seem to indicate that strong interactions (e.g. momentum exchange, photon heating and ionization) are occurring between the radio source and the nebular gas. Remarkably, processes associated with the radio source, which originates in a powerful nuclear engine, can affect the photometric properties of the host galaxy to large galactic distances.
Several emission mechanisms have been proposed to explain the extended, blue optical continuum including star formation, scattered light from an obliquely-directed active nucleus, synchrotron radiation, nebular continuum, and inverse Compton radiation (di Serego Alighieri 1989; Fabian 1989; De Young 1989; Rees 1989; Begelman & Cioffi 1989; Daly 1990, 1992; Dickson 1995). Although each of these emission mechanisms is interesting in its own right, collectively they imply essentially two important consequences for the evolution of radio galaxies. Star formation at the implied levels of several tens to over a thousand solar masses per year (e.g. 4C41.17, Dey 1997) would have a significant impact on the development of the stellar composition and structure of the host galaxy. The remaining emission mechanisms in general imply the presence of a powerful central engine, a dense interstellar medium, and strong magnetic fields, but may be otherwise inconsequential to the | Wise
title:' An Alignment Effect in FR I Radio Galaxies: U - band Polarimetry of the Abell 2597 Cluster Central Galaxy'
---
brmacro.tex psfig.sty
Introduction
= = = = = = = = = = = =
The well known tendency for the aristocratic ocular continuum in distant, luminous radio galax to be aligned with their radio informant represents a potentially important evolutionary phase of radio receiver galaxy. The blue ocular continuum associated with the Fanaroff - Riley class two (FR II) radio receiver sources in powerful radio galaxies often extends several ten of kiloparsecs into their halos along the axis of the radio reservoir. The extended radio and optical continuum is usually associate with bright nebular emission. This so - called “ alliance effect ” becomes increasingly prominent in powerful radio galaxies at redshifts $ z \gae 0.6 $, but is rarely strong in radio galaxies at redshifts less than $ z=0.1 $ (see McCarthy 1993 for a review).
Distant, FR II radio sources are frequently bended and falsify in the vicinity of strong cloudlike discharge. The cloudlike gas velocity fields are usually turbulent and shearing, with characteristic velocities of several hundreds of kilometer per second. Such properties seem to indicate that strong interactions (e.g. momentum exchange, photon heating and ionization) are occurring between the radio source and the cloudlike gas. signally, procedure associated with the radio source, which originates in a powerful nuclear engine, can affect the photometric place of the host galaxy to large galactic distance.
Several discharge mechanisms have been proposed to excuse the extended, blue optical continuum including ace formation, scatter light from an obliquely - address active nucleus, synchrotron radiotherapy, nebular continuum, and inverse Compton radiation (di Serego Alighieri 1989; Fabian 1989; De Young 1989; Rees 1989; Begelman & Cioffi 1989; Daly 1990, 1992; Dickson 1995). Although each of these emission mechanisms is interesting in its own right, collectively they imply essentially two important consequence for the evolution of radio galaxy. Star formation at the implied levels of several tens to over a thousand solar bulk per year (e.g. 4C41.17, Dey 1997) would have a significant impact on the development of the leading composition and structure of the horde galaxy. The remaining emission mechanisms in general imply the bearing of a powerful central engine, a dense interstellar medium, and strong magnetic fields, but may be otherwise inconsequential to the | Wide
title: 'An Alignment Efftct in FR I Radio Galaxixs: U-bans Polarioetry of the Abell 2597 Cluster Renteal Gqlaxy'
---
brmacro.tex psfig.sgy
Introdubtion
============
The qell jnown tendxhcy for the bmme opciral continuum ik distant, lgminous radio cauaries to be aligned with their radio fources rfpresents a pojentisjly jmportant evolutionary phase of rasio galexies. The blue pptical continuum associatfd wlth the Fanaroff-Ripey class tqo (FW II) radio sojrces in powerful radik galaxies often extends severau tenx of kilopqrwecd into their halof along the axis of dhe radoo source. The cxtenved eadio and optical convinuum is usually asfociated fich bright nebular emiwsuon. Tvis vo-cauoed “aljgimeht efffct” becomes ihcreasingly prominent in powerguj radio galaxiss at weqshifts $z \gae 0.6$, but is rarely strong in gadik galaxies at redshifts less than $z=0.1$ (see McCatthy 1993 for w review).
Distant, FR II radio sources are frequentlf benv xnd dlstofref in the vicinity of strong nebular emission. Ege nvbular gas velociby fields are usuaklj yorbulent and sfeariny, wjth characteristic velocijies od several huncreds of kilometers per secind. Such proiertues seem to indicace that stroug intgractipns (e.g. momentum exchangz, photkn heating wnd ionizzgion) are occurrivg netfeen the radio source and ehe nebuler gax. Remaryabli, procefses assoclated with the radio sougce, wkich mriginates in a powerful nuclear engine, cei affect the khodomvtric pro'ertiex of the host galaxy to larye galaccic dirtances.
Sevvral emisvion mechanysms have beet proposed to explain the extwnded, buje optical conyinuum inblbding stae formation, scatteved lkfht from an oblndowly-directed acyivd ntcpens, sygwhrotron radhatiun, vrbulaf continuum, xnd onverse Compton radidtioh (di Serego Alighirrl 1989; Fabian 1989; De Yougg 1989; Rees 1989; Begrlman & Cioffi 1989; Dalj 1990, 1992; Dmckson 1995). Altnoudh each of these emission mechznisms is inberesting in yts iwn right, coklectively they imply essentially two inportant consequenxes for the evolutipk of radio gelaxief. Star fosmation at the impliwd levels of seveval tens to over a thohsand volar masses per year (e.g. 4C41.17, Dey 1997) would have a significant impact on the developnent oh ehe stellar comlositpon cnd strtctuce of the host galsxy. The remaining emission mechenisms in ceueral imply the presence of a puwerful centrxl engine, a dense intersfellar mrdium, and strong magnetic fieldx, but may be otherwise ikconsequebtial to thv | Wise title: 'An Alignment Effect in FR Galaxies: Polarimetry of Abell 2597 Cluster Introduction The well known for the blue continuum in distant, luminous radio galaxies be aligned with their radio sources represents a potentially important evolutionary phase of galaxies. The blue optical continuum associated with the Fanaroff-Riley class two (FR II) sources powerful galaxies extends several tens of kiloparsecs into their halos along the axis of the radio source. The radio and optical continuum is usually associated with nebular emission. This so-called effect” becomes increasingly prominent in radio at redshifts \gae but rarely strong in galaxies at redshifts less than $z=0.1$ (see McCarthy 1993 for a review). Distant, FR II radio sources frequently bent in the of nebular The nebular gas are usually turbulent and shearing, with several hundreds of kilometers per second. Such properties to indicate strong interactions (e.g. momentum exchange, photon and ionization) are occurring between the radio source the nebular gas. Remarkably, processes associated with the radio source, which originates in a powerful can affect the photometric of the host to galactic Several mechanisms have proposed to explain the extended, blue optical continuum including star formation, light from an obliquely-directed active nucleus, synchrotron radiation, nebular continuum, Compton (di Serego Alighieri Fabian 1989; De Young Rees Begelman & Cioffi 1989; 1992; 1995). these mechanisms interesting in its own collectively they imply essentially two consequences for the evolution at the implied levels of several tens to a thousand solar masses per year (e.g. Dey 1997) would have a significant impact on the development of the composition and the host galaxy. The remaining emission mechanisms in imply the presence of powerful central engine, a dense interstellar medium, and strong fields, may be inconsequential to the | Wise
title: 'An Alignment EffecT in FR I RadiO GalaXieS: U-bAnD PolArimEtry of the Abell 2597 cLustEr Central Galaxy'
---
brmacro.Tex psFiG.Sty
INTrOductIon
============
The wELl KNOwn TeNdEncY fOR tHe bluE opTical coNtinuum in dIstAnT, luminous radIO gAlaxies to bE alIgned with theIr rAdio soUrCes REpresEntS a potEntialLY imporTant evoluTiONary phASe of radIO GaLaxiEs. The blue optical cONtINuum associated With thE FANaROFf-RIleY class two (Fr Ii) radiO Sources IN pOWERfuL Radio galaxies Often extendS SevEral teNs Of kILoparsEcs inTo THeiR halos along The aXis of the rAdio soURce. The eXTended rAdio anD opTicAl coNTiNuUm iS uSUalLY aSsoCIatEd with brIgHt NebulAr emISSIOn. ThIs sO-calLed “alIgnment effect” BecOmes INcrEasinGly prOminEnT in poWerful Radio GaLaxies at redshifTs $z \gAe 0.6$, but is raRelY sTroNg In radIO galaxIes At rEdshiftS less thAN $z=0.1$ (sEe mCcArThy 1993 for a review).
DistaNt, fr iI Radio souRces arE FrEqUEntly benT aNd dIstoRTEd in tHe viCInIty of strOng nebULaR eMission. thE nebulAr Gas VelOcity FIeldS are usUally turBulenT And shearing, witH CharacteristiC VeLOCiTIes oF seVeral hundreDs of KIlomEterS PeR seCOnd. SuCh proPeRTiES seem to indicate that StRong inTeracTions (e.g. momentUm exchange, PHOTon heatiNg anD IoNIzation) are occuRring Between the RAdio sourCe and The nebulAr gas. RemaRKAbly, procEssEs aSsoCiaTED wIth the radio soURCe, whIcH originAteS in a powErfUl nUclEar EnGine, can afFect the pHoToMeTrIc pRoperTIes of the HoSt gAlAxy To larGE galacTic diStanCeS.
SEVerAl emissIOn MEChanIsMs Have BeeN pRoposEd to EXplAin the eXtended, blUe oPTicaL cOnTinuum iNcluding star fOrMation, scatTeRed Light fROM an obliqUely-directed active nucleUS, synchrOtrOn radIatiOn, nebular ConTinuum, And INverse comptoN radiAtIon (DI seregO aLiGhiErI 1989; Fabian 1989; De YOUNg 1989; REes 1989; BeGeLman & cioffi 1989; DAly 1990, 1992; Dickson 1995). Although EAch Of these emissiOn mEchaNISmS is INtEResTiNG in ITS own right, collecTively they ImPLy EssentiallY Two ImPortant ConsequEnces FOr the evOlution of Radio galaXiEs. StAR ForMation at thE implied Levels of sEVeral TEnS to ovEr a ThousaNd SolAr masSes per YEar (E.g. 4C41.17, DeY 1997) would HaVe a sigNificAnT impact oN the development of the steLlar coMposiTioN and strucTurE Of tHe host galAxy. THe remaininG emIssIon meChaNIsms iN genERaL imPLy the PresENce of a powERfUl cENTrAl engine, a deNSE IntErsteLlaR Medium, And sTrong magnetic fielDS, but may be otherWise INConSeqUEntiAl To the | Wise
title: 'An Alignment Effect in FR I Ra dio G alax ies: U-band Polari m etry of the Abell 2597 Clu sterCe n tral Ga laxy'
---
b r ma c r o.t ex p sfi g. s ty
Int rod uction========== ==
T he well know n t endency fo r t he blue opti cal conti nu umi n dis tan t, lu minous radiogalaxiesto be ali g ned wit h th eirradio sources rep r es e nts a potentia lly im po r ta n t ev olu tionary ph as e ofr adio ga l ax i e s . T h e blue optica l continuum ass ociate dwit h the F anaro ff - Ril ey class tw o (F R II) rad io sou r ces inp owerful radio ga lax ieso ft en ex te n dss ev era l te ns of ki lo pa rsecs int o t h eirhal os a longthe axis of t heradi o so urce. Theexte nd ed ra dio an d opt ic al continuum is usu ally asso cia te d w it h bri g ht neb ula r e mission . Thiss o-c al l e d “ alignment effect”be c o me s increa singly pr om i nent inpo wer fulr a dio g alax i es at reds hifts$ z\g ae 0.6$ ,but is r are lystron g inradiogalaxies at r e dshifts less t h an $z=0.1$ (s e eM c Ca r thy199 3 for a rev iew) .
Di stan t ,FRI I rad io so ur c es are frequently bent a nd dis torte d in the vici nity of st r o n g nebula r em i ss i on. The nebula r gas velocityf ields ar e usu ally tur bulent an d shearing , w ith ch ara c t er istic velocit i e s of s everalhun dreds o f k ilo met ers p er second . Such p ro pe rt ie s s eem t o indicat etha tstr ong i n teract ions(e.g .mo m ent um exch a ng e , pho to nheat ing a nd io niza t ion ) are o ccurringbet w eenth eradio s ource and the n ebular gas .Rem arkabl y , process es associated with ther adio so urc e, wh ichoriginate s i n a po wer f ul nuc lear e ngine ,can a ffect t he ph ot ometric pr o p ert ies o fthehost ga laxy to large gala c tic distances.
Sev eral e mi ssi o nm ech an i sms h ave been propos ed to expl ai n t he extende d , b lu e optic al cont inuum includi ng star f ormation, s catt e r edlight from an obli quely-dir e cteda ct ive n ucl eus, s yn chr otron radia t ion , neb ular c on tinuum , and i nverse C ompton radiation (di Se rego A lighi eri 1989; Fa bia n 19 89; De Yo ung1989; Rees 19 89; Bege lma n & Ci offi 19 89; Daly1990 , 1992; Di c ks on1 99 5). Althoug h e ach of t hes e emiss ionmechanisms is int e resting in its own r igh t,c olle ct ively they imp lyes s e ntiallytw o important consequ en c es fo r theevolut ion ofr a di o galax ies. St ar format ion a t the im pl ie d level s of s everal tenst o ov e r a thousand sola r mas s e s per yea r (e. g. 4C41.1 7 , De y 1997) w ould have a signi fica nt im pact on t he dev elo pm ent of the stellar c ompos ition a nd str uct ure of the h ost g alax y. Th e remaini n g e m is si o n m echa nisms i n ge neral imp l y the pr ese n ce of a p owe r f ul cen t ra l engine, aden se in t e rstellar m e dium , a n d str ong ma gnetic fields , bu tmay beoth e r wise inco nsequenti a l t othe | Wise
title:_'An Alignment_Effect in FR I_Radio Galaxies:_U-band_Polarimetry of_the_Abell 2597 Cluster_Central Galaxy'
---
brmacro.tex psfig.sty
Introduction
============
The_well known tendency for_the blue optical_continuum_in distant, luminous radio galaxies to be aligned with their radio sources represents a_potentially_important evolutionary_phase_of_radio galaxies. The blue optical_continuum associated with the Fanaroff-Riley_class two_(FR II) radio sources in powerful radio galaxies_often_extends several tens_of kiloparsecs into their halos along the axis of_the radio source. The extended radio_and optical continuum_is_usually_associated with bright nebular_emission. This so-called “alignment effect” becomes_increasingly prominent in powerful radio galaxies_at redshifts $z \gae 0.6$, but is_rarely strong in radio galaxies at_redshifts less than $z=0.1$ (see_McCarthy 1993_for a review).
Distant, FR II_radio sources are_frequently bent_and distorted in_the vicinity of strong nebular emission._The nebular gas_velocity fields are usually turbulent and_shearing,_with characteristic velocities_of_several_hundreds of_kilometers per second._Such_properties seem_to_indicate that strong interactions (e.g. momentum_exchange,_photon heating and ionization) are occurring between_the radio source and_the_nebular gas. Remarkably, processes_associated with the radio source,_which originates in a powerful nuclear_engine, can_affect the_photometric properties of the host galaxy to large galactic distances.
Several emission_mechanisms have been proposed to explain_the extended, blue optical_continuum including_star_formation, scattered light_from_an obliquely-directed_active nucleus, synchrotron radiation, nebular continuum, and_inverse Compton_radiation (di Serego Alighieri 1989; Fabian_1989; De Young 1989;_Rees_1989; Begelman & Cioffi 1989; Daly_1990, 1992; Dickson 1995). Although each_of these emission mechanisms is_interesting_in_its own right, collectively they_imply essentially two important consequences for_the evolution of_radio galaxies. Star formation at the implied_levels_of several tens to over a_thousand_solar masses per year (e.g. 4C41.17,_Dey_ 1997)_would have a significant impact_on the development of the stellar_composition and structure of the host galaxy. The remaining_emission mechanisms in_general imply the presence of_a_powerful_central engine, a dense interstellar medium, and strong magnetic fields,_but may_be otherwise inconsequential_to the |
20) ( 30,-20)
(39,2)[(-1,2)[8]{}]{} (39,-2)[(-1,-2)[8]{}]{} (29,18)[(-1,-2)[8]{}]{} (29,-18)[(-1,2)[8]{}]{} (33,18)[(3,-2)[24.5]{}]{} (33,-18)[(3,2)[24.5]{}]{} (2.3,-1.4)[(3,-2)[25.6]{}]{} (2.3,1.4)[(3,2)[25.6]{}]{}
(15,14)[(0,0)[$P$]{}]{} (15,-14)[(0,0)[$p$]{}]{} (45,14)[(0,0)[$S$]{}]{} (45,-14)[(0,0)[$s$]{}]{}
(21,9)[(0,0)[$Q$]{}]{} (21,-9)[(0,0)[$q$]{}]{} (39,9)[(0,0)[$R$]{}]{} (39,-9)[(0,0)[$r$]{}]{}
\[r index shift invariance\] Theorem \[t pairs commute simple\] requires showing that $g_Sh_S =
h_Sg_S$ for any $S \subset \{1,\ldots, N\}$ assuming certain relations. Though the proof via Theorem \[t pairs commute simple\] only shows this in the case $S = \{1,\ldots, N\}$, this easily shows $g_Sh_S =
h_Sg_S$ for general $S = \{s_1 < \cdots < s_m\}$: set $p_i = g_{s_i}$ and $q_i = h_{s_i}$; then $p_1,\ldots, p_m, q_1, \ldots, q_m$ satisfy the relations,, (with $p$, $q$, $m$ in place of $g$, $h$, $N$, respectively), and hence $p_m \cdots p_1 q_m \cdots q_1 = q_m \cdots q_1 p_m \cdots
p_1$.
Many other results in this paper require showing that $g_Sh_S =
h_Sg_S$ given certain relations. Just as above, we only show the case $S = \{1,\ldots, N\}$ since this easily implies the general case.
\[thm:group-the | 20) (30,-20)
(39,2)[(-1,2)[8 ] { } ] { } (39,-2)[(-1,-2)[8 ] { } ] { } (29,18)[(-1,-2)[8 ] { } ] { } (29,-18)[(-1,2)[8 ] { } ] { } (33,18)[(3,-2)[24.5 ] { } ] { } (33,-18)[(3,2)[24.5 ] { } ] { } (2.3,-1.4)[(3,-2)[25.6 ] { } ] { } (2.3,1.4)[(3,2)[25.6 ] { } ] { }
(15,14)[(0,0)[$P$ ] { } ] { } (15,-14)[(0,0)[$p$ ] { } ] { } (45,14)[(0,0)[$S$ ] { } ] { } (45,-14)[(0,0)[$s$ ] { } ] { }
(21,9)[(0,0)[$Q$ ] { } ] { } (21,-9)[(0,0)[$q$ ] { } ] { } (39,9)[(0,0)[$R$ ] { } ] { } (39,-9)[(0,0)[$r$ ] { } ] { }
\[r index shift invariance\ ] Theorem \[t pairs commute simple\ ] requires showing that $ g_Sh_S =
h_Sg_S$ for any $ south \subset \{1,\ldots, N\}$ assume certain relative. Though the validation via Theorem \[t pairs commute simple\ ] only shows this in the case $ S = \{1,\ldots, N\}$, this well show $ g_Sh_S =
h_Sg_S$ for general $ S = \{s_1 < \cdots < s_m\}$: sic $ p_i = g_{s_i}$ and $ q_i = h_{s_i}$; then $ p_1,\ldots, p_m, q_1, \ldots, q_m$ satisfy the relations, , (with $ p$, $ q$, $ m$ in position of $ g$, $ h$, $ N$, respectively), and hence $ p_m \cdots p_1 q_m \cdots q_1 = q_m \cdots q_1 p_m \cdots
p_1$.
Many other solution in this newspaper require showing that $ g_Sh_S =
h_Sg_S$ given certain relations. merely as above, we only show the case $ S = \{1,\ldots, N\}$ since this well implies the general font.
\[thm: group - the | 20) ( 30,-20)
(39,2)[(-1,2)[8]{}]{} (39,-2)[(-1,-2)[8]{}]{} (29,18)[(-1,-2)[8]{}]{} (29,-18)[(-1,2)[8]{}]{} (33,18)[(3,-2)[24.5]{}]{} (33,-18)[(3,2)[24.5]{}]{} (2.3,-1.4)[(3,-2)[25.6]{}]{} (2.3,1.4)[(3,2)[25.6]{}]{}
(15,14)[(0,0)[$P$]{}]{} (15,-14)[(0,0)[$p$]{}]{} (45,14)[(0,0)[$S$]{}]{} (45,-14)[(0,0)[$s$]{}]{}
(21,9)[(0,0)[$Q$]{}]{} (21,-9)[(0,0)[$q$]{}]{} (39,9)[(0,0)[$R$]{}]{} (39,-9)[(0,0)[$r$]{}]{}
\[r inaex shift invarnqnce\] Tieorem \[t pairs cummute simple\] requires showiig tyat $g_Wh_S =
h_Sg_S$ for any $S \subret \{1,\ldots, N\}$ assumung rertain relations. Though the pdlof rie Theorem \[t pairx commute vimple\] only shmwr chis in the case $S = \{1,\ldots, N\}$, this easyly shoes $g_Sh_S =
h_Sg_S$ for gentraj $S = \{s_1 < \cdots < s_m\}$: set $p_i = g_{s_i}$ and $q_i = g_{s_i}$; thei $p_1,\ldots, p_m, q_1, \lcots, q_m$ satisfy the relatilns,, (aith $p$, $q$, $m$ in placf of $g$, $h$, $N$, tsspqxtively), and fence $p_m \cdots p_1 q_m \cdkts q_1 = q_m \cdots q_1 p_m \cdots
p_1$.
Many othex results ib rhid paper requmre shjwing that $g_Sh_S =
h_Sg_S$ given vertain relatipns. Juwt as above, we only siow the case $S = \{1,\ldotf, N\}$ since tkis easily implies thw tenerdl cdse.
\[tfn:gruup-uhe | 20) ( 30,-20) (39,2)[(-1,2)[8]{}]{} (39,-2)[(-1,-2)[8]{}]{} (29,18)[(-1,-2)[8]{}]{} (29,-18)[(-1,2)[8]{}]{} (2.3,-1.4)[(3,-2)[25.6]{}]{} (15,14)[(0,0)[$P$]{}]{} (15,-14)[(0,0)[$p$]{}]{} (45,-14)[(0,0)[$s$]{}]{} (21,9)[(0,0)[$Q$]{}]{} (21,-9)[(0,0)[$q$]{}]{} invariance\] \[t pairs commute requires showing that = h_Sg_S$ for any $S \subset N\}$ assuming certain relations. Though the proof via Theorem \[t pairs commute simple\] shows this in the case $S = \{1,\ldots, N\}$, this easily shows $g_Sh_S h_Sg_S$ general = < \cdots < s_m\}$: set $p_i = g_{s_i}$ and $q_i = h_{s_i}$; then $p_1,\ldots, p_m, q_1, q_m$ satisfy the relations,, (with $p$, $q$, $m$ place of $g$, $h$, respectively), and hence $p_m \cdots q_m q_1 = \cdots p_m p_1$. Many other in this paper require showing that $g_Sh_S = h_Sg_S$ given certain relations. Just as above, we only the case \{1,\ldots, N\}$ this implies general case. \[thm:group-the | 20) ( 30,-20)
(39,2)[(-1,2)[8]{}]{} (39,-2)[(-1,-2)[8]{}]{} (29,18)[(-1,-2)[8]{}]{} (29,-18)[(-1,2)[8]{}]{} (33,18)[(3,-2)[24.5]{}]{} (33,-18)[(3,2)[24.5]{}]{} (2.3,-1.4)[(3,-2)[25.6]{}]{} (2.3,1.4)[(3,2)[25.6]{}]{}
(15,14)[(0,0)[$P$]{}]{} (15,-14)[(0,0)[$p$]{}]{} (45,14)[(0,0)[$S$]{}]{} (45,-14)[(0,0)[$s$]{}]{}
(21,9)[(0,0)[$Q$]{}]{} (21,-9)[(0,0)[$q$]{}]{} (39,9)[(0,0)[$R$]{}]{} (39,-9)[(0,0)[$r$]{}]{}
\[r index shift invariAnce\] TheoreM \[t paiRs cOmmUtE simPle\] rEquires showing THat $g_sh_S =
h_Sg_S$ for any $S \subset \{1,\ldOts, N\}$ aSsUMing CErTain rElationS. thOUGh tHe PrOof ViA thEorem \[T paIrs commUte simple\] oNly ShOws this in the CAsE $S = \{1,\ldots, N\}$, thIs eAsily shows $g_SH_S =
h_sg_S$ for GeNerAL $S = \{s_1 < \cdOts < S_m\}$: set $P_i = g_{s_i}$ aND $q_i = h_{s_i}$; Then $p_1,\ldotS, p_M, Q_1, \ldots, Q_M$ satisfY THe RelaTions,, (with $p$, $q$, $m$ in plaCE oF $G$, $h$, $N$, respectivelY), and heNcE $P_m \CDOts P_1 q_m \Cdots q_1 = q_m \cdOtS q_1 p_m \cDOts
p_1$.
ManY OtHER ResULts in this papeR require shoWIng That $g_SH_S =
H_Sg_s$ Given cErtaiN rELatIons. Just as aBove, We only shoW the caSE $S = \{1,\ldots, n\}$ Since thIs easiLy iMplIes tHE gEnEraL cASe.
\[tHM:gRouP-The | 20) ( 30,-20)
(39,2)[(-1, 2)[8]{}]{} (39, -2) [(- 1, -2)[ 8]{} ]{} (29,18)[(- 1 ,-2) [8]{}]{} (29,-18)[(-1, 2)[8] {} ] {} ( 3 3, 18)[( 3,-2)[2 4 .5 ] { }]{ }(3 3,- 18 ) [( 3,2)[ 24. 5]{}]{} (2.3,-1.4 )[( 3, -2)[25.6]{}] { }(2.3,1.4)[ (3, 2)[25.6]{}]{ }
(15,14 )[ (0, 0 )[$P$ ]{} ]{} ( 15,-14 ) [(0,0) [$p$]{}]{ }( 45,14) [ (0,0)[$ S $ ]{ }]{} (45,-14)[(0,0)[$ s $] { }]{}
(21,9)[( 0,0)[$ Q$ ] {} ] { } ( 21, -9)[(0,0)[ $q $]{}] { } (39,9 ) [( 0 , 0 )[$ R $]{}]{} (39,- 9)[(0,0)[$r $ ]{} ]{}
\ [r in d ex shi ft in va r ian ce\] Theore m \[ t pairs c ommute simple\ ] requir es sho win g t hat$ g_ Sh _S=h _Sg _ S$ fo r an y $S \su bs et \{1, \ldo t s , N\}$ as sumi ng ce rtain relatio ns. Tho u ghthe p roofviaTh eorem \[t p airsco mmute simple\]only shows th isin th ecase$ S = \{ 1,\ ldo ts, N\} $, this eas il y s ho ws $g_Sh_S =
h_Sg_ S$ f or general $S =\ {s _1 < \cdots < s_ m\}$ : set $ p_i= g _{s_i}$and $q _ i=h_{s_i} $; then$p _1, \ld ots,p _m,q_1, \ ldots, q _m$ s a tisfy the rela t ions,, (with$ p$ , $q $ , $m $ i n place of$g$, $h$, $N$ , r esp e ctive ly),an d h e nce $p_m \cdots p_1 q _m \cd ots q _1 = q_m \cdo ts q_1 p_m \ c dots
p_1 $.
M an y other results in t his paperr equire s howin g that $ g_Sh_S =h _ Sg_S$ gi ven ce rta inr e la tions. Just a s abov e, we onl y s how the ca se$S= \ {1 ,\ldots,N\}$ sin ce t hi seas ily i m plies th egen er alcase.
\[thm :grou p-th e | 20) (_30,-20)
(39,2)[(-1,2)[8]{}]{} (39,-2)[(-1,-2)[8]{}]{}_(29,18)[(-1,-2)[8]{}]{} (29,-18)[(-1,2)[8]{}]{} (33,18)[(3,-2)[24.5]{}]{} (33,-18)[(3,2)[24.5]{}]{}_(2.3,-1.4)[(3,-2)[25.6]{}]{} (2.3,1.4)[(3,2)[25.6]{}]{}
(15,14)[(0,0)[$P$]{}]{}_(15,-14)[(0,0)[$p$]{}]{}_(45,14)[(0,0)[$S$]{}]{} (45,-14)[(0,0)[$s$]{}]{}
(21,9)[(0,0)[$Q$]{}]{}_(21,-9)[(0,0)[$q$]{}]{}_(39,9)[(0,0)[$R$]{}]{} (39,-9)[(0,0)[$r$]{}]{}
\[r index_shift invariance\] Theorem \[t_pairs commute simple\] requires_showing that $g_Sh_S_=
h_Sg_S$_for any $S \subset \{1,\ldots, N\}$ assuming certain relations. Though the proof via Theorem \[t_pairs_commute simple\]_only_shows_this in the case $S_= \{1,\ldots, N\}$, this easily_shows $g_Sh_S_=
h_Sg_S$ for general $S = \{s_1 < \cdots_<_s_m\}$: set $p_i_= g_{s_i}$ and $q_i = h_{s_i}$; then $p_1,\ldots, p_m,_q_1, \ldots, q_m$ satisfy the relations,,_(with $p$, $q$,_$m$_in_place of $g$, $h$,_$N$, respectively), and hence $p_m \cdots_p_1 q_m \cdots q_1 = q_m_\cdots q_1 p_m \cdots
p_1$.
Many other results in_this paper require showing that $g_Sh_S_=
h_Sg_S$ given certain relations. Just_as above,_we only show the case_$S = \{1,\ldots,_N\}$ since_this easily implies_the general case.
\[thm:group-the |
of branching random field, *Proc. Natl. Acad. Sci.*, Vol. 76, No. 2, pp 872–875, 1979.
[^1]: University of Chicago
[^2]: University of Chicago
[^3]: At any rate the argument is routine in the case where $f (1)$ and $g (1)$ take values in the unit interval $[0,1]$; in this case existence follows by a routine phase-portrait analysis for the associated first-order system, using the nonnegativity of the forcing term $\Psi (h)$. When $f (1)$ and $g (1)$ are complex-valued, however, other methods must be used. See the proof of Lemma \[lemma.analytic-continuation\] in section \[sec:number-killed\] below for a proof in the case needed for the theorems on the distribution of the number of killed particles.
---
abstract: 'The effect of various velocity boundary condition is studied in two-dimensional Rayleigh-Bénard convection. Combinations of no-slip, stress-free and periodic boundary conditions are used on both the sidewalls and the horizontal plates. For the studied Rayleigh numbers ${\text{Ra}}$ between $10^8$ and $10^{11}$ the heat transport is lower for $\Gamma = 0.33$ than for $\Gamma = 1$ in case of no-slip sidewalls. This is surprisingly opposite for stress-free sidewalls, where the heat transport increases for lower aspect-ratio. In wider cells the aspect-ratio dependence is observed to disappear for ${\text{Ra}}\ge 10^{10}$. Two distinct flow types with very different dynamics can be seen, mostly dependent on the plate velocity boundary condition, namely roll-like flow and horizontal zonal flow, which have a substantial effect on the dynamics and heat transport in the system. The predominantly horizontal zonal flow suppresses heat flux and is observed for stress-free and asymmetric plates. Low aspect-ratio periodic sidewall simulations with a no-slip boundary condition on the plates also exhibit zonal flow. In all the other cases, the flow is roll-like. In two-dimensional Rayleigh-Bénard convection, the velocity boundary conditions thus have large implications on both roll-like and zonal flow that have to be taken into consideration before the boundary conditions are imposed.'
author:
- 'Erwin P. van der Poel$^{1}$, Rodol | of branching random field, * Proc. Natl. Acad. Sci. *, Vol. 76, No. 2, pp 872–875, 1979.
[ ^1 ]: University of Chicago
[ ^2 ]: University of Chicago
[ ^3 ]: At any rate the controversy is everyday in the case where $ f (1)$ and $ g (1)$ bring value in the unit interval $ [ 0,1]$; in this character universe follows by a everyday phase - portrayal analysis for the associated beginning - order system, using the nonnegativity of the forcing condition $ \Psi (h)$. When $ f (1)$ and $ g (1)$ are complex - valued, however, other methods must be used. attend the proof of Lemma \[lemma.analytic - continuation\ ] in section \[sec: number - killed\ ] under for a proof in the case needed for the theorems on the distribution of the act of killed particles.
---
abstract:' The effect of various velocity limit condition is studied in two - dimensional Rayleigh - Bénard convection. Combinations of no - slip, stress - free and periodic boundary conditions are use on both the sidewall and the horizontal plates. For the studied Rayleigh numbers $ { \text{Ra}}$ between $ 10 ^ 8 $ and $ 10^{11}$ the heat transport is lower for $ \Gamma = 0.33 $ than for $ \Gamma = 1 $ in case of no - slip sidewall. This is surprisingly opposite for tension - spare sidewall, where the heat transport increase for lower aspect - ratio. In wider cell the aspect - ratio dependence is observed to disappear for $ { \text{Ra}}\ge 10^{10}$. Two distinct stream types with very different dynamics can be seen, largely dependent on the home plate velocity boundary condition, namely roll - like menstruation and horizontal zonal flow, which have a substantial consequence on the dynamics and heat transport in the system. The predominantly horizontal zonal stream suppresses heat flux and is observe for stress - complimentary and asymmetric plate. Low expression - ratio periodic sidewall simulations with a no - slip boundary condition on the plates besides exhibit zonal flow. In all the other case, the flow is roll - like. In two - dimensional Rayleigh - Bénard convection, the speed boundary conditions thus have large deduction on both roll - like and zonal flow that have to be taken into consideration before the boundary condition are imposed.'
author:
-' Erwin P. van der Poel$^{1}$, Rodol | of branching random field, *Kroc. Natl. Acad. Scn.*, Vol. 76, Io. 2, pp 872–875, 1979.
[^1]: Univerrity of Chicago
[^2]: University oh Chucago
[^3]: At any rate the argumdnt is roltine in rhe rase where $f (1)$ anv $g (1)$ takc valhcs in vhe unit intervsl $[0,1]$; in thiv case existenwe flllows by a routine phase-portrait agalysis flr the associajed fprft-orsvr system, using the nonnegativify of tie forcing term $\Psi (h)$. When $f (1)$ and $g (1)$ are fomppex-valued, however, lther methoes mtwt be used. Sde the proof of Lemma \[lgmma.analytic-continuation\] in sectkon \[see:number-killgb\] bepmw for a priof ig the case ncvded fos the tneorems on the divtrubution of the number of killed particlgs.
---
abstracd: 'Che effect of various vwlocijy bogndaft cundjtmon is stkdixd in two-dijensional Rqyleigh-Bénard convecuiog. Combinations of no-flyp, stress-free and periodic boundary congitjons are used on both tye sidewalls and the jorizontaj plates. For the studied Rayleigh numbers ${\text{Ra}}$ tetwexn $10^8$ aue $10^{11}$ thd hfat transport is lower for $\Gamma = 0.33$ than for $\Gwjms = 1$ in case of nj-slip sidewslps. Jhis is surpriringly oplosite for stress-fgee sidgwalls, where tht heay transport increases for liwer aspect-rctii. In wider cells tke aspect-ratno depgndencr is observed to disappzar fod ${\text{Ra}}\ge 10^{10}$. Two distjvct flow types wkth vary difftfent dynamics can be seen, nostky depevdenj on thq plate vepocitn boundary conditioj, namgly ronl-like floa and horizontal zonal flow, whiri have a subsjandian effect on tme dynamics and heat transporj in the fysteo. The predkminantny horizontwl zonal flow duppresses hxat flux wnd us ovserved wor stress-free and asymmetric plares. Low aspect-ratip pgrjodic sidewall wimooations with a no-rli[ honndarr condition ot thd pustes xlso exhibib zunal flow. In all the othar czses, the flow is rplj-like. In two-dimegsional Rayleogh-Bénard convectioj, the veuocitu bjundary conditions thus have lzrge impllcabions on both rolo-like and zoual flow that have to be taken into consmderation before the boyndary conditions ate lmposed.'
authoc:
- 'Erwig P. van dar Poel$^{1}$, Rodol | of branching random field, *Proc. Natl. Acad. 76, 2, pp 1979. [^1]: University Chicago At any rate argument is routine the case where $f (1)$ and (1)$ take values in the unit interval $[0,1]$; in this case existence follows a routine phase-portrait analysis for the associated first-order system, using the nonnegativity of forcing $\Psi When (1)$ and $g (1)$ are complex-valued, however, other methods must be used. See the proof of \[lemma.analytic-continuation\] in section \[sec:number-killed\] below for a proof the case needed for theorems on the distribution of number killed particles. abstract: effect various velocity boundary is studied in two-dimensional Rayleigh-Bénard convection. Combinations of no-slip, stress-free and periodic boundary conditions are used on the sidewalls horizontal plates. the Rayleigh ${\text{Ra}}$ between $10^8$ the heat transport is lower for than for $\Gamma = 1$ in case of sidewalls. This surprisingly opposite for stress-free sidewalls, where heat transport increases for lower aspect-ratio. In wider the aspect-ratio dependence is observed to disappear for ${\text{Ra}}\ge 10^{10}$. Two distinct flow types with dynamics can be seen, dependent on the velocity condition, roll-like and horizontal flow, which have a substantial effect on the dynamics and heat in the system. The predominantly horizontal zonal flow suppresses heat is for stress-free and plates. Low aspect-ratio periodic simulations a no-slip boundary condition plates exhibit all other the flow is roll-like. two-dimensional Rayleigh-Bénard convection, the velocity conditions thus have large zonal flow that have to be taken into before the boundary conditions are imposed.' author: 'Erwin P. van der Poel$^{1}$, Rodol | of branching random field, *ProC. Natl. Acad. SCi.*, Vol. 76, no. 2, pP 872–875, 1979.
[^1]: UnIvErsiTy of chicago
[^2]: UniversITy of chicago
[^3]: At any rate the argUment Is ROutiNE iN the cAse wherE $F (1)$ aND $G (1)$ taKe VaLueS iN ThE unit IntErval $[0,1]$; in This case exIstEnCe follows by a ROuTine phase-pOrtRait analysis For The assOcIatED firsT-orDer syStem, usINg the nOnnegativItY Of the fORcing teRM $\psI (h)$. WhEn $f (1)$ and $g (1)$ are complex-VAlUEd, however, other MethodS mUSt BE UseD. SeE the proof oF LEmma \[lEMma.analYTiC-CONtiNUation\] in sectiOn \[sec:number-KIllEd\] beloW fOr a PRoof in The caSe NEedEd for the theOremS on the disTributIOn of the NUmber of Killed ParTicLes.
---
aBStRaCt: 'THe EFfeCT oF vaRIouS velocitY bOuNdary CondITION is sTudIed iN two-dImensional RayLeiGh-BéNArd ConveCtion. combInAtionS of no-sLip, stReSs-free and periodIc boUndary conDitIoNs aRe Used oN Both thE siDewAlls and The horiZOntAl PLATeS. For the studied RaylEiGH NuMbers ${\texT{Ra}}$ betWEeN $10^8$ aND $10^{11}$ the heat TrAnsPort IS Lower For $\GAMmA = 0.33$ than for $\gamma = 1$ iN CaSe Of no-sliP sIdewalLs. thiS is SurprISingLy oppoSite for sTress-FRee sidewalls, whERe the heat tranSPoRT InCReasEs fOr lower aspeCt-raTIo. In WideR CeLls THe aspEct-raTiO DePEndence is observed to DiSappeaR for ${\tExt{Ra}}\ge 10^{10}$. Two disTinct flow tYPES with verY difFErENt dynamics can bE seen, Mostly depeNDent on thE platE velocitY boundary CONdition, nAmeLy rOll-LikE FLoW and horizontaL ZOnal FlOw, which HavE a substAntIal EffEct On The dynamiCs and heaT tRaNsPoRt iN the sYStem. The pReDomInAntLy horIZontal Zonal Flow SuPpREssEs heat fLUx AND is oBsErVed fOr sTrEss-frEe anD AsyMmetric Plates. Low AspECt-raTiO pEriodic Sidewall simulAtIons with a nO-sLip BoundaRY ConditioN on the plates also exhibit ZOnal floW. In All thE othEr cases, thE flOw is roLl-lIKe. In twO-dimenSionaL RAylEIGh-BénARD cOnvEcTion, the velOCIty BoundArY conDitions Thus have large impliCAtiOns on both roll-LikE and ZONaL flOW tHAt hAvE To bE TAken into consideRation befoRe THe Boundary coNDitIoNs are imPosed.'
auThor:
- 'ERWin P. van Der Poel$^{1}$, RoDol | of branching random field , *Proc. N atl.Aca d.Sc i.*, Vol . 76, No. 2, p p 872 –875, 1979.
[^1]: Uni versi ty of C h ic ago
[^2]: U n iv e r sit yof Ch ic a go
[^3 ]:At anyrate the a rgu me nt is routin e i n the case wh ere $f (1)$and $g (1 )$ ta k e val ues in t he uni t inter val $[0,1 ]$ ; in th i s casee x is tenc e follows by a ro u ti n e phase-portra it ana ly s is f orthe associate dfirst - order s y st e m , us i ng the nonneg ativity oft heforcin gter m $\Psi (h)$ .W hen $f (1)$ an d $g (1)$ are compl e x-value d , howev er, ot her me thod s m us t b eu sed . S eet heproof of L em ma \[ lemm a . a n alyt ic- cont inuat ion\] in sect ion \[s e c:n umber -kill ed\] b elowfor aproof i n the case need ed f or the th eor em s o nthe d i stribu tio n o f the n umber o f ki ll e d pa rticles.
---
abst ra c t :'The eff ect of va ri o us veloc it y b ound a r y con diti o nis studi ed int wo -d imensio na l Rayl ei gh- Bén ard c o nvec tion.Combinat ionso f no-slip, str e ss-free and p e ri o d ic boun dar y condition s ar e use d on bo tht he si dewal ls an d the horizontal pla te s. For thestudied Rayle igh number s $ {\text{R a}}$ be t ween $10^8$ an d $10 ^{11}$ the heat tra nspor t is low er for $\ G a mma = 0. 33$ th anfor $ \G amma = 1$ inc a se o fno-slip si dewalls . T his is su rp risinglyopposite f or s tr ess -free sidewall s, wh er e t he he a t tran sport inc re as e s f or lowe r a s p ect- ra ti o. I n w id er ce llst heaspect- ratio dep end e nceis o bserved to disappear f or ${\text {R a}} \ge 10 ^ { 10}$. Tw o distinct flow types w i th very di ffere nt d ynamics c anbe see n,m ostlydepend ent o nthe p latev e lo cit yboundary c o n dit ion,na mely roll-l ike flow and horiz o nta l zonal flow, wh ichh a ve as ub s tan ti a l e f f ect on the dyna mics and h ea t t ransport i n th esystem. The pr edomi n antly h orizontal zonal fl ow sup p r ess es heat fl ux and i s observe d fors tr ess-f ree and a sy mme tricplates . Lo w asp ect-ra ti o peri odicsi dewall s imulations with a no-sl ip bou ndary co ndition o n t h e p lates als o ex hibit zona l f low . Inall the o ther ca ses , theflow is roll-l i ke . I n tw o-dimension a l Ray leigh -Bé n ard co nvec tion, the velocit y boundary cond itio n s th ush avela rge implicatio nson b oth roll -l ike and zon al flowth a t hav e to b e take n intoc o ns i derati on b efo re the bo und ar y condit io ns are im pose d. '
auth or:
-' Erwi n P. van der Poel$ ^{1}$ , Rodol | of_branching random_field, *Proc. Natl. Acad._Sci.*, Vol._76,_No. 2,_pp_872–875, 1979.
[^1]: University_of Chicago
[^2]: University_of Chicago
[^3]: At any_rate the argument_is_routine in the case where $f (1)$ and $g (1)$ take values in the_unit_interval $[0,1]$;_in_this_case existence follows by a_routine phase-portrait analysis for the_associated first-order_system, using the nonnegativity of the forcing term_$\Psi_(h)$. When $f_(1)$ and $g (1)$ are complex-valued, however, other methods_must be used. See the proof_of Lemma \[lemma.analytic-continuation\] in_section \[sec:number-killed\]_below_for a proof in_the case needed for the theorems_on the distribution of the number_of killed particles.
---
abstract: 'The effect of_various velocity boundary condition is studied_in two-dimensional Rayleigh-Bénard convection. Combinations_of no-slip,_stress-free and periodic boundary conditions_are used on_both the_sidewalls and the_horizontal plates. For the studied Rayleigh_numbers ${\text{Ra}}$ between_$10^8$ and $10^{11}$ the heat transport_is_lower for $\Gamma_=_0.33$_than for_$\Gamma = 1$_in_case of_no-slip_sidewalls. This is surprisingly opposite for_stress-free_sidewalls, where the heat transport increases for_lower aspect-ratio. In wider_cells_the aspect-ratio dependence is_observed to disappear for ${\text{Ra}}\ge_10^{10}$. Two distinct flow types with_very different_dynamics can_be seen, mostly dependent on the plate velocity boundary condition, namely_roll-like flow and horizontal zonal flow,_which have a substantial_effect on_the_dynamics and heat_transport_in the_system. The predominantly horizontal zonal flow suppresses_heat flux_and is observed for stress-free and_asymmetric plates. Low aspect-ratio_periodic_sidewall simulations with a no-slip boundary_condition on the plates also exhibit_zonal flow. In all the_other_cases,_the flow is roll-like. In_two-dimensional Rayleigh-Bénard convection, the velocity boundary_conditions thus have_large implications on both roll-like and zonal_flow_that have to be taken into_consideration_before the boundary conditions are imposed.'
author:
-_'Erwin_P._van der Poel$^{1}$, Rodol |
$a\leq b$; conversely, if $a\leq b$, then both $a/b$ and $b\backslash a$ exist and $b=a\oplus(a/b)=(b\backslash a)\oplus a$. Also, if $b\ominus a$ exists, then $a,b\ominus a\leq b$, $a\perp(b\ominus a)$ and $a\oplus(b\ominus a)=(b\ominus a)\oplus a
=b$. We note that a commutative GPEA is the same thing as a *generalized effect algebra* [@Zdenka99].
The GPEA $E$ is partially ordered by $\leq$ and $0$ is the smallest element in $E$. The cancellation laws in (GPA3) are easily extended to $\leq$ as follows: $$\mbox{If }a\oplus b\leq a\oplus c,\mbox{or if }b\oplus a\leq c\oplus a,
\mbox{ then }b\leq c.$$ An existing supremum (resp. infimum) in the partially ordered set (poset) $E$ of elements $a$ and $b$ is denoted by $a\vee b$ (resp. by $a\wedge b$). We say that $a$ and $b$ are *disjoint* iff $a\wedge b=0$. We note that a GPEA-morphism preserves inequalities and corresponding left and right differences.
An important example of a GPEA ([@DvVepo], Example 2.3) is a subset of the positive cone in a partially ordered group (po-group). Let $(G,+,0,
\leq)$ be a po-group with $G^+:=\{g\in G: 0\leq g\}$. Let $G^0$ be a nonempty subset of $G^+$ such that for all $a,b\in G^0$, if $b\leq a$ then $-a+b,\ b-a\in G^0$. Then $(G^0,\oplus,0)$, where $\oplus$ is the group addition restricted to those pairs of elements whose sum is again in $G^0$, is a GPEA whose partial order coincides with the group partial order restricted to $G^0$.
\[le:SlashProps\] Let $a,b, | $ a\leq b$; conversely, if $ a\leq b$, then both $ a / b$ and $ b\backslash a$ exist and $ b = a\oplus(a / b)=(b\backslash a)\oplus a$. Also, if $ b\ominus a$ exists, then $ deoxyadenosine monophosphate, b\ominus a\leq b$, $ a\perp(b\ominus a)$ and $ a\oplus(b\ominus a)=(b\ominus a)\oplus a
= b$. We notice that a commutative GPEA is the same thing as a * generalized consequence algebra * [ @Zdenka99 ].
The GPEA $ E$ is partially order by $ \leq$ and $ 0 $ is the smallest element in $ E$. The cancellation law in (GPA3) are well extended to $ \leq$ as follows: $ $ \mbox{If } a\oplus b\leq a\oplus c,\mbox{or if } b\oplus a\leq c\oplus a,
\mbox { then } b\leq c.$$ An existing supremum (resp. infimum) in the partially ordered set (poset) $ E$ of element $ a$ and $ b$ is denoted by $ a\vee b$ (resp. by $ a\wedge b$). We say that $ a$ and $ b$ are * disjoint * iff $ a\wedge b=0$. We notice that a GPEA - morphism preserves inequalities and corresponding left and right differences.
An important model of a GPEA ([ @DvVepo ], Example 2.3) is a subset of the positive cone in a partially ordered group (po - group). lease $ (G,+,0,
\leq)$ be a po - group with $ G^+:=\{g\in G: 0\leq g\}$. Let $ G^0 $ be a nonempty subset of $ G^+$ such that for all $ a, b\in G^0 $, if $ b\leq a$ then $ -a+b,\ b - a\in G^0$. Then $ (G^0,\oplus,0)$, where $ \oplus$ is the group addition restrict to those pairs of elements whose sum is again in $ G^0 $, is a GPEA whose partial order coincide with the group partial order restricted to $ G^0$.
\[le: SlashProps\ ] Let $ a, b, | $a\lfq b$; conversely, if $a\leq n$, then both $a/b$ aue $b\bacnslash a$ exist and $b=a\oplus(a/b)=(b\backslash a)\oppuw a$. Aoso, if $b\ominus a$ existr, then $a,b\lminus a\oeq u$, $a\perp(b\ominus a)$ and $a\oplus(b\omjkus a)=(y\oninus a)\oplus a
=n$. We note tvat a commutathvd YPEA is the same thing as a *generalised effrch algebra* [@Zdenha99].
Tht G[EA $S$ is partially ordered by $\leq$ and $0$ is the smallest elekent in $E$. The cancellation lawd in (GPA3) are easilj extended jk $\lqw$ as follows: $$\mbox{If }a\oklbs b\leq a\opmus c,\mbox{or if }b\oplus a\leq c\opljs a,
\kbox{ then }v\lwq f.$$ An existinj suprvmum (resp. inflkum) in the pattially orderec sxt (piset) $E$ of elements $a$ end $b$ is denoted by $w\vee b$ (revp. by $a\wedge b$). We sqy that $a$ atd $b$ qre *diajpiht* iff $a\wxdge b=0$. We nkte that a TPEA-morphism preserfef inequalities and cjrwesponding left and right differences.
An imlortant example of a GPWA ([@DvVepo], Example 2.3) is a subset of the positive cone in a partially ordered grou[ (po-gcojp). Oeb $(G,+,0,
\ldw)$ he a po-group with $G^+:=\{g\in G: 0\leq g\}$. Let $G^0$ be a nonqjpuy xubset of $G^+$ sugh that for all $a,b\on G^0$, yf $b\leq a$ thev $-a+b,\ b-c\jn G^0$. Then $(G^0,\oplus,0)$, whege $\opluf$ is rhe group addotion restricted to those pqirs of elemvnts whose sum is agaiu in $G^0$, is a YPEA wnose lartial order coincides witg the group partial kfder restricted go $N^0$.
\[le:VlashProkr\] Let $a,b, | $a\leq b$; conversely, if $a\leq b$, then and a$ exist $b=a\oplus(a/b)=(b\backslash a)\oplus a$. then a\leq b$, $a\perp(b\ominus and $a\oplus(b\ominus a)=(b\ominus a =b$. We note that a GPEA is the same thing as a *generalized effect algebra* [@Zdenka99]. The GPEA is partially ordered by $\leq$ and $0$ is the smallest element in $E$. cancellation in are extended to $\leq$ as follows: $$\mbox{If }a\oplus b\leq a\oplus c,\mbox{or if }b\oplus a\leq c\oplus a, \mbox{ }b\leq c.$$ An existing supremum (resp. infimum) in partially ordered set (poset) of elements $a$ and $b$ denoted $a\vee b$ by b$). say that $a$ $b$ are *disjoint* iff $a\wedge b=0$. We note that a GPEA-morphism preserves inequalities and corresponding left and differences. An of a ([@DvVepo], 2.3) a subset of cone in a partially ordered group \leq)$ be a po-group with $G^+:=\{g\in G: 0\leq Let $G^0$ a nonempty subset of $G^+$ such for all $a,b\in G^0$, if $b\leq a$ then b-a\in G^0$. Then $(G^0,\oplus,0)$, where $\oplus$ is the group addition restricted to those pairs of sum is again in is a GPEA partial coincides the partial order to $G^0$. \[le:SlashProps\] Let $a,b, | $a\leq b$; conversely, if $a\leq b$, theN both $a/b$ and $B\backSlaSh a$ ExIst aNd $b=a\Oplus(a/b)=(b\backslASh a)\oPlus a$. Also, if $b\ominus a$ exiSts, thEn $A,B\omiNUs A\leq b$, $A\perp(b\oMInUS A)$ anD $a\OpLus(B\oMInUs a)=(b\oMinUs a)\opluS a
=b$. We note tHat A cOmmutative GPea iS the same thIng As a *generalizEd eFfect aLgEbrA* [@zdenkA99].
ThE GPEA $e$ is parTIally oRdered by $\lEq$ ANd $0$ is thE SmallesT ELeMent In $E$. The cancellatioN LaWS in (GPA3) are easilY extenDeD To $\LEQ$ as FolLows: $$\mbox{If }A\oPlus b\LEq a\opluS C,\mBOX{Or iF }B\oplus a\leq c\opLus a,
\mbox{ theN }B\leQ c.$$ An exIsTinG SupremUm (resP. iNFimUm) in the partIallY ordered sEt (poseT) $e$ of elemENts $a$ and $B$ is denOteD by $A\vee B$ (ReSp. By $a\WeDGe b$). wE sAy tHAt $a$ And $b$ are *dIsJoInt* ifF $a\weDGE B=0$. we noTe tHat a gPEA-mOrphism preserVes IneqUAliTies aNd corRespOnDing lEft and Right DiFferences.
An impoRtanT example oF a GpEa ([@DvvePo], ExaMPle 2.3) is a SubSet Of the poSitive cONe iN a PARTiAlly ordered group (po-GrOUP). LEt $(G,+,0,
\leq)$ be A po-groUP wItH $g^+:=\{g\in G: 0\leq G\}$. LEt $G^0$ Be a nONEmpty SubsET oF $G^+$ such thAt for aLL $a,B\iN G^0$, if $b\leQ a$ Then $-a+b,\ B-a\In G^0$. theN $(G^0,\oplUS,0)$, wheRe $\opluS$ is the grOup adDItion restricteD To those pairs oF ElEMEnTS whoSe sUm is again in $g^0$, is a gpEA wHose PArTiaL Order CoincIdES wITh the group partial orDeR restrIcted To $G^0$.
\[le:SlashProPs\] Let $a,b, | $a\leq b$; conversely, if $a\leq b$ , the n b oth $ a/b$ and $b\backslasha $ ex ist and $b=a\oplus(a/b )=(b\ ba c ksla s ha)\op lus a$. Al s o , i f$b \om in u sa$ ex ist s, then $a,b\omin usa\ leq b$, $a\p e rp (b\ominusa)$ and $a\oplu s(b \ominu sa)= ( b\omi nus a)\o plus a =b$. W e note th at a comm u tativeG P EA isthe same thing as a* generalized ef fect a lg e br a * [@ Zde nka99].
T he GPEA $E$ isp ar t i a lly ordered by $\ leq$ and $0 $ is the s ma lle s t elem ent i n$ E$. The cancel lati on laws i n (GPA 3 ) are e a sily ex tended to $\ leq$ as f oll ow s : $ $ \m box { If}a\oplus b \l eq a\ oplu s c , \mbo x{o r if }b\o plus a\leq c\ opl us a ,
\ mbox{ then }b\ le q c.$ $ An e xisti ng supremum (resp . in fimum) in th epar ti allyo rdered se t ( poset)$E$ ofe lem en t s $a $ and $b$ is denot ed b y$a\vee b $ (res p .by $a\wedge b $). Wes a y tha t $a $ a nd $b$ a re *di s jo in t* iff$a \wedge b =0$ . W e not e tha t a GP EA-morph ism p r eserves inequa l ities and cor r es p o nd i ng l eft and rightdiff e renc es.An im p ortan t exa mp l eo f a GPEA ([@DvVepo] ,Exampl e 2.3 ) is a subset of the po s i t ive cone ina p a rtially ordere d gro up (po-gro u p). Let$(G,+ ,0,
\leq )$ be a p o - group wi th$G^ +:= \{g \ i nG: 0\leq g\}$ . Let$G ^0$ bea n onempty su bse t o f $ G^ +$ such t hat foral l$a ,b \in G^0$ , if $b\l eq a$ t hen $-a+ b ,\ b-a \in G ^0$. T he n $( G^0,\op l us , 0 )$,wh er e $\ opl us $ istheg rou p addit ion restr ict e d to t ho se pair s of elements w hose sum i saga in in$ G ^0$, isa GPEA whose partial or d er coin cid es wi th t he grouppar tial o rde r restr ictedto $G ^0 $.\ [le:S l a sh Pro ps \] Let $a, b , | $a\leq_b$; conversely,_if $a\leq b$, then_both $a/b$_and_$b\backslash a$_exist_and $b=a\oplus(a/b)=(b\backslash a)\oplus_a$. Also, if_$b\ominus a$ exists, then_$a,b\ominus a\leq b$,_$a\perp(b\ominus_a)$ and $a\oplus(b\ominus a)=(b\ominus a)\oplus a
=b$. We note that a commutative GPEA is the_same_thing as_a_*generalized_effect algebra* [@Zdenka99].
The GPEA $E$_is partially ordered by $\leq$_and $0$_is the smallest element in $E$. The cancellation_laws_in (GPA3) are_easily extended to $\leq$ as follows: $$\mbox{If }a\oplus b\leq_a\oplus c,\mbox{or if }b\oplus a\leq c\oplus_a,
\mbox{ then_}b\leq_c.$$_An existing supremum (resp._infimum) in the partially ordered set_(poset) $E$ of elements $a$ and_$b$ is denoted by $a\vee b$ (resp._by $a\wedge b$). We say that_$a$ and $b$ are *disjoint*_iff $a\wedge_b=0$. We note that a_GPEA-morphism preserves inequalities_and corresponding_left and right_differences.
An important example of a GPEA_([@DvVepo], Example 2.3)_is a subset of the positive_cone_in a partially_ordered_group_(po-group). Let_$(G,+,0,
\leq)$ be a_po-group_with $G^+:=\{g\in_G:_0\leq g\}$. Let $G^0$ be a_nonempty_subset of $G^+$ such that for all_$a,b\in G^0$, if $b\leq_a$_then $-a+b,\ b-a\in G^0$._Then $(G^0,\oplus,0)$, where $\oplus$ is_the group addition restricted to those_pairs of_elements whose_sum is again in $G^0$, is a GPEA whose partial order_coincides with the group partial order_restricted to $G^0$.
\[le:SlashProps\] Let_$a,b, |
$-$0.67$\pm$0.42 & $+$0.25$\pm$0.42 & 0.72$\pm$0.60 & & & 15.0 & $\le$8 & 1.12$\pm$0.23\
& & $J$ & 1.5 & 2.3 & $+$0.09$\pm$0.42 & $-$0.41$\pm$0.42 & 1.21$\pm$0.60 & 0.42$\pm$0.60 & & & &\
USco128 & M7 & $J$ & 1.0 & 6.0 & $-$0.30$\pm$0.42 & $-$0.11$\pm$0.42 & 0.32$\pm$0.60 & & & $\le$5 & $\le$24 & 0.78$\pm$0.22\
& & $H$ & 1.1 & 6.0 & $-$0.17$\pm$0.42 & $+$0.09$\pm$0.42 & 0.19$\pm$0.60 & & & & &\
USco132 & M7 & $J$ & 1.0 & 5.0 & $-$0.21$\pm$0.42 & $-$0.22$\pm$0.42 & 0.31$\pm$0.60 & & & & & 0.33$\pm$0.22\
J0136$+$09 & T2.5 & $J$ & 3.0 & 15.5 & $+$0.04$\pm$0.21 & $-$0.32$\pm$0.21 & 1.21$\pm$0.60 & 0.33$\pm$0.30 & & $\le$50 & 2.4 & 0.32$\pm$0.22\
J0241$-$03 & L0 & $J$ & 2.2 & 11.0 & $-$2.03$\pm$0.21 & $-$2.26$\pm$0.21 & 3.67$\pm$0.60 & 3.04$\pm$0.30 & 110$\pm$10 & & & 0.73$\pm$0.23\
J0355$+$11 & L5 & $J$ & 3.2 & 10.5 & $+$0.60$\pm$0.42 & | $ -$0.67$\pm$0.42 & $ + $ 0.25$\pm$0.42 & 0.72$\pm$0.60 & & & 15.0 & $ \le$8 & 1.12$\pm$0.23\
& & $ J$ & 1.5 & 2.3 & $ + $ 0.09$\pm$0.42 & $ -$0.41$\pm$0.42 & 1.21$\pm$0.60 & 0.42$\pm$0.60 & & & & \
USco128 & M7 & $ J$ & 1.0 & 6.0 & $ -$0.30$\pm$0.42 & $ -$0.11$\pm$0.42 & 0.32$\pm$0.60 & & & $ \le$5 & $ \le$24 & 0.78$\pm$0.22\
& & $ H$ & 1.1 & 6.0 & $ -$0.17$\pm$0.42 & $ + $ 0.09$\pm$0.42 & 0.19$\pm$0.60 & & & & & \
USco132 & M7 & $ J$ & 1.0 & 5.0 & $ -$0.21$\pm$0.42 & $ -$0.22$\pm$0.42 & 0.31$\pm$0.60 & & & & & 0.33$\pm$0.22\
J0136$+$09 & T2.5 & $ J$ & 3.0 & 15.5 & $ + $ 0.04$\pm$0.21 & $ -$0.32$\pm$0.21 & 1.21$\pm$0.60 & 0.33$\pm$0.30 & & $ \le$50 & 2.4 & 0.32$\pm$0.22\
J0241$-$03 & L0 & $ J$ & 2.2 & 11.0 & $ -$2.03$\pm$0.21 & $ -$2.26$\pm$0.21 & 3.67$\pm$0.60 & 3.04$\pm$0.30 & 110$\pm$10 & & & 0.73$\pm$0.23\
J0355$+$11 & L5 & $ J$ & 3.2 & 10.5 & $ + $ 0.60$\pm$0.42 & | $-$0.67$\pm$0.42 & $+$0.25$\pm$0.42 & 0.72$\pm$0.60 & & & 15.0 & $\le$8 & 1.12$\pm$0.23\
& & $J$ & 1.5 & 2.3 & $+$0.09$\pm$0.42 & $-$0.41$\pm$0.42 & 1.21$\pm$0.60 & 0.42$\pm$0.60 & & & &\
USco128 & M7 & $J$ & 1.0 & 6.0 & $-$0.30$\pm$0.42 & $-$0.11$\pm$0.42 & 0.32$\pm$0.60 & & & $\le$5 & $\le$24 & 0.78$\pm$0.22\
& & $H$ & 1.1 & 6.0 & $-$0.17$\pm$0.42 & $+$0.09$\pm$0.42 & 0.19$\pm$0.60 & & & & &\
USco132 & O7 & $J$ & 1.0 & 5.0 & $-$0.21$\pm$0.42 & $-$0.22$\pm$0.42 & 0.31$\pm$0.60 & & & & & 0.33$\pm$0.22\
J0136$+$09 & T2.5 & $J$ & 3.0 & 15.5 & $+$0.04$\pm$0.21 & $-$0.32$\pj$0.21 & 1.21$\pm$0.60 & 0.33$\pm$0.30 & & $\le$50 & 2.4 & 0.32$\pm$0.22\
M0241$-$03 & L0 & $J$ & 2.2 & 11.0 & $-$2.03$\pm$0.21 & $-$2.26$\pm$0.21 & 3.67$\pm$0.60 & 3.04$\[m$0.30 & 110$\'m$10 & & & 0.73$\pm$0.23\
J0355$+$11 & L5 & $J$ & 3.2 & 10.5 & $+$0.60$\pm$0.42 & | $-$0.67$\pm$0.42 & $+$0.25$\pm$0.42 & 0.72$\pm$0.60 & & & & 1.12$\pm$0.23\ & $J$ & & & 1.21$\pm$0.60 & & & & USco128 & M7 & $J$ & & 6.0 & $-$0.30$\pm$0.42 & $-$0.11$\pm$0.42 & 0.32$\pm$0.60 & & & $\le$5 & & 0.78$\pm$0.22\ & & $H$ & 1.1 & 6.0 & $-$0.17$\pm$0.42 & $+$0.09$\pm$0.42 0.19$\pm$0.60 & & USco132 & M7 & $J$ & 1.0 & 5.0 & $-$0.21$\pm$0.42 & $-$0.22$\pm$0.42 & 0.31$\pm$0.60 & & & & 0.33$\pm$0.22\ J0136$+$09 & T2.5 & & 3.0 & 15.5 $+$0.04$\pm$0.21 & $-$0.32$\pm$0.21 & 1.21$\pm$0.60 0.33$\pm$0.30 & $\le$50 2.4 0.32$\pm$0.22\ & L0 & & 2.2 & 11.0 & $-$2.03$\pm$0.21 & $-$2.26$\pm$0.21 & 3.67$\pm$0.60 & 3.04$\pm$0.30 & 110$\pm$10 & & & J0355$+$11 & $J$ & & & & | $-$0.67$\pm$0.42 & $+$0.25$\pm$0.42 & 0.72$\pm$0.60 & & & 15.0 & $\le$8 & 1.12$\pm$0.23\
& & $J$ & 1.5 & 2.3 & $+$0.09$\pm$0.42 & $-$0.41$\pm$0.42 & 1.21$\pm$0.60 & 0.42$\pm$0.60 & & & &\
USco128 & M7 & $J$ & 1.0 & 6.0 & $-$0.30$\Pm$0.42 & $-$0.11$\pm$0.42 & 0.32$\pm$0.60 & & & $\le$5 & $\le$24 & 0.78$\Pm$0.22\
& & $H$ & 1.1 & 6.0 & $-$0.17$\pm$0.42 & $+$0.09$\Pm$0.42 & 0.19$\pM$0.60 & & & & &\
USCo132 & m7 & $J$ & 1.0 & 5.0 & $-$0.21$\pm$0.42 & $-$0.22$\Pm$0.42 & 0.31$\pm$0.60 & & & & & 0.33$\Pm$0.22\
J0136$+$09 & T2.5 & $J$ & 3.0 & 15.5 & $+$0.04$\pm$0.21 & $-$0.32$\pm$0.21 & 1.21$\pm$0.60 & 0.33$\pm$0.30 & & $\lE$50 & 2.4 & 0.32$\Pm$0.22\
J0241$-$03 & L0 & $j$ & 2.2 & 11.0 & $-$2.03$\pm$0.21 & $-$2.26$\pm$0.21 & 3.67$\pm$0.60 & 3.04$\pm$0.30 & 110$\pm$10 & & & 0.73$\pm$0.23\
J0355$+$11 & L5 & $J$ & 3.2 & 10.5 & $+$0.60$\pm$0.42 & | $-$0.67$\pm$0.42 & $+$0.2 5$\pm$0.42 & 0. 72$ \pm $0 .60& && 15.0 & $\le$ 8 & 1 .12$\pm$0.23\
& & $J$& 1.5 & 2.3& $ +$0.0 9$\pm$0 . 42 & $- $0 .4 1$\ pm $ 0. 42 &1.2 1$\pm$0 .60 & 0.42 $\p m$ 0.60 & & & & \
U Sco128 & M 7 & $J$ & 1.0 & 6. 0 & $- $0 .30 $ \pm$0 .42 & $- $0.11$ \ pm$0.4 2 & 0.32$ \p m $0.60& & & $\ l e $5 & $ \le$24 & 0.78$\pm $ 0. 2 2\
& & $H$ & 1 .1 & 6 .0 &$ - $0. 17$ \pm$0.42 & $ +$0.0 9 $\pm$0. 4 2& 0 .19 $ \pm$0.60 & && & &\
USco 1 32& M7 & $ J$& 1.0 & 5.0&$ -$0 .21$\pm$0.4 2 &$-$0.22$\ pm$0.4 2 & 0.31 $ \pm$0.6 0 & && & &0.33 $ \p m$ 0.2 2\ J01 3 6$ +$0 9 &T2.5 & $ J$ & 3.0& 15 . 5 & $+$ 0.0 4$\p m$0.2 1 & $-$0.32$\ pm$ 0.21 & 1 .21$\ pm$0. 60 & 0 .33$\ pm$0.3 0 & & $ \le$50 & 2.4 &0.32 $\pm$0.22 \
J 02 41$ -$ 03 &L 0 & $J $ & 2. 2 & 11. 0 & $-$ 2 .03 $\ p m $ 0. 21 & $-$2.26$\pm$0 .2 1 &3.67$\pm $0.60& 3 .0 4 $\pm$0.3 0& 1 10$\ p m $10 & & & 0. 73$\pm$0 .23\
J 0 35 5$ +$11 &L5 & $J$ & 3. 2 & 10.5 & $+ $0.60$ \pm$0.42 & | $-$0.67$\pm$0.42_& $+$0.25$\pm$0.42_& 0.72$\pm$0.60 & &_& 15.0_&_$\le$8 &_1.12$\pm$0.23\
&_& $J$ &_1.5 & 2.3_& $+$0.09$\pm$0.42 & $-$0.41$\pm$0.42_& 1.21$\pm$0.60 &_0.42$\pm$0.60_& & & &\
USco128 & M7 & $J$ & 1.0 & 6.0 & $-$0.30$\pm$0.42_&_$-$0.11$\pm$0.42 &_0.32$\pm$0.60_&_& & $\le$5 & $\le$24_& 0.78$\pm$0.22\
& & $H$ &_1.1 &_6.0 & $-$0.17$\pm$0.42 & $+$0.09$\pm$0.42 & 0.19$\pm$0.60 &_&_& & &\
USco132_& M7 & $J$ & 1.0 & 5.0 &_$-$0.21$\pm$0.42 & $-$0.22$\pm$0.42 & 0.31$\pm$0.60 &_& & &_&_0.33$\pm$0.22\
J0136$+$09_& T2.5 & $J$_& 3.0 & 15.5 & $+$0.04$\pm$0.21_& $-$0.32$\pm$0.21 & 1.21$\pm$0.60 & 0.33$\pm$0.30_& & $\le$50 & 2.4 & 0.32$\pm$0.22\
J0241$-$03_& L0 & $J$ & 2.2_& 11.0 & $-$2.03$\pm$0.21 &_$-$2.26$\pm$0.21 &_3.67$\pm$0.60 & 3.04$\pm$0.30 & 110$\pm$10_& & &_0.73$\pm$0.23\
J0355$+$11 &_L5 & $J$_& 3.2 & 10.5 & $+$0.60$\pm$0.42_& |
The payoff functions of the players are $$\begin{aligned}
\pi_A=&(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$ $$\begin{aligned}
\pi_B=&(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$ $$\begin{aligned}
\pi_E=&(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B],\end{aligned}$$ $$\pi_C=(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx_C-[(a-x_C-x_D-bx_A-bx_B-bx_E)x_D-c_Cx_D],$$ $$\pi_D=(a-x_C-x_D-bx_A-bx_B-bx_E)x_D-c_Cx_D-[(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx | The payoff functions of the players are $ $ \begin{aligned }
\pi_A=&(a - x_A - x_B - x_E - bx_C - bx_D)x_A - c_Ax_A-\frac{1}{2}[(a - x_A - x_B - x_E - bx_C - bx_D)x_B - c_Ax_B\\
& + (a - x_A - x_B - x_E - bx_C - bx_D)x_E - c_Ax_E],\end{aligned}$$ $ $ \begin{aligned }
\pi_B=&(a - x_A - x_B - x_E - bx_C - bx_D)x_B - c_Ax_B-\frac{1}{2}[(a - x_A - x_B - x_E - bx_C - bx_D)x_A - c_Ax_A\\
& + (a - x_A - x_B - x_E - bx_C - bx_D)x_E - c_Ax_E],\end{aligned}$$ $ $ \begin{aligned }
\pi_E=&(a - x_A - x_B - x_E - bx_C - bx_D)x_E - c_Ax_E-\frac{1}{2}[(a - x_A - x_B - x_E - bx_C - bx_D)x_A - c_Ax_A\\
& + (a - x_A - x_B - x_E - bx_C - bx_D)x_B - c_Ax_B],\end{aligned}$$ $ $ \pi_C=(a - x_C - x_D - bx_A - bx_B - bx_E)x_C - c_Cx_C-[(a - x_C - x_D - bx_A - bx_B - bx_E)x_D - c_Cx_D],$$ $ $ \pi_D=(a - x_C - x_D - bx_A - bx_B - bx_E)x_D - c_Cx_D-[(a - x_C - x_D - bx_A - bx_B - bx_E)x_C - c_Cx | Thf payoff functions of tht players are $$\begnb{alignxd}
\pi_A=&(a-x_Z-x_B-x_E-bx_C-cx_D)x_A-c_Ax_A-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_B-r_Ax_B\\
&+(q-x_A-x_B-z_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligndd}$$ $$\begin{apigned}
\pi_V=&(a-x_A-e_B-x_E-bx_C-bx_D)x_B-c_Ax_B-\hdac{1}{2}[(a-x_A-x_N-r_E-bx_C-gw_D)x_A-c_Cx_E\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_G-c_Ax_E],\end{aligted}$$ $$\begin{alignad}
\oi_Z=&(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_Q)x_A-c_Ax_A\\
&+(s-x_W-x_B-x_E-bx_C-bx_D)x_B-c_Wx_B],\emq{alifned}$$ $$\pi_C=(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx_C-[(a-x_C-x_D-bs_A-bx_B-bx_T)x_D-c_Cx_D],$$ $$\pi_D=(a-x_C-x_D-bx_S-bx_B-bx_E)x_D-c_Cx_D-[(a-x_C-x_D-bx_A-bx_B-bx_F)x_C-c_Fx | The payoff functions of the players are &+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$ \pi_B=&(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\ &+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$ \pi_E=&(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\ &+(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B],\end{aligned}$$ $$\pi_C=(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx_C-[(a-x_C-x_D-bx_A-bx_B-bx_E)x_D-c_Cx_D],$$ | The payoff functions of the plAyers are $$\beGin{alIgnEd}
\pI_A=&(A-x_A-x_b-x_E-bX_C-bx_D)x_A-c_Ax_A-\fraC{1}{2}[(A-x_A-x_b-x_E-bx_C-bx_D)x_B-c_Ax_B\\
&+(a-x_A-x_B-x_E-Bx_C-bx_d)x_e-C_Ax_E],\ENd{AlignEd}$$ $$\begin{ALiGNEd}
\pI_B=&(A-x_a-x_B-X_E-BX_C-Bx_D)x_B-C_Ax_b-\frac{1}{2}[(a-x_a-x_B-x_E-bx_C-bx_d)x_A-C_AX_A\\
&+(a-x_A-x_B-x_E-bx_C-BX_D)X_E-c_Ax_E],\end{aLigNed}$$ $$\begin{aligNed}
\Pi_E=&(a-x_A-X_B-X_E-bX_c-bx_D)x_e-c_AX_E-\fraC{1}{2}[(a-x_A-x_B-X_e-bx_C-bx_d)x_A-c_Ax_A\\
&+(a-x_a-x_b-X_E-bx_C-bX_d)x_B-c_Ax_B],\END{aLignEd}$$ $$\pi_C=(a-x_C-x_D-bx_A-bx_B-bX_e)x_c-C_Cx_C-[(a-x_C-x_D-bx_A-bx_b-bx_E)x_D-C_CX_d],$$ $$\pI_d=(A-x_C-X_D-bX_A-bx_B-bx_E)x_D-C_CX_D-[(a-x_C-X_d-bx_A-bx_B-BX_E)X_c-C_cx | The payoff functions of t he players are$$\ beg in {ali gned }
\pi_A=&(a-x_ A -x_B -x_E-bx_C-bx_D)x_A-c_A x_A-\ fr a c{1} { 2} [(a-x _A-x_B- x _E - b x_C -b x_ D)x _B - c_ Ax_B\ \
& +(a-x_A -x_B-x_E-b x_C -b x_D)x_E-c_Ax _ E] ,\end{alig ned }$$ $$\begin {al igned}
\ pi_ B =&(a- x_A -x_B- x_E-bx _ C-bx_D )x_B-c_Ax _B - \frac{ 1 }{2}[(a - x _A -x_B -x_E-bx_C-bx_D)x_ A -c _ Ax_A\\
&+(a-x_ A-x_B- x_ E -b x _ C-b x_D )x_E-c_Ax_ E] ,\end { aligned } $$ $ $ \be g in{aligned}
\ pi_E=&(a-x_ A -x_ B-x_E- bx _C- b x_D)x_ E-c_A x_ E -\f rac{1}{2}[( a-x_ A-x_B-x_E -bx_C- b x_D)x_A - c_Ax_A\ \
&+(a -x_ A-x _B-x _ E- bx _C- bx _ D)x _ B- c_A x _B] ,\end{al ig ne d}$$$$\p i _ C = (a-x _C- x_D- bx_A- bx_B-bx_E)x_C -c_ Cx_C - [(a -x_C- x_D-b x_A- bx _B-bx _E)x_D -c_Cx _D ],$$ $$\pi_D=(a -x_C -x_D-bx_A -bx _B -bx _E )x_D- c _Cx_D- [(a -x_ C-x_D-b x_A-bx_ B -bx _E ) x _ C- c_Cx | The_payoff functions_of the players are_$$\begin{aligned}
\pi_A=&(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$ $$\begin{aligned}
\pi_B=&(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E],\end{aligned}$$_$$\begin{aligned}
\pi_E=&(a-x_A-x_B-x_E-bx_C-bx_D)x_E-c_Ax_E-\frac{1}{2}[(a-x_A-x_B-x_E-bx_C-bx_D)x_A-c_Ax_A\\
&+(a-x_A-x_B-x_E-bx_C-bx_D)x_B-c_Ax_B],\end{aligned}$$_$$\pi_C=(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx_C-[(a-x_C-x_D-bx_A-bx_B-bx_E)x_D-c_Cx_D],$$ $$\pi_D=(a-x_C-x_D-bx_A-bx_B-bx_E)x_D-c_Cx_D-[(a-x_C-x_D-bx_A-bx_B-bx_E)x_C-c_Cx |
) - f(\xi ^{'}) | \leq \| \xi - \xi ^{'} \| _{*}. $
### For continuity of $g$: {#for-continuity-of-g.unnumbered}
We follow the same argument as above.
Let $\Phi$ be an internal function such that condition $(1)$ holds. To show that $(1) \Rightarrow (2)$, fix $\varepsilon \gg 0$. We shall show there exists a $\delta$ for this $\varepsilon$ as in condition $(2)$. Since $\Phi$ is internal, the set $$I = \left\lbrace \delta \in { }^{*} \mathbb{R}_{>0}: \text{ }\forall \omega^{'} \in { }^{*}\Omega \text{ } ( { }^{*} \| \omega - \omega^{'} \|_{\infty} < \delta \Rightarrow { }^{*} |\Phi (\omega)- \Phi (\omega ^{'})| < \varepsilon ) \right\rbrace,$$ is internal by the Internal Definition Principle and also contains every positive infinitesimal. By Overspill (cf. @Albeverio [Proposition $1.27$]) $I$ must then contain some positive $\delta \in \mathbb{R}$.\
Conversely, suppose condition (1) does not hold, that is, there exists some $\omega^{'} \in { }^{*}\Omega$ such that $${ }^{*} \| \omega - \omega^{'} \|_{\infty} \simeq 0 \text{ and } { }^{*} |\Phi (\omega)- \Phi (\omega ^{'})| \text{ is not infinitesimal}.$$ If $\varepsilon = \min(1, { }^{*}| \Phi (\omega)- \Phi (\omega ^{'})|/2),$ we know that for each standard $\delta >0,$ there is a point $\omega ^{'}$ within $\delta$ of $\omega$ at which $\Phi (\omega ^{'})$ is farther than $\varepsilon$ from $\Phi (\omega)$. This shows that condition $(2)$ cannot hold either.
[10]{}
, S., R. [H[ø]{}egh-Krohn]{}, J. [Fenstad]{}, and T. [Lindstr[ø]{}m]{} (1986)., Volume 122 of [*[Pure and Applied Mathematics]{}*]{}. Orlando: Academic Press.
, | ) - f(\xi ^ {' }) | \leq \| \xi - \xi ^ {' } \| _ { * }. $
# # # For continuity of $ g$: { # for-continuity-of-g.unnumbered }
We follow the same argument as above.
lease $ \Phi$ be an inner function such that condition $ (1)$ holds. To testify that $ (1) \Rightarrow (2)$, fix $ \varepsilon \gg 0$. We shall show there exist a $ \delta$ for this $ \varepsilon$ as in circumstance $ (2)$. Since $ \Phi$ is internal, the set $ $ I = \left\lbrace \delta \in { } ^ { * } \mathbb{R}_{>0 }: \text { } \forall \omega^ {' } \in { } ^{*}\Omega \text { } ({ } ^ { * } \| \omega - \omega^ {' } \|_{\infty } < \delta \Rightarrow { } ^ { * } |\Phi (\omega)- \Phi (\omega ^{'})| < \varepsilon) \right\rbrace,$$ is internal by the Internal Definition Principle and also contains every positivist infinitesimal. By Overspill (cf. @Albeverio [ Proposition $ 1.27 $ ]) $ I$ must then contain some incontrovertible $ \delta \in \mathbb{R}$.\
Conversely, suppose condition (1) does not hold, that is, there exists some $ \omega^ {' } \in { } ^{*}\Omega$ such that $ $ { } ^ { * } \| \omega - \omega^ {' } \|_{\infty } \simeq 0 \text { and } { } ^ { * } |\Phi (\omega)- \Phi (\omega ^{'})| \text { is not infinitesimal}.$$ If $ \varepsilon = \min(1, { } ^{*}| \Phi (\omega)- \Phi (\omega ^{'})|/2),$ we know that for each standard $ \delta > 0,$ there is a detail $ \omega ^{'}$ within $ \delta$ of $ \omega$ at which $ \Phi (\omega ^{'})$ is farther than $ \varepsilon$ from $ \Phi (\omega)$. This shows that condition $ (2)$ cannot agree either.
[ 10 ] { }
, S., R. [ H[ø]{}egh - Krohn ] { }, J. [ Fenstad ] { }, and T. [ Lindstr[ø]{}m ] { } (1986). , Volume 122 of [ * [ Pure and Applied Mathematics ] { } * ] { }. Orlando: Academic Press.
, | ) - f(\di ^{'}) | \leq \| \xi - \xi ^{'} \| _{*}. $
### Nor continuity oy $g$: {#for-rontinujty-of-g.unvumbered}
We follow the same acgumwnt aw above.
Let $\Phi$ be an ivternal flnction sych uhat condition $(1)$ holds. To show tgwt $(1) \Xijhtarrow (2)$, fix $\vsrepsilon \cg 0$. We shall svod chere exists a $\delta$ for this $\varepsylon$ as ij condition $(2)$. Synce $\[hi$ js internal, the set $$I = \left\lbrace \selta \ii { }^{*} \mathbb{R}_{>0}: \tect{ }\forall \omega^{'} \in { }^{*}\Omega \hext{ } ( { }^{*} \| \omega - \omega^{'} \|_{\infty} < \deljz \Rythtarrow { }^{*} |\Pfi (\omega)- \Khn (\omega ^{'})| < \vzrepsilon ) \right\rbrace,$$ is interval bv the Interbao Dfxinition Prmnciplv and also coknains eeery poxitive infinitcsiman. Bt Overspill (cf. @Albevecio [Proposition $1.27$]) $I$ most then cmncain some positive $\deotq \in \kathtb{R}$.\
Cubvefsemy, shppose coidition (1) doss not hold, that is, there exisus fime $\omega^{'} \in { }^{*}\Omega$ stch that $${ }^{*} \| \omega - \omega^{'} \|_{\infty} \simeq 0 \ttxt{ ahd } { }^{*} |\Phi (\omega)- \Phi (\omwga ^{'})| \text{ is not infijitesimal}.$$ If $\varepsilon = \min(1, { }^{*}| \Phi (\omega)- \Phi (\omega ^{'})|/2),$ we knof thav wor ewzy dtandard $\delta >0,$ there is a point $\omega ^{'}$ withig $\dtlts$ of $\omega$ at chich $\Phi (\omega ^{'})$ id gwrther than $\vxrepsilon$ from $\Phi (\omega)$. Thls showf thar conditijn $(2)$ vannot hold either.
[10]{}
, S., R. [H[ø]{}egh-Keohn]{}, J. [Fenstab]{}, abd T. [Lindstr[ø]{}m]{} (1986)., Volule 122 of [*[Pure and Apploed Mathematics]{}*]{}. Orlando: Acasemic Press.
, | ) - f(\xi ^{'}) | \leq \| \xi \| _{*}. ### For continuity the argument as above. $\Phi$ be an function such that condition $(1)$ holds. show that $(1) \Rightarrow (2)$, fix $\varepsilon \gg 0$. We shall show there a $\delta$ for this $\varepsilon$ as in condition $(2)$. Since $\Phi$ is internal, set = \delta { }^{*} \mathbb{R}_{>0}: \text{ }\forall \omega^{'} \in { }^{*}\Omega \text{ } ( { }^{*} \| \omega \omega^{'} \|_{\infty} < \delta \Rightarrow { }^{*} |\Phi \Phi (\omega ^{'})| < ) \right\rbrace,$$ is internal by Internal Principle and contains positive By Overspill (cf. [Proposition $1.27$]) $I$ must then contain some positive $\delta \in \mathbb{R}$.\ Conversely, suppose condition (1) does not that is, some $\omega^{'} { such $${ }^{*} \| \omega^{'} \|_{\infty} \simeq 0 \text{ and |\Phi (\omega)- \Phi (\omega ^{'})| \text{ is not If $\varepsilon \min(1, { }^{*}| \Phi (\omega)- \Phi ^{'})|/2),$ we know that for each standard $\delta there is a point $\omega ^{'}$ within $\delta$ of $\omega$ at which $\Phi (\omega ^{'})$ than $\varepsilon$ from $\Phi This shows that $(2)$ hold [10]{} S., R. J. [Fenstad]{}, and T. [Lindstr[ø]{}m]{} (1986)., Volume 122 of [*[Pure and Mathematics]{}*]{}. Orlando: Academic Press. , | ) - f(\xi ^{'}) | \leq \| \xi - \xi ^{'} \| _{*}. $
### For continuity of $G$: {#for-continUity-oF-g.uNnuMbEred}
we foLlow the same argUMent As above.
Let $\Phi$ be an interNal fuNcTIon sUCh That cOnditioN $(1)$ HoLDS. To ShOw ThaT $(1) \RIGhTarroW (2)$, fiX $\varepsIlon \gg 0$. We shAll ShOw there existS A $\dElta$ for thiS $\vaRepsilon$ as in ConDition $(2)$. siNce $\pHi$ is iNteRnal, tHe set $$I = \LEft\lbrAce \delta \iN { }^{*} \mAThbb{R}_{>0}: \tEXt{ }\foralL \OMeGa^{'} \in { }^{*}\omega \text{ } ( { }^{*} \| \omega - \omeGA^{'} \|_{\iNFty} < \delta \RightaRrow { }^{*} |\PhI (\oMEgA)- \pHi (\oMegA ^{'})| < \varepsiloN ) \rIght\rBRace,$$ is iNTeRNAL by THe Internal DefInition PrinCIplE and alSo ConTAins evEry poSiTIve InfinitesimAl. By overspill (Cf. @AlbeVErio [ProPOsition $1.27$]) $i$ must tHen ConTain SOmE pOsiTiVE $\deLTa \In \mAThbB{R}$.\
ConverSeLy, SuppoSe coNDITIon (1) dOes Not hOld, thAt is, there exisTs sOme $\oMEga^{'} \In { }^{*}\OmeGa$ sucH thaT $${ }^{*} \| \oMega - \oMega^{'} \|_{\inFty} \siMeQ 0 \text{ and } { }^{*} |\Phi (\omegA)- \Phi (\Omega ^{'})| \text{ Is nOt InfInItesiMAl}.$$ If $\vaRepSilOn = \min(1, { }^{*}| \PhI (\omega)- \PHI (\omEgA ^{'})|/2),$ WE KnOw that for each standArD $\DElTa >0,$ there iS a poinT $\OmEgA ^{'}$ Within $\deLtA$ of $\OmegA$ AT whicH $\Phi (\OMeGa ^{'})$ is fartHer thaN $\VaRePsilon$ fRoM $\Phi (\omEgA)$. ThIs sHows tHAt coNditioN $(2)$ cannot hOld eiTHer.
[10]{}
, S., R. [H[ø]{}egh-KrohN]{}, j. [Fenstad]{}, and T. [LINdSTR[ø]{}M]{} (1986)., voluMe 122 oF [*[Pure and AppLied mAtheMatiCS]{}*]{}. ORlaNDo: AcaDemic prESs.
, | ) - f(\xi ^{'}) | \leq \ | \xi - \ xi ^{ '}\|_{ *}.$
# ## For continu i ty o f $g$: {#for-continuit y-of- g. u nnum b er ed}
We foll o wt h e s am earg um e nt as a bov e.
Let $\Phi$ be an i nternal func t io n such tha t c ondition $(1 )$holds. T o s h ow th at$(1)\Right a rrow ( 2)$, fix$\ v arepsi l on \gg0 $ .We s hall show there e x is t s a $\delta$ f or thi s$ \v a r eps ilo n$ as in c on ditio n $(2)$. Si n c e $\ P hi$ is intern al, the set $$I = \le ft \lb r ace \d elta\i n { }^{*} \mat hbb{ R}_{>0}:\text{ }\foral l \omega ^{'} \ in{ } ^{*} \ Om eg a \ te x t{} ( {} ^{* } \| \om eg a- \om ega^ { ' } \|_{ \in fty} < \d elta \Rightar row { } ^ {*} |\Ph i (\o mega )- \Ph i (\om ega ^ {' })| < \varepsil on ) \right\r bra ce ,$$ i s int e rnal b y t heInterna l Defin i tio nP r i nc iple and also cont ai n s e very pos itivei nf in i tesimal. B y O vers p i ll (c f. @ A lb everio [ Propos i ti on $1.27$ ]) $I$ m us t t hen cont a in s ome po sitive $ \delt a \in \mathbb{R } $.\
Conversel y ,s u pp o se c ond ition (1) d oesn ot h old, th ati s, th ere e xi s ts some $\omega^{'} \i n{ }^{* }\Ome ga$ such that $${ }^{*} \ | \omega- \o m eg a ^{'} \|_{\inft y} \s imeq 0 \te x t{ and } { } ^{*} |\P hi (\omeg a ) - \Phi ( \om ega ^{ '}) | \t ext{ is not i n f init es imal}.$ $ I f $\var eps ilo n = \m in (1, { }^{ *}| \Phi ( \o me ga )-\Phi( \omega ^ {' })| /2 ),$ we k n ow tha t for eac hst a nda rd $\de l ta > 0,$th er e is apo int $ \ome g a ^ {'}$ wi thin $\de lta $ of$\ om ega$ at which $\Phi(\ omega ^{'} )$ is farth e r than $\ varepsilon$ from $\Phi( \omega) $.Thisshow s that co ndi tion $ (2) $ canno t hold eith er .
[ 1 0]{}, S .,R. [H[ø]{}eg h - Kro hn]{} ,J. [ Fenstad ]{}, and T. [Linds t r[ø ]{}m]{} (1986 )., Vol u m e122 of [*[ Pu r e a n d Applied Mathem atics]{}*] {} . O rlando: Ac a dem ic Press.
, | ) -_f(\xi ^{'})_ | \leq _\| \xi__- \xi_^{'}_\| _{*}. $
###_For continuity of_$g$: {#for-continuity-of-g.unnumbered}
We follow the_same argument as_above.
Let_$\Phi$ be an internal function such that condition $(1)$ holds. To show that $(1)_\Rightarrow_(2)$, fix_$\varepsilon_\gg_0$. We shall show there_exists a $\delta$ for this_$\varepsilon$ as_in condition $(2)$. Since $\Phi$ is internal, the_set_$$I = \left\lbrace_\delta \in { }^{*} \mathbb{R}_{>0}: \text{ }\forall \omega^{'}_\in { }^{*}\Omega \text{ } (_{ }^{*} \|_\omega_-_\omega^{'} \|_{\infty} < \delta_\Rightarrow { }^{*} |\Phi (\omega)- _\Phi (\omega ^{'})| < \varepsilon )_\right\rbrace,$$ is internal by the Internal Definition_Principle and also contains every positive_infinitesimal. By Overspill (cf. @Albeverio_[Proposition $1.27$])_$I$ must then contain some_positive $\delta \in_\mathbb{R}$.\
Conversely, suppose_condition (1) does_not hold, that is, there exists_some $\omega^{'} \in_{ }^{*}\Omega$ such that $${ }^{*}_\|_\omega - \omega^{'}_\|_{\infty}_\simeq_0 \text{_and } _{_}^{*} |\Phi_(\omega)-_\Phi (\omega ^{'})| \text{ is not_infinitesimal}.$$_If $\varepsilon = \min(1, { }^{*}| \Phi_(\omega)- \Phi (\omega ^{'})|/2),$_we_know that for each_standard $\delta >0,$ there is_a point $\omega ^{'}$ within $\delta$_of $\omega$_at which_$\Phi (\omega ^{'})$ is farther than $\varepsilon$ from $\Phi (\omega)$. This_shows that condition $(2)$ cannot hold_either.
[10]{}
, S., R. [H[ø]{}egh-Krohn]{}, J. [Fenstad]{},_and T. [Lindstr[ø]{}m]{}_(1986).,_Volume 122 of_[*[Pure_and Applied_Mathematics]{}*]{}. Orlando: Academic Press.
, |
) = 0$ by [@Kaw85b Theorem 8.2].
Arguing by contradiction, assume that there exists $p\geq1$ and an infinite set $\mathcal T\subseteq \Z$ such that $$H^0\big(Y,(\Omega^1_Y)^{\otimes p} \otimes \OO_Y(m\pi^*K_X)\big) \neq 0$$ for all $m\in\mathcal T$. Denote $\mathcal E=(\Omega^1_Y)^{\otimes p} $ and $Z=\PS(\mathcal E)$ with the projection $f\colon Z\to Y$. First note that $$H^0\big(Y,\mathcal E\otimes \OO_Y(m\pi^*K_X)\big)\simeq H^0\big(Z,\OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)\big).$$ Since $K_X$ is pseudoeffective and not numerically trivial, the line bundle $\OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)$ is not pseudoeffective for $m\ll0$, hence there are only finitely many negative integers in $\mathcal{T}$. Therefore, we may assume that $\mathcal T\subseteq\N$.
Every nontrivial section of $H^0\big(Y,\mathcal E \otimes \OO_Y(m\pi^*K_X)\big)$ gives an inclusion $\OO_Y(-m\pi^*K_X)\to \mathcal E$, and let $\mathcal F \subseteq\mathcal E$ be the image of the map $$\textstyle\bigoplus_{m\in \mathcal T}\OO_Y(-m\pi^*K_X)\to \mathcal E.$$ Then $\mathcal F$ is quasi-coherent by [@Har77 Proposition II.5.7], and therefore a torsion free coherent sheaf as it is a subsheaf of the torsion free coherent sheaf $\mathcal E$. Let $r$ be the rank of $\mathcal F$. We may assume that there exist infinitely many $r$-tuples $(m_1,\dots,m_r)$ such that the image of the map $$\label{eq:inclusion}
\OO_Y(-m_1\pi^*K_X)\oplus\cdots\oplus\OO_Y(-m_r\pi^*K_X)\to\mathcal F$$ | ) = 0 $ by [ @Kaw85b Theorem 8.2 ].
Arguing by contradiction, assume that there exists $ p\geq1 $ and an infinite specify $ \mathcal T\subseteq \Z$ such that $ $ H^0\big(Y,(\Omega^1_Y)^{\otimes p } \otimes \OO_Y(m\pi^*K_X)\big) \neq 0$$ for all $ m\in\mathcal T$. Denote $ \mathcal E=(\Omega^1_Y)^{\otimes phosphorus } $ and $ Z=\PS(\mathcal E)$ with the projection $ f\colon Z\to Y$. First note that $ $ H^0\big(Y,\mathcal E\otimes \OO_Y(m\pi^*K_X)\big)\simeq H^0\big(Z,\OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)\big).$$ Since $ K_X$ is pseudoeffective and not numerically fiddling, the cable bundle $ \OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)$ is not pseudoeffective for $ m\ll0 $, hence there be entirely finitely many negative integer in $ \mathcal{T}$. consequently, we may assume that $ \mathcal T\subseteq\N$.
Every nontrivial section of $ H^0\big(Y,\mathcal east \otimes \OO_Y(m\pi^*K_X)\big)$ gives an inclusion $ \OO_Y(-m\pi^*K_X)\to \mathcal E$, and lease $ \mathcal F \subseteq\mathcal E$ be the image of the map $ $ \textstyle\bigoplus_{m\in \mathcal T}\OO_Y(-m\pi^*K_X)\to \mathcal E.$$ Then $ \mathcal F$ is quasi - coherent by [ @Har77 Proposition II.5.7 ], and therefore a torsion free coherent bundle as it is a subsheaf of the torsion free coherent sheaf $ \mathcal E$. get $ r$ be the rank of $ \mathcal F$. We may assume that there exist boundlessly many $ r$-tuples $ (m_1,\dots, m_r)$ such that the image of the function $ $ \label{eq: inclusion }
\OO_Y(-m_1\pi^*K_X)\oplus\cdots\oplus\OO_Y(-m_r\pi^*K_X)\to\mathcal F$$ | ) = 0$ hy [@Kaw85b Theorem 8.2].
Arguing ny contradiction, assume that fhere exksts $p\geq1$ and an infinite sev $\marhcal T\subseteq \Z$ such that $$H^0\big(Y,(\Omeha^1_Y)^{\otimew p} \irimes \OO_Y(m\'j^*K_X)\big) \kzq 0$$ fkv all $n\in\mathcal T$. Dgnote $\mathcan E=(\Omega^1_Y)^{\otimev o} $ and $Z=\PS(\mathcal E)$ with the projectijn $f\colpn Z\to Y$. First njte uhae $$H^0\bjg(Y,\mathcal E\otimes \OO_Y(m\pi^*K_X)\big)\simes H^0\big(Z,\MO_Z(1)\otimes f^*\OO_U(m\pi^*K_X)\big).$$ Since $K_X$ is pseufoefvective and not nulerically ttjviwo, the line bjndle $\OO_Z(1)\ouikes f^*\OO_Y(m\pj^*K_X)$ is not pseudoeffective for $o\ll0$, hznce there qrw ojny finitely many negative inbvgers it $\mathcsl{T}$. Therefore, ee kay assume that $\mathcal V\subseteq\N$.
Every nonttivial secdikn of $H^0\big(Y,\mathcao W \otikes \MO_Y(m\ou^*K_X)\cig)$ gmvea an ijclnsion $\OO_Y(-m\pj^*K_X)\to \mathcql E$, and let $\mathcak S \subseteq\mathdal E$ fe the image of the map $$\textstyle\bigoplus_{k\in \mathcal T}\OO_Y(-m\pi^*K_X)\to \marhcal E.$$ Then $\mathcal V$ is quasy-coherent by [@Har77 Proposition II.5.7], and therefore a tmrsioi wret gjfwrfnt sheaf as it is a subsheaf of the torsion sdet cpherent sheaf $\iathcal E$. Lrt $r$ fe the rank ow $\mathezl F$. We may assume tjat thete exiwt infiniuely kany $r$-tuples $(m_1,\dots,m_r)$ such tyat the imagv of the map $$\label{eq:inelusion}
\OO_Y(-m_1\pn^*K_X)\oplos\cdotx\oplus\OO_Y(-m_r\pi^*K_X)\to\mathcau F$$ | ) = 0$ by [@Kaw85b Theorem 8.2]. contradiction, that there $p\geq1$ and an such $$H^0\big(Y,(\Omega^1_Y)^{\otimes p} \otimes \neq 0$$ for $m\in\mathcal T$. Denote $\mathcal E=(\Omega^1_Y)^{\otimes p} and $Z=\PS(\mathcal E)$ with the projection $f\colon Z\to Y$. First note that $$H^0\big(Y,\mathcal \OO_Y(m\pi^*K_X)\big)\simeq H^0\big(Z,\OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)\big).$$ Since $K_X$ is pseudoeffective and not numerically trivial, the line $\OO_Z(1)\otimes is pseudoeffective $m\ll0$, hence there are only finitely many negative integers in $\mathcal{T}$. Therefore, we may assume that T\subseteq\N$. Every nontrivial section of $H^0\big(Y,\mathcal E \otimes gives an inclusion $\OO_Y(-m\pi^*K_X)\to E$, and let $\mathcal F E$ the image the $$\textstyle\bigoplus_{m\in T}\OO_Y(-m\pi^*K_X)\to \mathcal E.$$ $\mathcal F$ is quasi-coherent by [@Har77 Proposition II.5.7], and therefore a torsion free coherent sheaf as it a subsheaf torsion free sheaf E$. $r$ be the $\mathcal F$. We may assume that many $r$-tuples $(m_1,\dots,m_r)$ such that the image of map $$\label{eq:inclusion} F$$ | ) = 0$ by [@Kaw85b Theorem 8.2].
Arguing by conTradiction, AssumE thAt tHeRe exIsts $P\geq1$ and an infinITe seT $\mathcal T\subseteq \Z$ such That $$H^0\BiG(y,(\OmeGA^1_Y)^{\OtimeS p} \otimeS \oO_y(M\Pi^*K_x)\bIg) \Neq 0$$ FoR AlL $m\in\mAthCal T$. DenOte $\mathcal e=(\OmEgA^1_Y)^{\otimes p} $ and $z=\pS(\Mathcal E)$ wiTh tHe projection $F\coLon Z\to y$. FIrsT Note tHat $$h^0\big(Y,\MathcaL e\otimeS \OO_Y(m\pi^*K_X)\BiG)\Simeq H^0\BIg(Z,\OO_Z(1)\oTIMeS f^*\OO_y(m\pi^*K_X)\big).$$ Since $K_X$ iS PsEUdoeffective anD not nuMeRIcALLy tRivIal, the line BuNdle $\Oo_z(1)\otimes F^*\oO_y(M\PI^*K_X)$ IS not pseudoeffEctive for $m\lL0$, HenCe therE aRe oNLy finiTely mAnY NegAtive integeRs in $\Mathcal{T}$. THerefoRE, we may aSSume thaT $\mathcAl T\SubSeteQ\n$.
EVeRy nOnTRivIAl SecTIon Of $H^0\big(Y,\mAtHcAl E \otImes \oo_y(M\Pi^*K_X)\Big)$ GiveS an inClusion $\OO_Y(-m\pi^*k_X)\tO \matHCal e$, and lEt $\matHcal f \sUbsetEq\mathCal E$ bE tHe image of the map $$\TextStyle\bigoPluS_{m\In \mAtHcal T}\oo_Y(-m\pi^*K_x)\to \MatHcal E.$$ ThEn $\mathcAL F$ iS qUASI-cOherent by [@Har77 ProposItION Ii.5.7], and therEfore a TOrSiON free cohErEnt SheaF AS it is A subSHeAf of the tOrsion FReE cOherent ShEaf $\matHcAl E$. let $R$ be thE Rank Of $\mathCal F$. We maY assuME that there exisT Infinitely manY $R$-tUPLeS $(M_1,\dotS,m_r)$ Such that the ImagE Of thE map $$\LAbEl{eQ:IncluSion}
\Oo_Y(-M_1\Pi^*k_x)\oplus\cdots\oplus\OO_Y(-M_r\Pi^*K_X)\to\MathcAl F$$ | ) = 0$ by [@Kaw85b Theorem 8.2].
Ar guing by co nt radi ctio n, assume that ther e exists $p\geq1$ andan in fi n ites et $\ma thcal T \ su b s ete q\Z $ s uc h t hat $ $H^ 0\big(Y ,(\Omega^1 _Y) ^{ \otimes p} \ o ti mes \OO_Y( m\p i^*K_X)\big) \n eq 0$$ f ora ll $m \in \math cal T$ . Denot e $\mathc al E=(\Om e ga^1_Y) ^ { \o time s p} $ and $Z=\PS ( \m a thcal E)$ with the p ro j ec t i on$f\ colon Z\to Y $. Fi r st note th a t $$H ^ 0\big(Y,\math cal E\otime s \O O_Y(m\ pi ^*K _ X)\big )\sim eq H^0 \big(Z,\OO_ Z(1) \otimes f ^*\OO_ Y (m\pi^* K _X)\big ).$$ S inc e $ K_X$ is p seu do e ffe c ti vea ndnot nume ri ca lly t rivi a l , thelin e bu ndle$\OO_Z(1)\oti mes f^* \ OO_ Y(m\p i^*K_ X)$is notpseudo effec ti ve for $m\ll0$, hen ce thereare o nly f inite l y many ne gat ive int egers i n $\ ma t h c al {T}$. Therefore, w em a yassume t hat $\ m at hc a l T\subs et eq\ N$.E verynont r iv ial sect ion of $H ^0 \big(Y, \m athcal E \o tim es \O O _Y(m \pi^*K _X)\big) $ giv e s an inclusion $\OO_Y(-m\pi^ * K_ X ) \t o \ma thc al E$, andlet$ \mat hcal F\su b seteq \math ca l E $ be the image of th emap $$ \text style\bigoplu s_{m\in \m a t h cal T}\O O_Y( - m\ p i^*K_X)\to \ma thcal E.$$ Then $\mathca l F$is quasi -coherent b y [@Har7 7 P rop osi tio n II .5.7], and th e r efor ea torsi onfree co her ent sh eaf a s it is a subshea fof t he to rsion free coh er ent s hea f $\m a thcalE$. L et $ r$ b e th e ranko f$ \ math ca lF$.Wema y ass umet hat thereexist inf ini t elyma ny $r$-tu ples $(m_1,\d ot s,m_r)$ su ch th at the i mage ofthe map $$\label{eq:inc l usion}\OO _Y(-m _1\p i^*K_X)\o plu s\cdot s\o p lus\OO _Y(-m_ r\pi^ *K _X) \ t o\mat h c al F$ $ | ) =_0$ by_[@Kaw85b Theorem 8.2].
Arguing by_contradiction, assume_that_there exists_$p\geq1$_and an infinite_set $\mathcal T\subseteq_\Z$ such that $$H^0\big(Y,(\Omega^1_Y)^{\otimes_p} \otimes \OO_Y(m\pi^*K_X)\big)_\neq_0$$ for all $m\in\mathcal T$. Denote $\mathcal E=(\Omega^1_Y)^{\otimes p} $ and $Z=\PS(\mathcal E)$ with_the_projection $f\colon_Z\to_Y$._First note that $$H^0\big(Y,\mathcal E\otimes_\OO_Y(m\pi^*K_X)\big)\simeq H^0\big(Z,\OO_Z(1)\otimes f^*\OO_Y(m\pi^*K_X)\big).$$ Since $K_X$_is pseudoeffective_and not numerically trivial, the line bundle $\OO_Z(1)\otimes_f^*\OO_Y(m\pi^*K_X)$_is not pseudoeffective_for $m\ll0$, hence there are only finitely many negative_integers in $\mathcal{T}$. Therefore, we may_assume that $\mathcal_T\subseteq\N$.
Every_nontrivial_section of $H^0\big(Y,\mathcal E_\otimes \OO_Y(m\pi^*K_X)\big)$ gives an inclusion $\OO_Y(-m\pi^*K_X)\to_\mathcal E$, and let $\mathcal F_\subseteq\mathcal E$ be the image of the_map $$\textstyle\bigoplus_{m\in \mathcal T}\OO_Y(-m\pi^*K_X)\to \mathcal E.$$_Then $\mathcal F$ is quasi-coherent_by [@Har77_Proposition II.5.7], and therefore a_torsion free coherent_sheaf as_it is a_subsheaf of the torsion free coherent_sheaf $\mathcal E$._Let $r$ be the rank of_$\mathcal_F$. We may_assume_that_there exist_infinitely many $r$-tuples_$(m_1,\dots,m_r)$_such that_the_image of the map $$\label{eq:inclusion}
\OO_Y(-m_1\pi^*K_X)\oplus\cdots\oplus\OO_Y(-m_r\pi^*K_X)\to\mathcal F$$ |
took full account of the suppression of scalar gradients by diffusive effects.
This study extends the above result to the three-dimensional case, with an improvement in the method of analysis by making use of an analytic inequality. For flows having bounded velocity gradients, the maximally achievable decay rate $2\kappa\langle|\nabla\theta|^2\rangle$ is shown to be bounded from above. Furthermore, in the limit $\kappa\to0$, the derived upper bound vanishes provided that the scalar field remains marginally smooth. More precisely, if there exists a positive $\alpha_0$ independent of $\kappa$ such that $\langle|(-\Delta)^{\alpha/2}\theta|^2\rangle<\infty$ for $\alpha\le\alpha_0$, then $\kappa\langle|\nabla\theta|^2\rangle\le C\kappa^{\alpha_0}$. Here, $\alpha_0\in(0,1]$ and $C$ is a constant, depending on $|\nabla\u|$ and $\langle|(-\Delta)^{\alpha_0/2}\theta|^2\rangle$, both being bounded and independent of $\kappa$. This constraint covers all scalar distribution having variance spectrum $\Theta(k)$ steeper than $k^{-1}$ at the spectral tail, i.e. at large wave numbers $k$. This is an improvement over the result of Tran and Dritschel$^{23}$ for the two-dimensional case, where power-law spectra were assumed explicitly in their calculations. When no such positive $\alpha_0$ exists, a plausible scenario consistent with Batchelor’s theory$^{3}$ is that $\Theta(k)$ becomes increasingly shallower for smaller $\kappa$, approaching the Batchelor scaling $k^{-1}$ in the limit $\kappa\to0$. For this classical case, the semi-analytic estimate of Tran and Dritschel can be applied, yielding a vanishing decay rate no slower than $(\ln P_r)^{-1}$. Hence, diffusion anomaly is ruled out for a broad class of scalar distribution, including power-law spectra no shallower than $k^{-1}$. The implication is that a non-vanishing decay rate in the limit $\kappa\to0$ necessarily requires the variance at small scales to be greater than that allowed by the Batchelor spectrum. This is consistent with recent studies$^{1,7,14,19}$ suggesting spectra shallower than $k^{-1}$. This analysis when applied to the forced case implies that for | took full account of the suppression of scalar gradient by diffusing effects.
This study cover the above solution to the three - dimensional case, with an improvement in the method of psychoanalysis by making use of an analytic inequality. For flows having jump velocity gradients, the maximally accomplishable decay rate $ 2\kappa\langle|\nabla\theta|^2\rangle$ is shown to be bounded from above. Furthermore, in the limit $ \kappa\to0 $, the derive upper bound vanishes provide that the scalar field remains marginally politic. More precisely, if there exists a positive $ \alpha_0 $ independent of $ \kappa$ such that $ \langle|(-\Delta)^{\alpha/2}\theta|^2\rangle<\infty$ for $ \alpha\le\alpha_0 $, then $ \kappa\langle|\nabla\theta|^2\rangle\le C\kappa^{\alpha_0}$. Here, $ \alpha_0\in(0,1]$ and $ C$ is a ceaseless, depending on $ |\nabla\u|$ and $ \langle|(-\Delta)^{\alpha_0/2}\theta|^2\rangle$, both being bounded and independent of $ \kappa$. This constraint cover all scalar distribution having variance spectrum $ \Theta(k)$ steeper than $ k^{-1}$ at the spectral tail, i.e. at big wave numbers $ k$. This is an improvement over the result of Tran and Dritschel$^{23}$ for the two - dimensional case, where power - law spectra were assumed explicitly in their calculations. When no such positive $ \alpha_0 $ exists, a plausible scenario consistent with Batchelor ’s theory$^{3}$ is that $ \Theta(k)$ becomes increasingly shallower for smaller $ \kappa$, approaching the Batchelor scaling $ k^{-1}$ in the limit $ \kappa\to0$. For this classical event, the semi - analytic estimate of Tran and Dritschel can be applied, concede a fly decay rate no slower than $ (\ln P_r)^{-1}$. Hence, dissemination anomaly is ruled out for a broad class of scalar distribution, including baron - law spectra no shallower than $ k^{-1}$. The implication is that a non - vanishing decay rate in the terminus ad quem $ \kappa\to0 $ necessarily requires the variance at small scales to be greater than that allowed by the Batchelor spectrum. This is reproducible with recent studies$^{1,7,14,19}$ suggesting spectra shallower than $ k^{-1}$. This analysis when applied to the forced shell implies that for | tolk full account of the smppression of sccoar gredients by diffjsive effects.
This study exteids rhe avove result to the thrde-dimensilnal casw, wiuh an improvement in the method lf auaoysis by makinn use of an analytic ineqgauicy. For flows having bounded velocity gradiemtd, the maximalli achpedabls decay rate $2\kappa\langle|\nabla\theta|^2\dangle$ ps shown to be boinded from above. Furthermoge, ij the limit $\kappa\tl0$, the derivgs u[per bound vavishes pronnded that tge scalar field remains marginauly skooth. More peeclvely, if thece exifts a positiyv $\alpha_0$ indepemdent of $\kappa$ surh tyat $\langle|(-\Delta)^{\alpha/2}\tieta|^2\rangle<\infty$ for $\wlpha\le\al[hc_0$, then $\kappa\langle|\naboa\rheta|^2\tangla\le Z\japoa^{\ampia_0}$. Gere, $\apphe_0\in(0,1]$ and $C$ ia a constanr, depending on $|\nabls\u|$ qnd $\langle|(-\Delfa)^{\alphw_0/2}\treta|^2\rangle$, both being bounded and indeptndenf of $\kappa$. This constraunt covers all scalar distribueion having variance spectrum $\Theta(k)$ steeper than $k^{-1}$ at ghe siectfql tail, i.e. at large wave numbers $k$. This is an iilrpvvment over the refult of Tram wnc Dritschel$^{23}$ fot the tck-djmensional case, whfre powgr-law wpectra wtre axsumed explicitly in their xalculations. Qhen no such positnve $\alpha_0$ exnsts, a plauxible scenario consisteut witg Batchelor’d theory$^{3}$ jr that $\Theta(k)$ bezomvs itcreasingly shallower for fmaller $\keppa$, cpproachkng jhe Batshelor scaping $k^{-1}$ in the limit $\kappw\to0$. Flr this clasdical case, the semi-analytic estmkate of Tran atd Gritschej can be applied, yyelding a vanixhing dzcay rxte no sloser thai $(\ln P_r)^{-1}$. Hencq, diffusion atlmaly is rulxd out fow a vroae class uf scalar disttibution, pneluding piwer-law spectra no shxmlower than $k^{-1}$. Tkt inplication is tnat a goj-venishytg decay rata in thd limig $\kappa\uj0$ kecdssatily requires the vasiande at small scales tj be greqter thag that allowec by the Batchelor specvrum. Tiis is cogsistent with recent studies$^{1,7,14,19}$ shggesting spcctra shallowqr tman $h^{-1}$. This anakysis when applied to the forced case inplies that for | took full account of the suppression of by effects. This extends the above with improvement in the of analysis by use of an analytic inequality. For having bounded velocity gradients, the maximally achievable decay rate $2\kappa\langle|\nabla\theta|^2\rangle$ is shown to bounded from above. Furthermore, in the limit $\kappa\to0$, the derived upper bound vanishes that scalar remains smooth. More precisely, if there exists a positive $\alpha_0$ independent of $\kappa$ such that $\langle|(-\Delta)^{\alpha/2}\theta|^2\rangle<\infty$ for then $\kappa\langle|\nabla\theta|^2\rangle\le C\kappa^{\alpha_0}$. Here, $\alpha_0\in(0,1]$ and $C$ is constant, depending on $|\nabla\u|$ $\langle|(-\Delta)^{\alpha_0/2}\theta|^2\rangle$, both being bounded and of This constraint all distribution variance spectrum $\Theta(k)$ than $k^{-1}$ at the spectral tail, i.e. at large wave numbers $k$. This is an improvement over result of Dritschel$^{23}$ for two-dimensional where spectra were assumed their calculations. When no such positive plausible scenario consistent with Batchelor’s theory$^{3}$ is that becomes increasingly for smaller $\kappa$, approaching the Batchelor $k^{-1}$ in the limit $\kappa\to0$. For this classical the semi-analytic estimate of Tran and Dritschel can be applied, yielding a vanishing decay rate than $(\ln P_r)^{-1}$. Hence, anomaly is ruled for broad of distribution, including spectra no shallower than $k^{-1}$. The implication is that a non-vanishing rate in the limit $\kappa\to0$ necessarily requires the variance at to greater than that by the Batchelor spectrum. is with recent studies$^{1,7,14,19}$ suggesting than This to forced implies that for | took full account of the supprEssion of scAlar gRadIenTs By diFfusIve effects.
This STudy Extends the above result tO the tHrEE-dimENsIonal Case, witH An IMProVeMeNt iN tHE mEthod Of aNalysis By making usE of An Analytic ineqUAlIty. For flowS haVing bounded vEloCity grAdIenTS, the mAxiMally AchievABle decAy rate $2\kapPa\LAngle|\nABla\thetA|^2\RAnGle$ iS shown to be bounded FRoM Above. FurthermoRe, in thE lIMiT $\KAppA\to0$, The derived UpPer boUNd vanisHEs PROVidED that the scalaR field remaiNS maRginalLy SmoOTh. More PreciSeLY, if There exists A posItive $\alphA_0$ indepENdent of $\KAppa$ sucH that $\lAngLe|(-\DElta)^{\ALpHa/2}\TheTa|^2\RAngLE<\iNftY$ For $\Alpha\le\aLpHa_0$, Then $\kAppa\LANGLe|\naBla\ThetA|^2\rangLe\le C\kappa^{\alpHa_0}$. HEre, $\aLPha_0\In(0,1]$ and $c$ is a cOnstAnT, depeNding oN $|\nablA\u|$ And $\langle|(-\Delta)^{\aLpha_0/2}\Theta|^2\rangLe$, bOtH beInG bounDEd and iNdePenDent of $\kAppa$. ThiS ConStRAINt Covers all scalar disTrIBUtIon havinG variaNCe SpECtrum $\TheTa(K)$ stEepeR THan $k^{-1}$ aT the SPeCtral taiL, i.e. at lARgE wAve numbErS $k$. This Is An iMprOvemeNT oveR the reSult of TrAn and dRitschel$^{23}$ for the TWo-dimensional CAsE, WHeRE powEr-lAw spectra weRe asSUmed ExplICiTly IN theiR calcUlATiONs. When no such positivE $\aLpha_0$ exIsts, a Plausible scenArio consisTENT with BatChelOR’s THeory$^{3}$ is that $\TheTa(k)$ beComes increASingly shAllowEr for smaLler $\kappa$, APProachinG thE BaTchEloR SCaLing $k^{-1}$ in the limIT $\KappA\tO0$. For thiS clAssical CasE, thE seMi-aNaLytic estiMate of TrAn AnD DRiTscHel caN Be applieD, yIelDiNg a VanisHIng decAy ratE no sLoWeR ThaN $(\ln P_r)^{-1}$. HeNCe, DIFfusIoN aNomaLy iS rUled oUt foR A brOad clasS of scalar DisTRibuTiOn, IncludiNg power-law speCtRa no shalloWeR thAn $k^{-1}$. The IMPlicatioN is that a non-vanishing decAY rate in The Limit $\KappA\to0$ necessAriLy requIreS The varIance aT smalL sCalES To be gREAtEr tHaN that allowED By tHe BatChElor SpectruM. This is consistent wITh rEcent studies$^{1,7,14,19}$ sUggEstiNG SpEctRA sHAllOwER thAN $K^{-1}$. This analysis whEn applied tO tHE fOrced case iMPliEs That for | took full account of thesuppressio n ofsca lar g radi ents by diffusivee ffec ts.
This study extend s the a b over es ult t o the t h re e - dim en si ona lc as e, wi than impr ovement in th emethod of an a ly sis by mak ing use of an a nal ytic i ne qua l ity.For flow s havi n g boun ded veloc it y gradi e nts, th e ma xima lly achievable de c ay rate $2\kappa\ langle |\ n ab l a \th eta |^2\rangle $is sh o wn to b e b o u n ded from above. F urthermore, inthe li mi t $ \ kappa\ to0$, t h e d erived uppe r bo und vanis hes pr o vided t h at thescalar fi eld rem a in smar gi n all y s moo t h.More pre ci se ly, i f th e r e exis tsa po sitiv e $\alpha_0$ind epen d ent of $ \kapp a$ s uc h tha t $\la ngle| (- \Delta)^{\alpha /2}\ theta|^2\ ran gl e<\ in fty$f or $\a lph a\l e\alpha _0$, th e n $ \k a p p a\ langle|\nabla\thet a| ^ 2 \r angle\le C\kap p a^ {\ a lpha_0}$ .Her e, $ \ a lpha_ 0\in ( 0, 1]$ and$C$ is aco nstant, d ependi ng on $| \nabl a \u|$ and $ \langle| (-\De l ta)^{\alpha_0/ 2 }\theta|^2\ra n gl e $ ,b othbei ng boundedandi ndep ende n tof$ \kapp a$. T hi s c o nstraint covers all s calardistr ibution havin g variance s p ectrum $ \The t a( k )$ steeper tha n $k^ {-1}$ at t h e spectr al ta il, i.e. at large w ave numb ers $k $.Thi s is an improveme n t ove rthe res ult of Tra n a ndDri tsc he l$^{23}$for thetw o- di me nsi onalc ase, whe re po we r-l aw sp e ctra w ere a ssum ed e x pli citly i n t h e ir c al cu lati ons .Whenno s u chpositiv e $\alpha _0$ exis ts ,a plaus ible scenario c onsistentwi thBatche l o r’s theo ry$^{3}$ is that $\Thet a (k)$ be com es in crea singly sh all ower f ors maller $\kap pa$,ap pro a c hingt h eBat ch elor scali n g $k ^{-1} $in t he limi t $\kappa\to0$. Fo r th is classicalcas e, t h e s emi - an a lyt ic est i m ate of Tran and Dritschel c a nbe applied , yi el ding avanishi ng de c ay rate no slowe r than $( \l n P_ r ) ^{- 1}$. Hence , diffus ion anoma l y isr ul ed ou t f or a b ro adclass of sc a lar dist ributi on , incl uding p ower-law spectra no shallower t han $k ^{-1} $.The impli cat i onis that a non -vanishing de cay rate in the l imit $\ kap p a\to0 $ ne c essarilyr eq uir e s t he variance a t sm all s cal e s to b e gr eater than that a l lowed by the B atch e l orspe c trum .This is consis ten tw i th recen tstudies$^{1 ,7,14,19 }$ sugge stingspectr a shall o w er than $ k^{- 1}$ . This an aly si s when a pp li e d to t he f or ced ca se imp l iest h at for | took_full account_of the suppression of_scalar gradients_by_diffusive effects.
This_study_extends the above_result to the_three-dimensional case, with an_improvement in the_method_of analysis by making use of an analytic inequality. For flows having bounded velocity_gradients,_the maximally_achievable_decay_rate $2\kappa\langle|\nabla\theta|^2\rangle$ is shown to_be bounded from above. Furthermore,_in the_limit $\kappa\to0$, the derived upper bound vanishes provided_that_the scalar field_remains marginally smooth. More precisely, if there exists a_positive $\alpha_0$ independent of $\kappa$ such_that $\langle|(-\Delta)^{\alpha/2}\theta|^2\rangle<\infty$ for_$\alpha\le\alpha_0$,_then_$\kappa\langle|\nabla\theta|^2\rangle\le C\kappa^{\alpha_0}$. Here, $\alpha_0\in(0,1]$_and $C$ is a constant, depending_on $|\nabla\u|$ and $\langle|(-\Delta)^{\alpha_0/2}\theta|^2\rangle$, both being_bounded and independent of $\kappa$. This constraint_covers all scalar distribution having variance_spectrum $\Theta(k)$ steeper than $k^{-1}$_at the_spectral tail, i.e. at large_wave numbers $k$._This is_an improvement over_the result of Tran and Dritschel$^{23}$_for the two-dimensional_case, where power-law spectra were assumed_explicitly_in their calculations._When_no_such positive_$\alpha_0$ exists, a_plausible_scenario consistent_with_Batchelor’s theory$^{3}$ is that $\Theta(k)$ becomes_increasingly_shallower for smaller $\kappa$, approaching the Batchelor_scaling $k^{-1}$ in the_limit_$\kappa\to0$. For this classical_case, the semi-analytic estimate of_Tran and Dritschel can be applied,_yielding a_vanishing decay_rate no slower than $(\ln P_r)^{-1}$. Hence, diffusion anomaly is ruled_out for a broad class of_scalar distribution, including power-law_spectra no_shallower_than $k^{-1}$. The_implication_is that_a non-vanishing decay rate in the limit_$\kappa\to0$ necessarily_requires the variance at small scales_to be greater than_that_allowed by the Batchelor spectrum. This_is consistent with recent studies$^{1,7,14,19}$ suggesting_spectra shallower than $k^{-1}$. This_analysis_when_applied to the forced case_implies that for |
21-734.
Wang, Y, and Wagner, S.: Towards applying a safety analysis and verification method based on STPA to agile software development. IEEE/ACM International Workshop on Continuous Software Evolution and Delivery (CSED). IEEE, 2016.
Wang, Y, Bogicevic, I. and Wagner, S.: A study of safety documentation in a Scrum development process. Proceedings of the XP2017 Scientific Workshops. ACM, 2017.
Theocharis, G., Kuhrmann, M., M[ü]{}nch, J., and Diebold, P.: Is water-scrum-fall reality? on the use of agile and traditional development practices. In International Conference on Product-Focused Software Process Improvement (pp. 149-166). Springer International Publishing.
Falessi, D., Juristo, N., Wohlin, C., Turhan, B., M[ü]{}nch, J., Jedlitschka, A., and Oivo, M.: Empirical software engineering experts on the use of students and professionals in experiments. In Journal of Empirical Software Engineering. Springer, 2017.
[^1]: https://www.kaaproject.org/overview/
[^2]: https://github.com/ywISTE/student-project—Smart-Home
[^3]: The questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^4]: http://www.xstampp.de/
[^5]: The interview guideline is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^6]: The questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^7]: The interview guideline is available: https://zenodo.org/record/439696\#.WODCovl96Uk
---
address: |
Service de Physique Théorique, CEA, CE-Saclay\
F-91191 Gif-sur-Yvette Cedex, France
author:
- 'S. MUNIER'
title: '$k_\bot$-FACTORIZATION PREDICTIONS FOR $F_2^c$ AT HERA[^1] ' | 21 - 734.
Wang, Y, and Wagner, S.: Towards applying a safety analysis and verification method acting free-base on STPA to agile software development. IEEE / ACM International Workshop on Continuous Software Evolution and Delivery (CSED). IEEE, 2016.
Wang, Y, Bogicevic, I. and Wagner, S.: A study of condom software documentation in a Scrum development process. Proceedings of the XP2017 Scientific Workshops. ACM, 2017.
Theocharis, G., Kuhrmann, M., M[ü]{}nch, J., and Diebold, P.: Is body of water - scrum - fall reality? on the function of agile and traditional development practices. In International Conference on intersection - Focused Software Process Improvement (pp. 149 - 166). Springer International Publishing.
Falessi, D., Juristo, N., Wohlin, C., Turhan, B., M[ü]{}nch, J., Jedlitschka, A., and Oivo, M.: Empirical software engineering expert on the use of students and professionals in experiments. In Journal of Empirical Software Engineering. Springer, 2017.
[ ^1 ]: https://www.kaaproject.org/overview/
[ ^2 ]: https://github.com/ywISTE/student-project—Smart-Home
[ ^3 ]: The questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[ ^4 ]: http://www.xstampp.de/
[ ^5 ]: The interview guideline is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[ ^6 ]: The questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[ ^7 ]: The interview road map is available: https://zenodo.org/record/439696\#.WODCovl96Uk
---
address: |
Service de Physique Théorique, CEA, CE - Saclay\
F-91191 Gif - sur - Yvette Cedex, France
author:
-' S. MUNIER'
title:' $ k_\bot$-FACTORIZATION prediction FOR $ F_2^c$ AT HERA[^1 ]' | 21-734.
Wanh, Y, and Wagner, S.: Towards applying a safgtt analbsis ans verifization method based on STPA vo atile woftware development. IDEE/ACM Innernationql Wiekshop on Rkntinuomf Sorbware Xvolution and Dglivery (CSED). IEEE, 2016.
Wang, Y, Bmgkczvic, I. and Wagner, S.: A study of safetr documrnhation in a Sctum dtvejopmsnt process. Proceedings of the XP2017 Acientihic Workshops. AVM, 2017.
Theocharis, G., Kuhrmann, M., M[ü]{}ncj, J., and Diebold, P.: Ls water-scroj-fajo reality? on the use of agile and jraditional development practicer. In Nnternationql Cojxerence on 'roducn-Focused Software Prowess Imlrovement (pp. 149-166). Xprmngee International Publivhing.
Falessi, D., Jurysto, N., Wovlnn, C., Turhan, B., M[ü]{}nch, J., Heelitswhka, A., ave Okvo, M.: Ejpiricwl aoftware ehgineering wxperts on the use pf wtudents and lrofesfijnals in experiments. In Journal of Empigicam Software Engineering. Wpringer, 2017.
[^1]: https://www.kaakroject.org/jverview/
[^2]: https://github.com/ywISTE/student-project—Smart-Hmme
[^3]: Tie qutsblonnxurf is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^4]: rftk://wwe.xstampp.de/
[^5]: The interview buldrjine is availxble: hcfpa://zenodo.org/record/439696\#.WOFCovl96Uk
[^6]: The wuestionnwire is available: https://zenodo.ort/record/439696\#.WODConl96Uk
[^7]: The interview guibeline is avcilablg: httpx://zenodo.org/record/439696\#.WODCovl96Bk
---
adddess: |
Segvice de Lfysique Théorique, CES, WE-Saclay\
F-91191 Gif-sur-Yvette Cedex, Frence
abthor:
- 'S. OUNIGR'
title: '$k_\bot$-FACTOGIZATLMN PREDICTIONS FOR $F_2^c$ AJ HERA[^1] ' | 21-734. Wang, Y, and Wagner, S.: Towards safety and verification based on STPA International on Continuous Software and Delivery (CSED). 2016. Wang, Y, Bogicevic, I. and S.: A study of safety documentation in a Scrum development process. Proceedings of XP2017 Scientific Workshops. ACM, 2017. Theocharis, G., Kuhrmann, M., M[ü]{}nch, J., and Diebold, Is reality? the of agile and traditional development practices. In International Conference on Product-Focused Software Process Improvement (pp. 149-166). International Publishing. Falessi, D., Juristo, N., Wohlin, C., B., M[ü]{}nch, J., Jedlitschka, and Oivo, M.: Empirical software experts the use students professionals experiments. In Journal Empirical Software Engineering. Springer, 2017. [^1]: https://www.kaaproject.org/overview/ [^2]: https://github.com/ywISTE/student-project—Smart-Home [^3]: The questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk [^4]: http://www.xstampp.de/ The interview available: https://zenodo.org/record/439696\#.WODCovl96Uk The is https://zenodo.org/record/439696\#.WODCovl96Uk [^7]: The is available: https://zenodo.org/record/439696\#.WODCovl96Uk --- address: | Théorique, CEA, CE-Saclay\ F-91191 Gif-sur-Yvette Cedex, France author: 'S. MUNIER' '$k_\bot$-FACTORIZATION PREDICTIONS FOR $F_2^c$ AT HERA[^1] | 21-734.
Wang, Y, and Wagner, S.: Towards appLying a safeTy anaLysIs aNd VeriFicaTion method baseD On STpA to agile software develOpmenT. IeeE/ACm inTernaTional WORkSHOp oN COnTinUoUS SOftwaRe EVolutioN and DeliveRy (CsEd). IEEE, 2016.
Wang, Y, BoGIcEvic, I. and WaGneR, S.: A study of saFetY documEnTatIOn in a scrUm devElopmeNT proceSs. ProceedInGS of the xp2017 ScientIFIc workShops. ACM, 2017.
TheochariS, g., KUHrmann, M., M[ü]{}nch, J., aNd DiebOlD, p.: IS WAteR-scRum-fall reaLiTy? on tHE use of aGIlE AND trADitional develOpment practICes. in InteRnAtiONal ConFerenCe ON PrOduct-FocuseD SofTware ProcEss ImpROvement (PP. 149-166). SpringEr InteRnaTioNal PUBlIsHinG.
FALesSI, D., jurISto, n., Wohlin, C., tuRhAn, B., M[ü]{}Nch, J., jEDLItscHka, a., and oivo, M.: empirical softWarE engINeeRing eXpertS on tHe Use of StudenTs and PrOfessionals in exPeriMents. In JoUrnAl Of EMpIricaL softwaRe ENgiNeering. springeR, 2017.
[^1]: HttPs://WWW.KaAproject.org/overvieW/
[^2]: hTTPs://Github.coM/ywISTe/StUdENt-projecT—SMarT-HomE
[^3]: tHe queStioNNaIre is avaIlable: HTtPs://Zenodo.oRg/Record/439696\#.wOdCoVl96UK
[^4]: http://WWw.xsTampp.dE/
[^5]: The inteRview GUideline is avaiLAble: https://zenoDO.oRG/ReCOrd/439696\#.WoDCOvl96Uk
[^6]: The queStioNNairE is aVAiLabLE: httpS://zenoDo.ORg/REcord/439696\#.WODCovl96Uk
[^7]: The inTeRview gUidelIne is availablE: https://zenoDO.ORg/record/439696\#.wODCOVl96uK
---
address: |
ServicE de PhYsique ThéoRIque, CEA, Ce-SaclAy\
F-91191 Gif-suR-Yvette CeDEX, France
aUthOr:
- 'S. mUNiER'
TITlE: '$k_\bot$-FACTORIZatiON PrEdICTIONs FOr $F_2^c$ AT HErA[^1] ' | 21-734.
Wang, Y, and Wagn er, S.: To wards ap ply in g asafe ty analysis an d ver ification method based on S TP A toa gi le so ftwared ev e l opm en t. IE EE / AC M Int ern ational WorkshoponCo ntinuous Sof t wa re Evoluti onand Delivery (C SED).IE EE, 2016.
W ang,Y, Bog i cevic, I. and W ag n er, S. : A stud y of saf ety documentation in a Scrum develo pmentpr o ce s s . P roc eedings of t he XP 2 017 Sci e nt i f i c W o rkshops. ACM, 2017.
The o cha ris, G ., Ku h rmann, M.,M[ ü ]{} nch, J., an d Di ebold, P. : Is w a ter-scr u m-fallrealit y?ontheu se o f a gi l e a n dtra d iti onal dev el op mentprac t i c e s. I n I nter natio nal Conferenc e o n Pr o duc t-Foc usedSoft wa re Pr ocessImpro ve ment (pp. 149-1 66). Springer In te rna ti onalP ublish ing .
Falessi , D., J u ris to , N ., Wohlin, C., Turha n, B ., M[ü]{}n ch, J. , J ed l itschka, A .,andO i vo, M .: E m pi rical so ftware en gi neering e xperts o n t heuse o f stu dentsand prof essio n als in experim e nts. In Journ a lo f E m piri cal Software E ngin e erin g. S p ri nge r , 201 7.
[ ^1 ] :h ttps://www.kaaproje ct .org/o vervi ew/
[^2]: ht tps://gith u b . com/ywIS TE/s t ud e nt-project—Sma rt-Ho me
[^3]:T he quest ionna ire is a vailable: h ttps://z eno do. org /re c o rd /439696\#.WOD C o vl96 Uk
[^4]: ht tp://ww w.x sta mpp .de /
[^5]: Th e interv ie wgu id eli ne is availabl e: ht tp s:/ /zeno d o.org/ recor d/43 96 96 \ #.W ODCovl9 6 Uk [^6] :Th e qu est io nnair e is ava ilable: https:// zen o do.o rg /r ecord/4 39696\#.WODCo vl 96Uk
[^7] :The inter v i ew guide line is available: http s ://zeno do. org/r ecor d/439696\ #.W ODCovl 96U k
---addres s: | S e r viced e P hys iq ue Théoriq u e , C EA, C E- Sacl ay\
F-91191 Gif-sur-Y v ett e Cedex, Fran ceauth o r :- ' S .M UNI ER '
ti t l e: '$k_\bot$-FA CTORIZATIO NP RE DICTIONS F O R $ F_ 2^c$ AT HERA[^ 1] ' | 21-734.
Wang, Y,_and Wagner,_S.: Towards applying a_safety analysis_and_verification method_based_on STPA to_agile software development._IEEE/ACM International Workshop on_Continuous Software Evolution_and_Delivery (CSED). IEEE, 2016.
Wang, Y, Bogicevic, I. and Wagner, S.: A study of safety_documentation_in a_Scrum_development_process. Proceedings of the XP2017_Scientific Workshops. ACM, 2017.
Theocharis, G.,_Kuhrmann, M.,_M[ü]{}nch, J., and Diebold, P.: Is water-scrum-fall reality?_on_the use of_agile and traditional development practices. In International Conference on_Product-Focused Software Process Improvement (pp. 149-166)._Springer International Publishing.
Falessi,_D.,_Juristo,_N., Wohlin, C., Turhan,_B., M[ü]{}nch, J., Jedlitschka, A., and_Oivo, M.: Empirical software engineering experts_on the use of students and professionals_in experiments. In Journal of Empirical_Software Engineering. Springer, 2017.
[^1]: https://www.kaaproject.org/overview/
[^2]:_https://github.com/ywISTE/student-project—Smart-Home
[^3]: The_questionnaire is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^4]: http://www.xstampp.de/
[^5]:_The interview guideline_is available:_https://zenodo.org/record/439696\#.WODCovl96Uk
[^6]: The questionnaire_is available: https://zenodo.org/record/439696\#.WODCovl96Uk
[^7]: The interview guideline_is available: https://zenodo.org/record/439696\#.WODCovl96Uk
_---
address: |
Service_de_Physique Théorique, CEA,_CE-Saclay\
__ _F-91191 Gif-sur-Yvette Cedex,_France
author:
-_'S. MUNIER'
title:_'$k_\bot$-FACTORIZATION_PREDICTIONS FOR $F_2^c$ AT HERA[^1] ' |
n=\sqrt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, there exists a constant $K_2>0$ such that $\|\xi'_n\|_\infty\leq K_2 n^{5/12}$.
Consider the positive self-adjoint operator $A$ (whose inverse is Hilbert–Schmidt) given by $A=-\frac{\mathrm{d}^2}{\mathrm{d}x^2}+(1+x^2)$. We have $A\xi_n=(2n+2)\xi_n$.
For any $p\ge 0$, define the space $\mathcal{S}_p({\mathbb{R}})$ to be the domain of the closure of $A^p$. Endowed with the norm $|\xi|_p:=|A^p\xi|_0$, it is a Hilbert space. Note that the norm $|\cdot|_p$ can be expressed as follows, if one uses the orthonormal basis $(\xi_n)$: $$|\xi|_p^2=\sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{2p}.$$ We denote by $\mathcal{S}'_p({\mathbb{R}})$ the dual of $\mathcal{S}_p({\mathbb{R}})$. The norm in $S'_p({\mathbb{R}})$ is given by (see [@ob], Lemma 1.2.8) $$|\xi|^2_{-p}=\sum_{n=0} ^\infty |\la \xi,A^{-p}\xi_n\ra |^2
=\sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{-2p}$$ for any $\xi\in \mathcal{S}'_p({\mathbb{R}})$. One can show that the projective limit of the spaces $\mathcal{S}_{p}({\mathbb{R}})$, $p\geq 0$, is $\mathcal{S}({\mathbb{R}})$, that the inductive limit of the spaces $ \mathcal{S}_p({\mathbb{R}})'$, $p\ge 0$, is $\mathcal{S}'({\mathbb{R}})$ and that $$\mathcal{S}({\mathbb{R}})\subset L^2({\mathbb{R}})\subset \mathcal{S}'({\mathbb{R}})$$ is a Gel’fand triple.
We | n=\sqrt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, there exists a constant $ K_2>0 $ such that $ \|\xi'_n\|_\infty\leq K_2 n^{5/12}$.
Consider the positive self - adjoint hustler $ A$ (whose inverse is Hilbert – Schmidt) render by $ A=-\frac{\mathrm{d}^2}{\mathrm{d}x^2}+(1+x^2)$. We have $ A\xi_n=(2n+2)\xi_n$.
For any $ p\ge 0 $, define the quad $ \mathcal{S}_p({\mathbb{R}})$ to be the world of the closure of $ A^p$. Endowed with the average $ |\xi|_p:=|A^p\xi|_0 $, it is a Hilbert space. Note that the norm $ |\cdot|_p$ can be express as follows, if one uses the orthonormal footing $ (\xi_n)$: $ $ |\xi|_p^2=\sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{2p}.$$ We denote by $ \mathcal{S}'_p({\mathbb{R}})$ the dual of $ \mathcal{S}_p({\mathbb{R}})$. The norm in $ S'_p({\mathbb{R}})$ is given by (see [ @ob ], Lemma 1.2.8) $ $ |\xi|^2_{-p}=\sum_{n=0 } ^\infty |\la \xi, A^{-p}\xi_n\ra |^2
= \sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{-2p}$$ for any $ \xi\in \mathcal{S}'_p({\mathbb{R}})$. One can usher that the projective limit of the spaces $ \mathcal{S}_{p}({\mathbb{R}})$, $ p\geq 0 $, is $ \mathcal{S}({\mathbb{R}})$, that the inductive limit of the spaces $ \mathcal{S}_p({\mathbb{R}})'$, $ p\ge 0 $, is $ \mathcal{S}'({\mathbb{R}})$ and that $ $ \mathcal{S}({\mathbb{R}})\subset L^2({\mathbb{R}})\subset \mathcal{S}'({\mathbb{R}})$$ is a Gel’fand triple.
We | n=\sqgt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, tmere exists a couwtant $N_2>0$ such that $\|\xi'_v\|_\infty\leq K_2 n^{5/12}$.
Consider the podirive welf-adjoint operator $A$ (whose innerse is Yilbtrt–Schmidt) given ug $A=-\frac{\mathrm{s}^2}{\lathxm{v}x^2}+(1+x^2)$. We have $A\xi_k=(2n+2)\xi_n$.
For anf $p\ge 0$, define dhd dpace $\mathcal{S}_p({\mathbb{R}})$ to be the doiain of tje closure of $W^p$. Emqowes with the norm $|\xi|_p:=|A^p\xi|_0$, it is a Himbert skace. Note that the norm $|\cdot|_p$ can be expressfd ad follows, if one udes the ortyonownal basis $(\xi_v)$: $$|\xi|_p^2=\sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{2p}.$$ We denote by $\matfcal{S}'_'({\mathbb{R}})$ thg suwn of $\mathcao{S}_p({\manhbb{R}})$. The norm in $S'_p({\mdthbb{R}})$ os given by (sec [@ob], Nemna 1.2.8) $$|\xi|^2_{-p}=\sum_{n=0} ^\infty |\la \ei,A^{-p}\xi_n\ra |^2
=\sum_{n=0}^\infty \langle \xh,\xn_n\rangle^2(2n+2)^{-2p}$$ for any $\xi\un \mathwal{S}'_[({\matfvb{R}})$. Ont cen ahow tjat the projedtive limit of the spaces $\mathvaj{W}_{p}({\mathbb{R}})$, $p\ges 0$, is $\iaehcal{S}({\mathbb{R}})$, that the inductive limit mf fhe spaces $ \mathcal{S}_p({\marhbb{R}})'$, $p\ge 0$, is $\mathcal{D}'({\mathbb{R}})$ wnd that $$\mathcal{S}({\mathbb{R}})\subset L^2({\mathbb{R}})\subset \matvcal{S}'({\jxthyn{V}})$$ is q Hel’fand triple.
We | n=\sqrt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, there exists a constant $K_2>0$ such K_2 Consider the self-adjoint operator $A$ by We have $A\xi_n=(2n+2)\xi_n$. any $p\ge 0$, the space $\mathcal{S}_p({\mathbb{R}})$ to be the of the closure of $A^p$. Endowed with the norm $|\xi|_p:=|A^p\xi|_0$, it is a space. Note that the norm $|\cdot|_p$ can be expressed as follows, if one the basis $$|\xi|_p^2=\sum_{n=0}^\infty \xi,\xi_n\rangle^2(2n+2)^{2p}.$$ We denote by $\mathcal{S}'_p({\mathbb{R}})$ the dual of $\mathcal{S}_p({\mathbb{R}})$. The norm in $S'_p({\mathbb{R}})$ is given by [@ob], Lemma 1.2.8) $$|\xi|^2_{-p}=\sum_{n=0} ^\infty |\la \xi,A^{-p}\xi_n\ra |^2 \langle \xi,\xi_n\rangle^2(2n+2)^{-2p}$$ for any \mathcal{S}'_p({\mathbb{R}})$. One can show that projective of the $\mathcal{S}_{p}({\mathbb{R}})$, 0$, $\mathcal{S}({\mathbb{R}})$, that the limit of the spaces $ \mathcal{S}_p({\mathbb{R}})'$, $p\ge 0$, is $\mathcal{S}'({\mathbb{R}})$ and that $$\mathcal{S}({\mathbb{R}})\subset L^2({\mathbb{R}})\subset \mathcal{S}'({\mathbb{R}})$$ is a triple. We | n=\sqrt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, There existS a conStaNt $K_2>0$ SuCh thAt $\|\xi'_N\|_\infty\leq K_2 n^{5/12}$.
ConSIder The positive self-adjoint OperaToR $a$ (whoSE iNversE is HilbERt–sCHmiDt) GiVen By $a=-\FrAc{\matHrm{D}^2}{\mathrm{D}x^2}+(1+x^2)$. We have $A\Xi_n=(2N+2)\xI_n$.
For any $p\ge 0$, dEFiNe the space $\MatHcal{S}_p({\mathbb{r}})$ to Be the dOmAin OF the cLosUre of $a^p$. EndoWEd with The norm $|\xi|_P:=|A^P\Xi|_0$, it is A hilbert SPAcE. NotE that the norm $|\cdot|_p$ CAn BE expressed as foLlows, iF oNE uSES thE orThonormal bAsIs $(\xi_n)$: $$|\XI|_p^2=\sum_{n=0}^\iNFtY \LANglE \Xi,\xi_n\rangle^2(2n+2)^{2p}.$$ we denote by $\mAThcAl{S}'_p({\maThBb{R}})$ THe dual Of $\matHcAL{S}_p({\Mathbb{R}})$. The nOrm iN $S'_p({\mathbb{r}})$ is givEN by (see [@oB], lemma 1.2.8) $$|\xi|^2_{-P}=\sum_{n=0} ^\iNftY |\la \Xi,A^{-p}\XI_n\Ra |^2
=\Sum_{N=0}^\iNFty \LAnGle \XI,\xi_N\rangle^2(2n+2)^{-2P}$$ fOr Any $\xi\In \maTHCAL{S}'_p({\mAthBb{R}})$. ONe can Show that the prOjeCtivE LimIt of tHe spaCes $\mAtHcal{S}_{P}({\mathbB{R}})$, $p\geQ 0$, iS $\mathcal{S}({\mathbb{r}})$, thaT the inducTivE lImiT oF the sPAces $ \maThcAl{S}_P({\mathbb{r}})'$, $p\ge 0$, is $\mAThcAl{s}'({\MAThBb{R}})$ and that $$\mathcal{S}({\MaTHBb{r}})\subset L^2({\Mathbb{r}})\SuBsET \mathcal{s}'({\mAthBb{R}})$$ iS A gel’faNd trIPlE.
We | n=\sqrt{\frac{n}2}\xi_{n-1 }-\sqrt{\f rac{n +1} {2} }\ xi_{ n+1} $, there exist s a c onstant $K_2>0$ such t hat $ \| \ xi'_ n \| _\inf ty\leqK _2 n ^{5 /1 2} $.
C o ns iderthe positi ve self-ad joi nt operator $A $ ( whose inve rse is Hilbert– Sch midt)gi ven by $A =-\ frac{ \mathr m {d}^2} {\mathrm{ d} x ^2}+(1 + x^2)$.W e h ave$A\xi_n=(2n+2)\xi _ n$ .
For any $p\g e 0$,de f in e the sp ace $\math ca l{S}_ p ({\math b b{ R } } )$t o be the doma in of the c l osu re of$A ^p$ . Endow ed wi th the norm $|\xi |_p: =|A^p\xi| _0$, i t is a H i lbert s pace.Not e t hatt he n orm $ | \cd o t| _p$ can be expr es se d asfoll o w s , ifone use s the orthonormalbas is $ ( \xi _n)$: $$|\ xi|_ p^ 2=\su m_{n=0 }^\in ft y \langle \xi, \xi_ n\rangle^ 2(2 n+ 2)^ {2 p}.$$ We den ote by $\math cal{S}' _ p({ \m a t h bb {R}})$ the dual of $ \ m at hcal{S}_ p({\ma t hb b{ R }})$. Th enor m in $ S'_p( {\ma t hb b{R}})$is giv e nby (see [ @o b], Le mm a 1 .2. 8) $$ | \xi| ^2_{-p }=\sum_{ n=0}^ \infty |\la \x i ,A^{-p}\xi_n\ r a| ^ 2= \sum _{n =0}^\infty \la n gle\xi, \ xi _n\ r angle ^2(2n +2 ) ^{ - 2p}$$ for any $\xi\ in \math cal{S }'_p({\mathbb {R}})$. On e c an showthat th e projective li mit o f the spac e s $\math cal{S }_{p}({\ mathbb{R} } ) $, $p\ge q 0 $,is$\m a t hc al{S}({\mathb b { R}}) $, that t heinducti velim itofth e spaces$ \mathc al {S }_ p( {\m athbb { R}})'$,$p \ge 0 $,is $\ m athcal {S}'( {\ma th bb { R}} )$ andt ha t $$\m at hc al{S }({ \m athbb {R}} ) \su bset L^ 2({\mathb b{R } })\s ub se t \math cal{S}'({\mat hb b{R}})$$ i sa G el’fan d triple.
We | n=\sqrt{\frac{n}2}\xi_{n-1}-\sqrt{\frac{n+1}{2}}\xi_{n+1}$, there_exists a_constant $K_2>0$ such that_$\|\xi'_n\|_\infty\leq K_2_n^{5/12}$.
Consider_the positive_self-adjoint_operator $A$ (whose_inverse is Hilbert–Schmidt)_given by $A=-\frac{\mathrm{d}^2}{\mathrm{d}x^2}+(1+x^2)$. We_have $A\xi_n=(2n+2)\xi_n$.
For any_$p\ge_0$, define the space $\mathcal{S}_p({\mathbb{R}})$ to be the domain of the closure of $A^p$._Endowed_with the_norm_$|\xi|_p:=|A^p\xi|_0$,_it is a Hilbert space._Note that the norm $|\cdot|_p$_can be_expressed as follows, if one uses the orthonormal_basis_$(\xi_n)$: $$|\xi|_p^2=\sum_{n=0}^\infty _\langle \xi,\xi_n\rangle^2(2n+2)^{2p}.$$ We denote by $\mathcal{S}'_p({\mathbb{R}})$ the dual of_$\mathcal{S}_p({\mathbb{R}})$. The norm in $S'_p({\mathbb{R}})$ is_given by (see_[@ob],_Lemma_1.2.8) $$|\xi|^2_{-p}=\sum_{n=0} ^\infty |\la_\xi,A^{-p}\xi_n\ra |^2
=\sum_{n=0}^\infty \langle \xi,\xi_n\rangle^2(2n+2)^{-2p}$$ for_any $\xi\in \mathcal{S}'_p({\mathbb{R}})$. One can show_that the projective limit of the spaces_$\mathcal{S}_{p}({\mathbb{R}})$, $p\geq 0$, is $\mathcal{S}({\mathbb{R}})$, that_the inductive limit of the_spaces $_\mathcal{S}_p({\mathbb{R}})'$, $p\ge 0$, is $\mathcal{S}'({\mathbb{R}})$_and that $$\mathcal{S}({\mathbb{R}})\subset_L^2({\mathbb{R}})\subset \mathcal{S}'({\mathbb{R}})$$_is a Gel’fand_triple.
We |
}],
\quad
h(x) = \hat a x +\hat b.$$
We immediately get $$\hat a = -\hat b, \quad \hat a = \frac1{2b}$$ and $$\label{eq:alphabetacayley3}
\alpha = 2t+1,$$ $$\label{eq:alphabetacayley4}
\gamma=\frac{2b}{\beta -1}
$$ where $t=b/a$.
Now we get the following
\[propZolot\] The polynomial $\hat p_3$ can be expressed through the Zolotarev polynomial $z_3$ up to a nonessential constant factor: $$\hat p_3 (s) \sim z_3( 2bs+1).$$
To verify the proposition, we should certify that the definition of $\alpha$ and $\beta$ from [eq:alphabetan]{} for $n=3$ and the relations [eq:alphabetacayley3]{}, [eq:alphabetacayley4]{} are compatible with the formula for $\gamma$ we got before as Cayley condition, see [CaleyEq1]{}.
In order to do that we will use well-known identities for the Jacobi elliptic functions:
$$\begin{gathered}
\sn^2u+\cn^2u=1,
\\
\kappa^2\sn^2u+\dn^2u=1,
\\
\sn(u+v)=\frac{\sn\, u\cn\, v\dn \,v + \sn\, v\cn\, u\dn\, u}{1-\kappa^2\sn^2 u \sn ^2 v},
\\
\sn (K-u)=\frac{\cn\, u}{\dn\, u}.\end{gathered}$$
In particular, we get $$\begin{gathered}
\sn \left(\frac {2K}{3}\right)= \frac{2\sn\, \frac{K}{3}\cn \,\frac{K}{3}\dn \,\frac{K}{3}}{1-\kappa^2\sn^4\frac{K}{3}},
\\
\sn \left(\frac {2}{3}K\right)= \sn \left(K-\frac {K}{3}\right)=\frac{\cn \,\frac{K}{3}}{\dn\, \frac{K | } ],
\quad
h(x) = \hat a x + \hat b.$$
We immediately get $ $ \hat a = -\hat b, \quad \hat a = \frac1{2b}$$ and $ $ \label{eq: alphabetacayley3 }
\alpha = 2t+1,$$ $ $ \label{eq: alphabetacayley4 }
\gamma=\frac{2b}{\beta -1 }
$ $ where $ t = b / a$.
Now we draw the take after
\[propZolot\ ] The polynomial $ \hat p_3 $ can be expressed through the Zolotarev polynomial $ z_3 $ up to a nonessential constant divisor: $ $ \hat p_3 (s) \sim z_3 (2bs+1).$$
To verify the proposition, we should license that the definition of $ \alpha$ and $ \beta$ from [ eq: alphabetan ] { } for $ n=3 $ and the relation [ eq: alphabetacayley3 ] { }, [ eq: alphabetacayley4 ] { } are compatible with the formula for $ \gamma$ we get before as Cayley condition, experience [ CaleyEq1 ] { }.
In order to do that we will use well - know identities for the Jacobi elliptic functions:
$ $ \begin{gathered }
\sn^2u+\cn^2u=1,
\\
\kappa^2\sn^2u+\dn^2u=1,
\\
\sn(u+v)=\frac{\sn\, u\cn\, v\dn \,v + \sn\, v\cn\, u\dn\, u}{1-\kappa^2\sn^2 u \sn ^2 v },
\\
\sn (K - u)=\frac{\cn\, u}{\dn\, u}.\end{gathered}$$
In particular, we get $ $ \begin{gathered }
\sn \left(\frac { 2K}{3}\right)= \frac{2\sn\, \frac{K}{3}\cn \,\frac{K}{3}\dn \,\frac{K}{3}}{1-\kappa^2\sn^4\frac{K}{3 } },
\\
\sn \left(\frac { 2}{3}K\right)= \sn \left(K-\frac { K}{3}\right)=\frac{\cn \,\frac{K}{3}}{\dn\, \frac{K | }],
\quaf
h(x) = \hat a x +\hat b.$$
We imoediately get $$\hcr a = -\het b, \quzd \hat a = \frac1{2b}$$ and $$\label{eq:alphabetaraylwy3}
\alpya = 2t+1,$$ $$\label{eq:alphabetazayley4}
\gamla=\frac{2b}{\bwta -1}
$$ qhere $t=b/a$.
Now we geb the nollociig
\[propZolot\] The polynomian $\hat p_3$ can be ebpxessed through the Zolotarev polynomyal $z_3$ ul ho a nonessentyal bogstahn nactor: $$\hat p_3 (s) \sim z_3( 2bs+1).$$
To verify the prmposition, we xhould certify that the devinihion of $\alpha$ and $\heta$ from [ew:alprqbetan]{} for $n=3$ and the rtlctions [eq:alkhabetacayley3]{}, [eq:alphabetacayley4]{} xre cpmpatible qirh hve formula hor $\gaima$ we got bcgore av Cayleu condition, sec [CalxyEq1]{}.
Un order to do that wx will use well-known identitias for the Jacobi eoluptic funwtiovw:
$$\beein{favheded}
\sn^2u+\fn^2u=1,
\\
\iappa^2\sn^2u+\dn^2h=1,
\\
\sn(u+v)=\frac{\sn\, u\cn\, v\dn \,v + \sn\, v\cn\, i\dg\, u}{1-\kappa^2\sn^2 u \sh ^2 v},
\\
\sn (K-t)=\frac{\cn\, u}{\dn\, u}.\end{gathered}$$
In particular, wt get $$\begin{gathered}
\sn \left(\frqc {2K}{3}\right)= \frac{2\sn\, \frac{N}{3}\cn \,\frac{K}{3}\qn \,\frac{K}{3}}{1-\kappa^2\sn^4\frac{K}{3}},
\\
\sn \left(\frac {2}{3}K\right)= \sn \left(K-\frdc {K}{3}\rmgft)=\fxqc{\cn \,\weaf{K}{3}}{\dn\, \frac{K | }], \quad h(x) = \hat a x We get $$\hat = -\hat b, and \alpha = 2t+1,$$ \gamma=\frac{2b}{\beta -1} $$ $t=b/a$. Now we get the following The polynomial $\hat p_3$ can be expressed through the Zolotarev polynomial $z_3$ up a nonessential constant factor: $$\hat p_3 (s) \sim z_3( 2bs+1).$$ To verify the we certify the of $\alpha$ and $\beta$ from [eq:alphabetan]{} for $n=3$ and the relations [eq:alphabetacayley3]{}, [eq:alphabetacayley4]{} are compatible with formula for $\gamma$ we got before as Cayley see [CaleyEq1]{}. In order do that we will use identities the Jacobi functions: \sn^2u+\cn^2u=1, \kappa^2\sn^2u+\dn^2u=1, \\ \sn(u+v)=\frac{\sn\, v\dn \,v + \sn\, v\cn\, u\dn\, u}{1-\kappa^2\sn^2 u \sn ^2 v}, \\ \sn (K-u)=\frac{\cn\, u}{\dn\, u}.\end{gathered}$$ In we get \left(\frac {2K}{3}\right)= \frac{K}{3}\cn \,\frac{K}{3}}{1-\kappa^2\sn^4\frac{K}{3}}, \sn \left(\frac {2}{3}K\right)= {K}{3}\right)=\frac{\cn \,\frac{K}{3}}{\dn\, \frac{K | }],
\quad
h(x) = \hat a x +\hat b.$$
We immediatEly get $$\hat a = -\Hat b, \qUad \Hat A = \fRac1{2b}$$ And $$\lAbel{eq:alphabetACaylEy3}
\alpha = 2t+1,$$ $$\label{eq:alphabeTacayLeY4}
\GammA=\FrAc{2b}{\beTa -1}
$$ where $T=B/a$.
nOW we GeT tHe fOlLOwIng
\[prOpZOlot\] The Polynomial $\Hat P_3$ cAn be expresseD ThRough the ZoLotArev polynomiAl $z_3$ Up to a nOnEssENtial ConStant Factor: $$\HAt p_3 (s) \siM z_3( 2bs+1).$$
To verIfY The proPOsition, WE ShOuld Certify that the defINiTIon of $\alpha$ and $\bEta$ froM [eQ:AlPHAbeTan]{} For $n=3$ and the ReLatioNS [eq:alphABeTACAylEY3]{}, [eq:alphabetacAyley4]{} are comPAtiBle witH tHe fORmula fOr $\gamMa$ WE goT before as CaYley Condition, See [CalEYEq1]{}.
In orDEr to do tHat we wIll Use Well-KNoWn IdeNtITieS FoR thE jacObi ellipTiC fUnctiOns:
$$\bEGIN{GathEreD}
\sn^2u+\Cn^2u=1,
\\
\kaPpa^2\sn^2u+\dn^2u=1,
\\
\sn(u+v)=\FraC{\sn\, u\CN\, v\dN \,v + \sn\, v\Cn\, u\dn\, U}{1-\kapPa^2\Sn^2 u \sn ^2 V},
\\
\sn (K-u)=\fRac{\cn\, U}{\dN\, u}.\end{gathered}$$
In PartIcular, we gEt $$\bEgIn{gAtHered}
\SN \left(\fRac {2k}{3}\riGht)= \frac{2\Sn\, \frac{K}{3}\CN \,\frAc{k}{3}\DN \,\FrAc{K}{3}}{1-\kappa^2\sn^4\frac{K}{3}},
\\
\sn \lEfT(\FRaC {2}{3}K\right)= \sN \left(K-\FRaC {K}{3}\RIght)=\frac{\Cn \,\FraC{K}{3}}{\dn\, \FRAc{K | }],
\quad
h(x) = \hat a x +\hat b.$ $
We im med ia tely get $$\hat a = -\ h at b , \quad \hat a = \frac 1{2b} $$ and$ $\ label {eq:alp h ab e t aca yl ey 3}\a l ph a = 2 t+1 ,$$ $$\ label{eq:a lph ab etacayley4}\ ga mma=\frac{ 2b} {\beta -1}
$ $ w here $ t= b/a $ .
No w w e get the f o llowin g
\[prop Zo l ot\] T h e polyn o m ia l $\ hat p_3$ can be e x pr e ssed through t he Zol ot a re v pol yno mial $z_3$ u p toa noness e nt i a l co n stant factor: $$\hat p_3 (s) \simz_ 3(2 bs+1). $$
T ov eri fy the prop osit ion, we s houldc ertifyt hat the defin iti onof $ \ al ph a$an d $\ b et a$f rom [eq:alp ha be tan]{ } fo r $ n =3$and the rela tions [eq:alp hab etac a yle y3]{} , [eq :alp ha betac ayley4 ]{} a re compatible wit h th e formula fo r$\g am ma$ w e got b efo reas Cayl ey cond i tio n, s e e[CaleyEq1]{}.
Inor d e rto do th at wew il lu se well- kn own ide n t ities for th e Jacobi ellip t ic f unction s:
$$\b eg in{ gat hered }
\sn ^2u+\c n^2u=1,\\
\k a ppa^2\sn^2u+\d n ^2u=1,
\\
\sn ( u+ v ) =\ f rac{ \sn \, u\cn\, v \dn\ ,v + \sn \ ,v\c n \, u\ dn\,u} { 1- \ kappa^2\sn^2 u \sn^2 v},
\ \
\sn (K-u)=\frac{ \cn\, u}{\ d n \ , u}.\en d{ga t he r ed}$$
In part icula r, we get$ $\begin{ gathe red}
\sn \left(\f r a c {2K}{3 }\r igh t)= \f r a c{ 2\sn\, \frac{ K } {3}\ cn \,\fra c{K }{3}\dn \, \fr ac{ K}{ 3} }{1-\kapp a^2\sn^4 \f ra c{ K} {3} },
\\ \sn \lef t( \fr ac {2 }{3}K \ right) = \sn \le ft (K - \fr ac {K}{ 3 }\ r i ght) =\ fr ac{\ cn\, \frac {K}{ 3 }}{ \dn\, \ frac{K | }],
\quad
h(x)_= \hat_a x +\hat b.$$
We_immediately get_$$\hat_a =_-\hat_b, \quad \hat_a = \frac1{2b}$$_and $$\label{eq:alphabetacayley3}
\alpha = 2t+1,$$_$$\label{eq:alphabetacayley4}
\gamma=\frac{2b}{\beta -1}
$$ where_$t=b/a$.
Now_we get the following
\[propZolot\] The polynomial $\hat p_3$ can be expressed through the Zolotarev_polynomial_$z_3$ up_to_a_nonessential constant factor: $$\hat p_3_(s) \sim z_3( 2bs+1).$$
To verify_the proposition,_we should certify that the definition of $\alpha$_and_$\beta$ from [eq:alphabetan]{}_for $n=3$ and the relations [eq:alphabetacayley3]{}, [eq:alphabetacayley4]{} are compatible_with the formula for $\gamma$ we_got before as_Cayley_condition,_see [CaleyEq1]{}.
In order to_do that we will use well-known_identities for the Jacobi elliptic functions:
$$\begin{gathered}
\sn^2u+\cn^2u=1,
\\
\kappa^2\sn^2u+\dn^2u=1,
\\
\sn(u+v)=\frac{\sn\,_u\cn\, v\dn \,v + \sn\, v\cn\, u\dn\,_u}{1-\kappa^2\sn^2 u \sn ^2 v},
\\
\sn (K-u)=\frac{\cn\,_u}{\dn\, u}.\end{gathered}$$
In particular, we get_$$\begin{gathered}
\sn \left(\frac_{2K}{3}\right)= \frac{2\sn\, \frac{K}{3}\cn \,\frac{K}{3}\dn \,\frac{K}{3}}{1-\kappa^2\sn^4\frac{K}{3}},
\\
\sn_\left(\frac {2}{3}K\right)= \sn_\left(K-\frac {K}{3}\right)=\frac{\cn_\,\frac{K}{3}}{\dn\, \frac{K |
hof] The Fermi surface of holon is a small one with Fermi wave-vector $k_F\approx\pi\delta$.[@marchetti2004b] Due to the staggered $\pi$-flux, we divide the square lattice into two sublattices, A(even sites) and B(odd sites). On each sublattice, the holon’s annihilation operators are denoted by $\hat{a}$ and $\hat{b}$, respectively. The Hamiltonian Eq. of free holon can then be recast in a quadratic form $$\begin{aligned}
\label{eq:13} \hat{H}_{0}^h \sim \sum_{\vec{k}} (t_{\vec{k}}
\hat{a}^{\dagger}_{\vec{k}} \hat{b}_{\vec{k}} + h.c.) -
\mu\sum_{\vec{k}} (\hat{a}_{\vec{k}}^{\dagger} \hat{a}_{\vec{k}} +
\hat{b}_{\vec{k}}^{\dagger} \hat{b}_{\vec{k}} ),\end{aligned}$$ where the momentum runs within the magnetic Brillouin zone (MBZ) and $t_{\vec{k}} =2t(\cos k_x e^{i\pi/4} + \cos k_y e^{-i\pi/4})$. It is straightforward to obtain the spectrum $\epsilon(\vec{k}) =
\pm |t_{\vec{k}}|$, with the Fermi surface consisting of four half circles around $(\pm\pi/2,\pm\pi/2)$, as shown in Fig. \[fig:1\]a, where the blue lines are the boundary of MBZ. The Fermi energy is $t\delta$ approximately. There are two primitive reciprocal vectors, $\vec{\pi}_{\pm}\equiv (\pm\pi,\pi)$ by which we can translate the MBZ in the 3rd and 4th quadrants to get another equivalent rectangular one as shown in Fig. \[fig:1\]b, which consists of two Dirac cones centered around $\vec{Q}_L=(-\pi/2,\pi/2)$ (left) and $\vec{Q}_R=
(\pi/2,\pi/2)$ (right), respectively. In this transformation, we note that $ \hat{a}_{\vec{k}+\vec{\pi}_{\pm}} = \hat{a}_{\vec{k}}$ and | hof ] The Fermi surface of holon is a small one with Fermi wave - vector $ k_F\approx\pi\delta$.[@marchetti2004b ] Due to the staggered $ \pi$-flux, we divide the square wicket into two sublattices, A(even web site) and B(odd sites). On each sublattice, the holon ’s annihilation operators are announce by $ \hat{a}$ and $ \hat{b}$, respectively. The Hamiltonian Eq. of detached holon can then be recast in a quadratic form $ $ \begin{aligned }
\label{eq:13 } \hat{H}_{0}^h \sim \sum_{\vec{k } } (t_{\vec{k } }
\hat{a}^{\dagger}_{\vec{k } } \hat{b}_{\vec{k } } + h.c .) -
\mu\sum_{\vec{k } } (\hat{a}_{\vec{k}}^{\dagger } \hat{a}_{\vec{k } } +
\hat{b}_{\vec{k}}^{\dagger } \hat{b}_{\vec{k } }), \end{aligned}$$ where the momentum runs within the charismatic Brillouin zone (MBZ) and $ t_{\vec{k } } = 2t(\cos k_x e^{i\pi/4 } + \cos k_y e^{-i\pi/4})$. It is straightforward to obtain the spectrum $ \epsilon(\vec{k }) =
\pm |t_{\vec{k}}|$, with the Fermi surface dwell of four half circles around $ (\pm\pi/2,\pm\pi/2)$, as shown in Fig. \[fig:1\]a, where the aristocratic lines are the boundary of MBZ. The Fermi department of energy is $ t\delta$ approximately. There are two primitive reciprocal vector, $ \vec{\pi}_{\pm}\equiv (\pm\pi,\pi)$ by which we can translate the MBZ in the 3rd and fourth quadrants to get another equivalent rectangular one as prove in Fig. \[fig:1\]b, which consists of two Dirac cones focus on around $ \vec{Q}_L=(-\pi/2,\pi/2)$ (left) and $ \vec{Q}_R=
(\pi/2,\pi/2)$ (right field), respectively. In this transformation, we note that $ \hat{a}_{\vec{k}+\vec{\pi}_{\pm } } = \hat{a}_{\vec{k}}$ and | hof] The Fermi surface of houon is a small one witi Fermi wave-vecgor $k_F\approx\pi\delta$.[@marchetti2004u] Duw to uke staggered $\pi$-flux, de divide the squqre oqttice into two sunjattjges, A(zvxn sites) and B(ocd sites). Ot each sublatthcd, che holon’s annihilation operators arq denotrd by $\hat{a}$ and $\hwt{b}$, gefpecfpvtly. The Hamiltonian Eq. of free homon can then be recaxt in a quadratic form $$\begln{allgned}
\label{eq:13} \hat{H}_{0}^h \sim \sum_{\vec{j}} (t_{\vqx{k}}
\hat{a}^{\dagger}_{\xec{k}} \hat{b}_{\vtc{n}} + h.c.) -
\mu\suj_{\vec{k}} (\hat{a}_{\vec{k}}^{\dagger} \hat{a}_{\vec{k}} +
\hxt{b}_{\vee{k}}^{\dagger} \haj{y}_{\cec{n}} ),\end{aligned}$$ wherv the momentum runs whthin tne magnetic Brllloumn zine (MBZ) and $t_{\vec{k}} =2t(\cov k_x e^{i\pi/4} + \cos k_y g^{-i\pi/4})$. It is scraightforward to obtqib the spewtruo $\eprilkn(\tec{i}) =
\pm |t_{\gec{i}}|$, with the Fermi surfqce consisting of fpuw half circles arounq $(\[m\pi/2,\pm\pi/2)$, as shown in Fig. \[fig:1\]a, where the tlus lines are the boundart of MBZ. The Fermi engrgy is $t\dqlta$ approximately. There are two primitive reciprmcal teztoxw, $\vec{\ou}_{\pl}\equiv (\pm\pi,\pi)$ by which we can translate the MFA on the 3rd and 4th quadrants yo grj another equixalent redtangular one as sjown in Fig. \[fug:1\]b, which conxists of two Dirac cones cebtered arounb $\vwc{Q}_L=(-\pi/2,\pi/2)$ (left) and $\rec{Q}_R=
(\pi/2,\pi/2)$ (riyht), rexpectovely. In this transformction, se note thah $ \hat{a}_{\ved{y}+\vec{\pi}_{\pm}} = \hat{a}_{\vez{k}}$ sng | hof] The Fermi surface of holon is one Fermi wave-vector Due to the square into two sublattices, sites) and B(odd On each sublattice, the holon’s annihilation are denoted by $\hat{a}$ and $\hat{b}$, respectively. The Hamiltonian Eq. of free holon then be recast in a quadratic form $$\begin{aligned} \label{eq:13} \hat{H}_{0}^h \sim \sum_{\vec{k}} (t_{\vec{k}} \hat{b}_{\vec{k}} h.c.) \mu\sum_{\vec{k}} \hat{a}_{\vec{k}} + \hat{b}_{\vec{k}}^{\dagger} \hat{b}_{\vec{k}} ),\end{aligned}$$ where the momentum runs within the magnetic Brillouin zone (MBZ) and =2t(\cos k_x e^{i\pi/4} + \cos k_y e^{-i\pi/4})$. It straightforward to obtain the $\epsilon(\vec{k}) = \pm |t_{\vec{k}}|$, with Fermi consisting of half around as shown in \[fig:1\]a, where the blue lines are the boundary of MBZ. The Fermi energy is $t\delta$ approximately. There two primitive $\vec{\pi}_{\pm}\equiv (\pm\pi,\pi)$ which can the MBZ in and 4th quadrants to get another as shown in Fig. \[fig:1\]b, which consists of Dirac cones around $\vec{Q}_L=(-\pi/2,\pi/2)$ (left) and $\vec{Q}_R= (\pi/2,\pi/2)$ respectively. In this transformation, we note that $ = \hat{a}_{\vec{k}}$ and | hof] The Fermi surface of holon Is a small onE with ferMi wAvE-vecTor $k_f\approx\pi\delta$.[@MArchEtti2004b] Due to the staggered $\Pi$-fluX, wE DiviDE tHe squAre lattICe INTo tWo SuBlaTtICeS, A(eveN siTes) and B(Odd sites). On EacH sUblattice, the HOlOn’s annihilAtiOn operators aRe dEnoted By $\Hat{A}$ And $\haT{b}$, rEspecTively. tHe HamiLtonian Eq. Of FRee holON can theN BE rEcasT in a quadratic form $$\BEgIN{aligned}
\label{eQ:13} \hat{H}_{0}^h \SiM \SuM_{\VEc{k}} (T_{\veC{k}}
\hat{a}^{\daggEr}_{\Vec{k}} \hAT{b}_{\vec{k}} + h.C.) -
\Mu\SUM_{\Vec{K}} (\Hat{a}_{\vec{k}}^{\daggeR} \hat{a}_{\vec{k}} +
\haT{B}_{\veC{k}}^{\daggEr} \Hat{B}_{\Vec{k}} ),\enD{aligNeD}$$ WheRe the momentUm ruNs within tHe magnETic BrilLOuin zonE (MBZ) anD $t_{\vEc{k}} =2T(\cos K_X e^{I\pI/4} + \coS k_Y E^{-i\pI/4})$. it Is sTRaiGhtforwaRd To ObtaiN the SPECTrum $\EpsIlon(\Vec{k}) =
\pM |t_{\vec{k}}|$, with the ferMi suRFacE consIstinG of fOuR half CircleS arouNd $(\Pm\pi/2,\pm\pi/2)$, as shown In FiG. \[fig:1\]a, wherE thE bLue LiNes arE The bouNdaRy oF MBZ. The fermi enERgy Is $T\DELtA$ approximately. TherE aRE TwO primitiVe reciPRoCaL Vectors, $\vEc{\Pi}_{\pM}\equIV (\Pm\pi,\pI)$ by wHIcH we can trAnslatE ThE MbZ in the 3Rd And 4th qUaDraNts To get ANothEr equiValent reCtangULar one as shown iN fig. \[fig:1\]b, which cONsISTs OF two dirAc cones centEred ARounD $\vec{q}_l=(-\pI/2,\pi/2)$ (LEft) anD $\vec{Q}_r=
(\pI/2,\Pi/2)$ (RIght), respectively. In tHiS transFormaTion, we note thaT $ \hat{a}_{\vec{k}+\vEC{\PI}_{\pm}} = \hat{a}_{\vEc{k}}$ aND | hof] The Fermi surface ofholon is a smal l o newi th F ermi wave-vector $ k _F\a pprox\pi\delta$.[@marc hetti 20 0 4b]D ue to t he stag g er e d $\ pi $- flu x, we divi dethe squ are lattic e i nt o two sublat t ic es, A(even si tes) and B(o ddsites) .One ach s ubl attic e, the holon’ s annihil at i on ope r ators a r e d enot ed by $\hat{a}$ a n d$ \hat{b}$, resp ective ly . T h e Ha mil tonian Eq. of fr e e holon ca n t hen be recast ina quadratic for m $$\b eg in{ a ligned }
\la be l {eq :13} \hat{H }_{0 }^h \sim\sum_{ \ vec{k}} (t_{\ve c{k}}\ha t{a }^{\ d ag ge r}_ {\ v ec{ k }} \h a t{b }_{\vec{ k} }+ h.c .) - \ m u \sum _{\ vec{ k}} ( \hat{a}_{\vec {k} }^{\ d agg er} \ hat{a }_{\ ve c{k}} +
\ha t{b}_ {\ vec{k}}^{\dagge r} \ hat{b}_{\ vec {k }}), \end{ a ligned }$$ wh ere the moment u m r un s w it hin the magnetic B ri l l ou in zone(MBZ)a nd $ t _{\vec{k }} =2 t(\c o s k_xe^{i \ pi /4} + \c os k_y e^ {- i\pi/4} )$ . It i sstr aig htfor w ardto obt ain thespect r um $\epsilon(\ v ec{k}) =
\pm| t_ { \ ve c {k}} |$, with the F ermi surf acec on sis t ing o f fou rh al f circles around $(\ pm \pi/2, \pm\p i/2)$, as sho wn in Fig. \ [ fig:1\]a , wh e re the blue lines arethe bounda r y of MBZ . The Fermi e nergy is$ t \delta$app rox ima tel y . T here are twop r imit iv e recip roc al vect ors , $ \ve c{\ pi }_{\pm}\e quiv (\p m\ pi ,\ pi )$by wh i ch we ca ntra ns lat e the MBZ in the3rdan d4 thquadran t st o get a no ther eq ui valen t re c tan gular o ne as sho wni n Fi g. \ [fig:1\ ]b, which con si sts of two D ira c cone s centered around $\vec{Q}_L=(-\p i /2,\pi/ 2)$ (lef t) a nd $\vec{ Q}_ R=
(\p i/2 , \pi/2) $ (rig ht),re spe c t ively . In th is transform a t ion , weno te t hat $ \ hat{a}_{\vec{k}+\v e c{\ pi}_{\pm}} =\ha t{a} _ { \v ec{ k }} $ an d | hof] The_Fermi surface_of holon is a_small one_with_Fermi wave-vector_$k_F\approx\pi\delta$.[@marchetti2004b]_Due to the_staggered $\pi$-flux, we_divide the square lattice_into two sublattices,_A(even_sites) and B(odd sites). On each sublattice, the holon’s annihilation operators are denoted by_$\hat{a}$_and $\hat{b}$,_respectively._The_Hamiltonian Eq. of free holon_can then be recast in_a quadratic_form $$\begin{aligned}
\label{eq:13} \hat{H}_{0}^h \sim \sum_{\vec{k}} (t_{\vec{k}}
\hat{a}^{\dagger}_{\vec{k}} \hat{b}_{\vec{k}} +_h.c.)_-
\mu\sum_{\vec{k}} (\hat{a}_{\vec{k}}^{\dagger} \hat{a}_{\vec{k}}_+
\hat{b}_{\vec{k}}^{\dagger} \hat{b}_{\vec{k}} ),\end{aligned}$$ where the momentum runs within the_magnetic Brillouin zone (MBZ) and $t_{\vec{k}}_=2t(\cos k_x e^{i\pi/4}_+_\cos_k_y e^{-i\pi/4})$. It is_straightforward to obtain the spectrum $\epsilon(\vec{k})_=
\pm |t_{\vec{k}}|$, with the Fermi surface_consisting of four half circles around $(\pm\pi/2,\pm\pi/2)$,_as shown in Fig. \[fig:1\]a, where the_blue lines are the boundary_of MBZ._The Fermi energy is $t\delta$_approximately. There are_two primitive_reciprocal vectors, $\vec{\pi}_{\pm}\equiv_(\pm\pi,\pi)$ by which we can translate_the MBZ in_the 3rd and 4th quadrants to_get_another equivalent rectangular_one_as_shown in_Fig. \[fig:1\]b, which consists_of_two Dirac_cones_centered around $\vec{Q}_L=(-\pi/2,\pi/2)$ (left) and $\vec{Q}_R=
(\pi/2,\pi/2)$_(right),_respectively. In this transformation, we note that_$ \hat{a}_{\vec{k}+\vec{\pi}_{\pm}} = \hat{a}_{\vec{k}}$_and |
momentum and kinetic-energy budgets. Further, the drag force for the perpendicular motion shows quantitative agreement with the infinite series solution of [@bar1968] with an error of less than $0.043\%$. The drag force for the parallel motion approaches the wide-gap solution of [@mag2003] with increasing $\varepsilon$. To make sure of numerical stability and accuracy, we set the clearance parameter in a range of $10^{-3}\leq \varepsilon\leq 9$. We performed the convergence tests with varying the size $R_{\rm max}$ of the computational domain, the number $N_\xi$ of nodes in the gap between the bubble and the wall, and the number $N_\eta$ of nodes describing the bubble surface. We confirmed that the relative errors in the migration velocity $V_{B3}$ and the drag forces for the perpendicular and parallel bubble motions to those obtained on the base meshes $N_\xi\times N_\eta = 200 \times 200$ for various $\varepsilon$ decrease with increasing the size $R_{\rm max}$ and the resolutions $N_\xi$ and $N_\eta$. From the convergence behaviors, we deduce to obtain the migration velocities $V_{B3}$ with errors of much less than $1\%$ using the present base meshes, which are accurate enough for the subsequent discussion.
Migration velocity
------------------
As shown in the inset of figure \[fig:migvel\_comp\_exp\], the ratio of the simulation to the analytical migration velocity becomes close to unity as $\kappa (=(1+\varepsilon)^{-1})$ approaches zero, indicating the simulation result is consistent with the analytical solution ([@mag2003]) for small $\kappa (< \sim 0.3)$. Such a consistency between the different approaches may deny the erroneous prediction of the migration velocity in [@mag2003], of which the possibility was pointed out by [@wan2006]. By contrast, the simulated migration velocity for the bubble closer to the wall with the clearance shorter than the bubble radius (i.e., for $\kappa > \sim 0.5$ presumably beyond the theoretical precondition $\kappa\ll 1$) is considerably higher than the analytical solution. Although the simulation result reveals lower velocity than the experimental one, the tendency of the higher velocity than the theoretical one for the narrow gap is qualitatively similar to that in the experiment. Thus, the present simulation also indicates the presence of the additional narrow-gap repulsive force.
For an undeformed spherical bubble at small but non-zero Reynolds numbers $0 | momentum and kinetic - energy budgets. Further, the puff power for the perpendicular motion testify quantitative agreement with the countless series solution of [ @bar1968 ] with an error of less than $ 0.043\%$. The drag force out for the parallel motion approaches the across-the-board - gap solution of [ @mag2003 ] with increasing $ \varepsilon$. To take sure of numeral constancy and accuracy, we set the clearance argument in a range of $ 10^{-3}\leq \varepsilon\leq 9$. We perform the overlap tests with varying the size $ R_{\rm max}$ of the computational domain, the phone number $ N_\xi$ of nodes in the gap between the house of cards and the wall, and the phone number $ N_\eta$ of nodes identify the bubble surface. We confirmed that the proportional errors in the migration velocity $ V_{B3}$ and the drag forces for the perpendicular and parallel bubble apparent motion to those obtained on the base meshes $ N_\xi\times N_\eta = 200 \times 200 $ for various $ \varepsilon$ decrease with increasing the size $ R_{\rm max}$ and the resolutions $ N_\xi$ and $ N_\eta$. From the convergence behaviors, we deduce to obtain the migration velocities $ V_{B3}$ with mistake of much less than $ 1\%$ using the present infrastructure mesh topology, which are accurate enough for the subsequent discussion.
Migration velocity
------------------
As shown in the inset of digit \[fig: migvel\_comp\_exp\ ], the proportion of the simulation to the analytical migration velocity become close to unity as $ \kappa (= (1+\varepsilon)^{-1})$ approaches zero, indicating the simulation result is consistent with the analytical solution ([ @mag2003 ]) for small $ \kappa (< \sim 0.3)$. Such a consistency between the different approaches may traverse the erroneous prediction of the migration velocity in [ @mag2003 ], of which the hypothesis was pointed out by [ @wan2006 ]. By contrast, the simulated migration velocity for the bubble closer to the rampart with the clearance shorter than the bubble radius (i.e., for $ \kappa > \sim 0.5 $ presumably beyond the theoretical precondition $ \kappa\ll 1 $) is considerably higher than the analytical solution. Although the simulation solution reveals lower speed than the experimental one, the tendency of the higher velocity than the theoretical one for the narrow opening is qualitatively similar to that in the experiment. therefore, the present simulation also indicates the bearing of the additional narrow - gap repulsive force.
For an undeformed spherical bubble at small but non - zero Reynolds numbers $ 0 | molentum and kinetic-energy budgets. Furthet, rhe dreg forcs for thd perpendicular motion shows qyantiuctive agreement with the infijite serues wilution of [@bar1968] witm an svror mh less than $0.043\%$. Thg drag force for the paralneu lotion approaches the wide-gap solutyon of [@kah2003] with increasyng $\nawepsjlon$. To make sure of numerical stagility end accuracy, we set the clearance parametfr ij a range of $10^{-3}\leq \vwrepsilon\lew 9$. Wq performed tfe convergtnee tests wijh varying the size $R_{\rm max}$ of tfe cokputationao eomwhn, the numbxr $N_\xi$ of nodes in the gap betweem the bubble akd thx waol, and the number $N_\ete$ of nodes describind the bubtlz surface. We confirmee rhat jhe ralatkce drrkrx jn the mijration velkcity $V_{B3}$ ane the drag forces fpr nne perpendichlar agd parallel bubble motions to those obtaited on the base meshes $N_\xi\rimes N_\eta = 200 \times 200$ flr variouf $\varepsilon$ decrease with increasing the size $R_{\rk max}$ xnd tme rdwoputions $N_\xi$ and $N_\eta$. From the convergence behwbiprx, we deduce to obtain the mlgtwtion velocitkes $V_{B3}$ wifh errors of much pess thwn $1\%$ uwing the kresemt base meshes, which are acxurate enougk fir the subsequent biscussion.
Miyratiom velpcity
------------------
As shown in the inret kf figure \[flg:migvel\_ckop\_exp\], the ratio uf nhe vimulation to the analyticwl migratmon vzlocity cecokes cljse to unihy as $\kappa (=(1+\varepsilon)^{-1})$ aoproaehes sero, indicwting the simulation result is rpnsistent wijh dhe analytieal sokution ([@mag2003]) fow small $\kappa (< \sim 0.3)$. Sbch a zonsistencj between the differqnt approachev may deny thx erroneots peedixtion ow the migration velocity in [@mag2003], of which the possibikiti sas pointed out by [@wan2006]. By contrasy, tfe filukaead migration velucigu for the bubble zlosrr to the wall with dhe dlearance shorter yhwn the bobble radyus (i.e., for $\kalpa > \sim 0.5$ presumabpy bebond tie theprejical precondition $\kappa\ll 1$) is donsiderahly higher than the analytical xolution. Although the simulation result reveals lower velicity than the expetimtntal one, the tendqncy of tve higher velocity tyan the theoreticsl one for the narrow gap iv quapitatively similar to that in the experiment. Thus, the present simulation qlso iidycates the lresrnce mf che addytioiau narrow-gap repuksive force.
For an undeformed spierical butbje at small but non-zero Reynplas numbers $0 | momentum and kinetic-energy budgets. Further, the drag the motion shows agreement with the with error of less $0.043\%$. The drag for the parallel motion approaches the solution of [@mag2003] with increasing $\varepsilon$. To make sure of numerical stability and we set the clearance parameter in a range of $10^{-3}\leq \varepsilon\leq 9$. We the tests varying size $R_{\rm max}$ of the computational domain, the number $N_\xi$ of nodes in the gap between bubble and the wall, and the number $N_\eta$ nodes describing the bubble We confirmed that the relative in migration velocity and drag for the perpendicular parallel bubble motions to those obtained on the base meshes $N_\xi\times N_\eta = 200 \times 200$ for $\varepsilon$ decrease the size max}$ the $N_\xi$ and $N_\eta$. convergence behaviors, we deduce to obtain $V_{B3}$ with errors of much less than $1\%$ the present meshes, which are accurate enough for subsequent discussion. Migration velocity ------------------ As shown in inset of figure \[fig:migvel\_comp\_exp\], the ratio of the simulation to the analytical migration velocity becomes unity as $\kappa (=(1+\varepsilon)^{-1})$ zero, indicating the result consistent the solution ([@mag2003]) small $\kappa (< \sim 0.3)$. Such a consistency between the different may deny the erroneous prediction of the migration velocity in which possibility was pointed by [@wan2006]. By contrast, simulated velocity for the bubble the with than bubble (i.e., for $\kappa > 0.5$ presumably beyond the theoretical $\kappa\ll 1$) is considerably Although the simulation result reveals lower velocity than experimental one, the tendency of the higher than the theoretical one for the narrow gap is qualitatively similar to in the the present simulation also indicates the presence of additional narrow-gap repulsive force. an undeformed spherical bubble at small but non-zero Reynolds $0 | momentum and kinetic-energy bUdgets. FurtHer, thE drAg fOrCe foR the Perpendicular mOTion Shows quantitative agreeMent wItH The iNFiNite sEries soLUtION of [@BaR1968] wIth An ERrOr of lEss Than $0.043\%$. The Drag force fOr tHe Parallel motiON aPproaches tHe wIde-gap solutiOn oF [@mag2003] wiTh IncREasinG $\vaRepsiLon$. To mAKe sure Of numericAl STabiliTY and accURAcY, we sEt the clearance parAMeTEr in a range of $10^{-3}\leQ \varepSiLOn\LEQ 9$. We PerFormed the cOnVergeNCe tests WItH VARyiNG the size $R_{\rm maX}$ of the compuTAtiOnal doMaIn, tHE numbeR $N_\xi$ oF nODes In the gap betWeen The bubble And the WAll, and tHE number $n_\eta$ of NodEs dEscrIBiNg The BuBBle SUrFacE. we cOnfirmed ThAt The reLatiVE ERRors In tHe miGratiOn velocity $V_{B3}$ aNd tHe drAG foRces fOr the PerpEnDiculAr and pArallEl Bubble motions to ThosE obtained On tHe BasE mEshes $n_\Xi\timeS N_\eTa = 200 \tImes 200$ for Various $\VArePsILON$ dEcrease with increasInG THe Size $R_{\rm mAx}$ and tHE rEsOLutions $N_\Xi$ And $n_\eta$. fROm the ConvERgEnce behaViors, wE DeDuCe to obtAiN the miGrAtiOn vElociTIes $V_{b3}$ with eRrors of mUch leSS than $1\%$ using the pREsent base meshES, wHICh ARe acCurAte enough foR the SUbseQuenT DiScuSSion.
MIgratIoN VeLOcity
------------------
As shown in the inSeT of figUre \[fiG:migvel\_comp\_exP\], the ratio oF THE simulatIon tO ThE Analytical migrAtion Velocity beCOmes closE to unIty as $\kapPa (=(1+\varepsiLON)^{-1})$ approacHes ZerO, inDicATInG the simulatioN REsulT iS consisTenT with thE anAlyTicAl sOlUtion ([@mag2003]) fOr small $\kApPa (< \SiM 0.3)$. SUch A consIStency beTwEen ThE diFfereNT approAches May dEnY tHE erRoneous PReDICtioN oF tHe miGraTiOn velOcitY In [@mAg2003], of whiCh the possIbiLIty wAs PoInted ouT by [@wan2006]. By contrAsT, the simulaTeD miGratioN VElocity fOr the bubble closer to the wALl with tHe cLearaNce sHorter thaN thE bubblE raDIus (i.e., fOr $\kappA > \sim 0.5$ pReSumABLy beyOND tHe tHeOretical prECOndItion $\KaPpa\lL 1$) is consIderably higher than THe aNalytical soluTioN. AltHOUgH thE SiMUlaTiON reSULt reveals lower vElocity thaN tHE eXperimentaL One, ThE tendenCy of the HigheR VelocitY than the tHeoreticaL oNe foR THe nArrow gap is QualitatIvely simiLAr to tHAt In the ExpErimenT. THus, The prEsent sIMulAtion Also inDiCates tHe preSeNce of the Additional narrow-gap repuLsive fOrce.
FOr aN undeformEd sPHerIcal bubblE at sMall but non-ZerO ReYnoldS nuMBers $0 | momentum and kinetic-ener gy budgets . Fur the r,th e dr ag f orce for the p e rpen dicular motion shows q uanti ta t ivea gr eemen t witht he i nfi ni te se ri e ssolut ion of [@b ar1968] wi than error of le s sthan $0.04 3\% $. The dragfor ce for t hep arall elmotio n appr o achesthe wide- ga p solut i on of [ @ m ag 2003 ] with increasing $\ v arepsilon$. To makesu r eo f nu mer ical stabi li ty an d accura c y, w e se t the clearanc e parameter ina rang eof$ 10^{-3 }\leq \ v are psilon\leq9$.We perfor med th e conver g ence te sts wi thvar ying th esiz e$ R_{ \ rm ma x }$of the c om pu tatio nald o m a in,the num ber $ N_\xi$ of nod esin t h e g ap be tween the b ubble and t he wa ll , and the numbe r $N _\eta$ of no de s d es cribi n g thebub ble surfac e. We c o nfi rm e d th at the relative er ro r s i n the mi gratio n v el o city $V_ {B 3}$ and t he dr ag f o rc es for t he per p en di cular a nd paral le l b ubb le mo t ions to th ose obta inedo n the base mes h es $N_\xi\tim e sN _ \e t a =200 \times 200 $ fo r var ious $\ var e psilo n$ de cr e as e with increasing th esize $ R_{\r m max}$ and t he resolut i o n s $N_\xi $ an d $ N _\eta$. From t he co nvergenceb ehaviors , wededuce t o obtaint h e migrat ion ve loc iti e s $ V_{B3}$ withe r rors o f muchles s than$1\ %$usi ngth e present base me sh es ,wh ich area ccurateen oug hfor thes ubsequ ent d iscu ss io n .
Migrati o nv e loci ty
- ---- --- -- ----- ---Asshown i n the ins eto f fi gu re \[fig: migvel\_comp\ _e xp\], thera tio of th e simulati on to the analytical mi g rationvel ocity bec omes clos e t o unit y a s $\kap pa (=( 1+\va re psi l o n)^{- 1 } )$ ap pr oaches zer o , in dicat in g th e simul ation result is co n sis tent with the an alyt i c al so l ut i on([ @ mag 2 0 03]) for small$\kappa (< \ s im 0.3)$. Su c h a c onsiste ncy bet weent he diff erent app roaches m ay den y the erroneous predict ion of th e migr a ti on ve loc ity in [ @ma g2003 ], ofw hic h the possi bi lity w as po in ted outby [@wan2006]. By contr ast, t he si mul ated migr ati o n v elocity f or t he bubbleclo ser to t hew all w itht he cl e aranc e sh o rter than th e b u b bl e radius (i . e . , f or $\ kap p a > \s im 0 .5$ presumably be y ond the theore tica l pre con d itio n$\kappa\ll 1$) is c o n siderabl yhigher than the ana ly t icalsoluti on. Al thought h es imulat ionres ult revea lslo w er velo ci ty than t he e xp erimen tal on e , th e tendency of thehighe r veloc i tythanth e theor e tica l one forthe narrowgap is qua litat ively s im ilar t o t ha t in the e x periment. Thus , the p re sent si mulati on a l s o ind icat es th e presenc e of th ea ddi tion al na rr ow-g ap repuls i ve force .
F or an u nd efo r m ed sph e ri c a l bubble a t s mallb u t non-zero Reyn o ld s numb ers $0 | momentum_and kinetic-energy_budgets. Further, the drag_force for_the_perpendicular motion_shows_quantitative agreement with_the infinite series_solution of [@bar1968] with_an error of_less_than $0.043\%$. The drag force for the parallel motion approaches the wide-gap solution of_[@mag2003]_with increasing_$\varepsilon$._To_make sure of numerical stability_and accuracy, we set the_clearance parameter_in a range of $10^{-3}\leq \varepsilon\leq 9$. We_performed_the convergence tests_with varying the size $R_{\rm max}$ of the computational_domain, the number $N_\xi$ of nodes_in the gap_between_the_bubble and the wall,_and the number $N_\eta$ of nodes_describing the bubble surface. We confirmed_that the relative errors in the migration_velocity $V_{B3}$ and the drag forces_for the perpendicular and parallel_bubble motions_to those obtained on the_base meshes $N_\xi\times_N_\eta =_200 \times 200$_for various $\varepsilon$ decrease with increasing_the size $R_{\rm_max}$ and the resolutions $N_\xi$ and_$N_\eta$._From the convergence_behaviors,_we_deduce to_obtain the migration_velocities_$V_{B3}$ with_errors_of much less than $1\%$ using_the_present base meshes, which are accurate enough_for the subsequent discussion.
Migration_velocity
------------------
As_shown in the inset_of figure \[fig:migvel\_comp\_exp\], the ratio_of the simulation to the analytical_migration velocity_becomes close_to unity as $\kappa (=(1+\varepsilon)^{-1})$ approaches zero, indicating the simulation result_is consistent with the analytical solution_([@mag2003]) for small $\kappa_(< \sim_0.3)$._Such a consistency_between_the different_approaches may deny the erroneous prediction of_the migration_velocity in [@mag2003], of which the_possibility was pointed out_by_[@wan2006]. By contrast, the simulated migration_velocity for the bubble closer to_the wall with the clearance_shorter_than_the bubble radius (i.e., for_$\kappa > \sim 0.5$ presumably beyond_the theoretical precondition_$\kappa\ll 1$) is considerably higher than the_analytical_solution. Although the simulation result reveals_lower_velocity than the experimental one, the_tendency_of_the higher velocity than the_theoretical one for the narrow gap_is qualitatively similar to that in the experiment. Thus,_the present simulation_also indicates the presence of_the_additional_narrow-gap repulsive force.
For an undeformed spherical bubble at small but_non-zero Reynolds_numbers $0 |
}\[eqX\]\
&&\
Q\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2 \[eqq2\]\
&&\
Q\_1’&=&nT -(++ -)Q\_1+e\^Q\_2.\[eqq1\] It must be noticed that the first four equations are formally identical to those of Lindblom & Detweiler [@LD; @DL], but in terms of the new variables defined by Eq. (\[Wh\])–(\[Eh\]).
Boundary conditions
-------------------
We seek for solutions of the perturbation equations which are regular at the origin. Assuming that all variables near the center of the star have the form $x(r)=x(0)+O(r^2)$, and expanding equations (\[exH0\])–(\[eqq1\]) we find that the following relations must be satisfied at the origin: H\_[0]{}(0)&=&K(0)\[H00\]\
[V]{}(0)&=&-(0)\[V0\]\
[X]{}(0)&=&(\_0+p\_0)e\^[\_0/2]{} {(0)+K(0)}\
&&\[X0\]\
H\_[1]{}(0)&=&\[H10\]\
Q\_[2]{}(0)&=&Q\_[1]{}(0).\[q20\] The values of $\hat E_{lm}$ and $\hat\Sigma_{lm}$ at the origin are given by (\[exE\]), (\[exSigma\]) evaluated at $r=0$. Therefore, our equations admit three independent solutions, which are reduced to one by imposing appropriate boundary conditions at the surface of the star. The first is the vanishing of the Lagrangian pressure perturbation X(R\_s)=0;\[Xs0\] and the second is the vanishing of the radial flux Q\_1(R\_s)=0. Once we know the only independent solution at the surface of the star, the Zerilli function and its derivative can be computed in terms of the $H_1$ and $K$, as usual (see for example [@IP] and [@koj] for the expression of the Zerilli function in terms of $H_0$, $K$; the expression in terms of $H_1$ follows by applying Eq.(\[exH0\]) ): Z&=&r\^l(K-e\^H\_[1]{})\
Z’&=&r\^l(K.\
| } \[eqX\]\
& & \
Q\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2 \[eqq2\]\
& & \
Q\_1’&=&nT -(++ -)Q\_1+e\^Q\_2.\[eqq1\ ] It must be noticed that the first four equations are formally identical to those of Lindblom & Detweiler [ @LD; @DL ], but in terms of the newfangled variable star defined by Eq. (\[Wh\])–(\[Eh\ ]).
Boundary conditions
-------------------
We seek for solution of the perturbation equations which are even at the origin. Assuming that all variable near the center of the star get the form $ x(r)=x(0)+O(r^2)$, and expanding equations (\[exH0\])–(\[eqq1\ ]) we find that the following relations must be quenched at the origin: H\_[0]{}(0)&=&K(0)\[H00\]\
[ V]{}(0)&=&-(0)\[V0\]\
[ X]{}(0)&=&(\_0+p\_0)e\^[\_0/2 ] { } { (0)+K(0)}\
& & \[X0\]\
H\_[1]{}(0)&=&\[H10\]\
Q\_[2]{}(0)&=&Q\_[1]{}(0).\[q20\ ] The values of $ \hat E_{lm}$ and $ \hat\Sigma_{lm}$ at the origin are yield by (\[exE\ ]), (\[exSigma\ ]) evaluated at $ r=0$. Therefore, our equation admit three autonomous solutions, which are reduce to one by imposing appropriate boundary conditions at the surface of the asterisk. The first is the vanishing of the Lagrangian pressure perturbation X(R\_s)=0;\[Xs0\ ] and the second is the vanishing of the radial flux Q\_1(R\_s)=0. Once we know the only independent solution at the surface of the star, the Zerilli function and its derivative can be computed in terms of the $ H_1 $ and $ K$, as usual (see for example [ @IP ] and [ @koj ] for the expression of the Zerilli affair in terms of $ H_0 $, $ K$; the expression in term of $ H_1 $ follow by applying Eq.(\[exH0\ ]) ): Z&=&r\^l(K - e\^H\_[1]{})\
Z’&=&r\^l(K.\ | }\[eqX\]\
&&\
E\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2 \[eqq2\]\
&&\
Q\_1’&=&nT -(++ -)Q\_1+e\^Q\_2.\[eqq1\] Lt must be noticgd that vhe firat four dquations are formally identmcal to tyose of Lindblom & Detwdiler [@LD; @FL], but ib tecms of the new vediables definsf by Xq. (\[Wh\])–(\[Eh\]).
Boundary conditionv
-------------------
We seek for smljtnons of the perturbation equations wrich arr gegular at the oribyn. Aasuming that all variables near ths centeg of the star havr the form $x(r)=x(0)+O(r^2)$, and expanfing equations (\[exH0\])–(\[eqq1\]) ae find thaj thq following rdlations mlvt be satiafied at the origin: H\_[0]{}(0)&=&K(0)\[H00\]\
[V]{}(0)&=&-(0)\[V0\]\
[X]{}(0)&=&(\_0+p\_0)e\^[\_0/2]{} {(0)+K(0)}\
&&\[X0\]\
F\_[1]{}(0)&=&\[H10\]\
Q\_[2]{}(0)&=&Q\_[1]{}(0).\[q20\] The valuew if $\jdt E_{lm}$ and $\iat\Sigia_{lm}$ at the origin ase givem by (\[exE\]), (\[exSigka\]) xvalyated at $r=0$. Therefore, mur equations admij three inge'endent solutions, whixh are teducad tu ond bg mmpksing wppcopriate bohndary condutions at the surfave if the star. Tge firft is the vanishing of the Lagrangian prevsude perturbation X(R\_s)=0;\[Xs0\] abd the second is the ganishing of the radial flux Q\_1(R\_s)=0. Once we know the only indapendxng silmtiov ah the surface of the star, the Zerilli functiog amd its derivativc can be computed on trtms of the $H_1$ avd $K$, as uaual (see for examppe [@IP] agd [@koh] for the exptession of the Zerilli funcrion in termf of $H_0$, $K$; the expresdion in teros og $H_1$ fpllows by applying Eq.(\[exK0\]) ): Z&=&r\^l(I-e\^H\_[1]{})\
Z’&=&r\^l(K.\
| }\[eqX\]\ &&\ Q\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2 \[eqq2\]\ &&\ Q\_1’&=&nT It be noticed the first four those Lindblom & Detweiler @DL], but in of the new variables defined by (\[Wh\])–(\[Eh\]). Boundary conditions ------------------- We seek for solutions of the perturbation equations which regular at the origin. Assuming that all variables near the center of the have form and equations (\[exH0\])–(\[eqq1\]) we find that the following relations must be satisfied at the origin: H\_[0]{}(0)&=&K(0)\[H00\]\ [V]{}(0)&=&-(0)\[V0\]\ {(0)+K(0)}\ &&\[X0\]\ H\_[1]{}(0)&=&\[H10\]\ Q\_[2]{}(0)&=&Q\_[1]{}(0).\[q20\] The values of $\hat and $\hat\Sigma_{lm}$ at the are given by (\[exE\]), (\[exSigma\]) at Therefore, our admit independent which are reduced one by imposing appropriate boundary conditions at the surface of the star. The first is the vanishing the Lagrangian X(R\_s)=0;\[Xs0\] and second the of the radial Once we know the only independent surface of the star, the Zerilli function and derivative can computed in terms of the $H_1$ $K$, as usual (see for example [@IP] and for the expression of the Zerilli function in terms of $H_0$, $K$; the expression in $H_1$ follows by applying ): Z&=&r\^l(K-e\^H\_[1]{})\ Z’&=&r\^l(K.\ | }\[eqX\]\
&&\
Q\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2 \[eqq2\]\
&&\
Q\_1’&=&nT -(++ -)Q\_1+e\^Q\_2.\[eqq1\] It mUst be noticEd thaT thE fiRsT fouR equAtions are formaLLy idEntical to those of LindblOm & DetWeILer [@Ld; @dL], But in Terms of THe NEW vaRiAbLes DeFInEd by EQ. (\[Wh\])–(\[eh\]).
BoundAry conditiOns
-------------------
we Seek for solutIOnS of the pertUrbAtion equatioNs wHich arE rEguLAr at tHe oRigin. assumiNG that aLl variablEs NEar the CEnter of THE sTar hAve the form $x(r)=x(0)+O(r^2)$, anD ExPAnding equationS (\[exH0\])–(\[eqQ1\]) wE FiND ThaT thE following ReLatioNS must be SAtISFIed AT the origin: H\_[0]{}(0)&=&K(0)\[H00\]\
[v]{}(0)&=&-(0)\[V0\]\
[X]{}(0)&=&(\_0+p\_0)e\^[\_0/2]{} {(0)+K(0)}\
&&\[X0\]\
H\_[1]{}(0)&=&\[H10\]\
Q\_[2]{}(0)&=&Q\_[1]{}(0).\[Q20\] the Values Of $\Hat e_{Lm}$ and $\hAt\SigMa_{LM}$ at The origin arE givEn by (\[exE\]), (\[exsigma\]) eVAluated AT $r=0$. ThereFore, ouR eqUatIons ADmIt ThrEe INdePEnDenT SolUtions, whIcH aRe redUced TO ONE by iMpoSing ApproPriate boundarY coNditIOns At the SurfaCe of ThE star. the firSt is tHe Vanishing of the LAgraNgian presSurE pErtUrBatioN x(R\_s)=0;\[Xs0\] aNd tHe sEcond is The vaniSHinG oF THE rAdial flux Q\_1(R\_s)=0. Once we KnOW ThE only indEpendeNT sOlUTion at thE sUrfAce oF THe staR, the zErIlli funcTion anD ItS dErivatiVe Can be cOmPutEd iN termS Of thE $H_1$ and $K$, As usual (sEe for EXample [@IP] and [@koj] FOr the expressiON oF THe zErilLi fUnction in teRms oF $h_0$, $K$; thE expREsSioN In terMs of $H_1$ FoLLoWS by applying Eq.(\[exH0\]) ): Z&=&r\^l(k-e\^h\_[1]{})\
Z’&=&r\^l(K.\
| }\[eqX\]\
&&\
Q\_2’&=&-Q\_ 2+Q\_1 -nT \_1p’ Q\_ 2 \ [e qq2\ ]\
& &\
Q\_1’&=&nT- (++-)Q\_1+e\^Q\_2.\[eqq1\ ] Itmu s t be no ticed that t h ef i rst f ou r e qu a ti ons a reformall y identica l t othose of Lin d bl om & Detwe ile r [@LD; @DL] , b ut inte rms of th e n ew va riable s defin ed by Eq. ( \ [Wh\]) – (\[Eh\] ) .
Boun dary conditions
- - -- - --------------
We s ee k f o r so lut ions of th epertu r batione qu a t i ons which are reg ular at the ori gin. A ss umi n g that allva r iab les near th e ce nter of t he sta r have t h e form$x(r)= x(0 )+O (r^2 ) $, a ndex p and i ng eq u ati ons (\[e xH 0\ ])–(\ [eqq 1 \ ] ) wefin d th at th e following r ela tion s mu st be sati sfie dat th e orig in: H \_ [0]{}(0)&=&K(0) \[H0 0\]\
[V]{ }(0 )& =&- (0 )\[V0 \ ]\
[X] {}( 0)& =&(\_0+ p\_0)e\ ^ [\_ 0/ 2 ] { }{(0)+K(0)}\
&&\[X0 \] \ H\ _[1]{}(0 )&=&\[ H 10 \] \
Q\_[2]{ }( 0)& =&Q\ _ [ 1]{}( 0).\ [ q2 0\] Thevalues of $ \hat E_ {l m}$ an d$\h at\ Sigma _ {lm} $ at t he origi n are given by (\[ex E \]), (\[exSig m a\ ] ) e v alua ted at $r=0$.Ther e fore , ou r e qua t ionsadmit t h re e independent soluti on s, whi ch ar e reduced toone by imp o s i ng appro pria t eb oundary condit ionsat the sur f ace of t he st ar. Thefirst ist h e vanish ing of th e L a g ra ngian pressur e pert ur bationX(R \_s)=0; \[X s0\ ] a ndth e secondis the v an is hi ng of ther adial fl ux Q\ _1 (R\ _s)=0 . Oncewe kn ow t he o n lyindepen d en t solu ti on atthe s urfac e of the star,the Zeril lif unct io nand its derivative c an be comput ed in terms o f the $H _1$ and $K$, as usual ( s ee forexa mple[@IP ] and [@k oj] for t hee xpress ion of theZe ril l i func t i on in t erms of $H _ 0 $,$K$;th e ex pressio n in terms of $H_1 $ fo llows by appl yin g Eq . ( \[ exH 0 \] ) ): Z & =&r \ ^ l(K-e\^H\_[1]{} )\
Z’&=&r\ ^l ( K. \
| }\[eqX\]\
&&\
Q\_2’&=&-Q\_2+Q\_1 -nT\_1p’Q\_2_\[eqq2\]\
&&\
Q\_1’&=&nT -(++_-)Q\_1+e\^Q\_2.\[eqq1\] It must be_noticed that_the_first four_equations_are formally identical_to those of_Lindblom & Detweiler [@LD;_@DL], but in_terms_of the new variables defined by Eq. (\[Wh\])–(\[Eh\]).
Boundary conditions
-------------------
We seek for solutions of the_perturbation_equations which_are_regular_at the origin. Assuming that_all variables near the center_of the_star have the form $x(r)=x(0)+O(r^2)$, and expanding equations_(\[exH0\])–(\[eqq1\])_we find that_the following relations must be satisfied at the origin:_H\_[0]{}(0)&=&K(0)\[H00\]\
[V]{}(0)&=&-(0)\[V0\]\
[X]{}(0)&=&(\_0+p\_0)e\^[\_0/2]{} {(0)+K(0)}\
&&\[X0\]\
H\_[1]{}(0)&=&\[H10\]\
Q\_[2]{}(0)&=&Q\_[1]{}(0).\[q20\] The values of $\hat_E_{lm}$ and $\hat\Sigma_{lm}$_at_the_origin are given by_(\[exE\]), (\[exSigma\]) evaluated at $r=0$. Therefore,_our equations admit three independent solutions,_which are reduced to one by imposing_appropriate boundary conditions at the surface_of the star. The first_is the_vanishing of the Lagrangian pressure_perturbation X(R\_s)=0;\[Xs0\] and_the second_is the vanishing_of the radial flux Q\_1(R\_s)=0. Once_we know the_only independent solution at the surface_of_the star, the_Zerilli_function_and its_derivative can be_computed_in terms_of_the $H_1$ and $K$, as usual_(see_for example [@IP] and [@koj] for the_expression of the Zerilli_function_in terms of $H_0$,_$K$; the expression in terms_of $H_1$ follows by applying Eq.(\[exH0\])_): Z&=&r\^l(K-e\^H\_[1]{})\
Z’&=&r\^l(K.\
|
C. Chamon, R. Jackiw, Y. Nishida, S. Y. Pi and L. Santos, Phys. Rev. B [**81**]{}, 224515 (2010)
[^1]: In the case of isovector fermion, the corresponding conjugated field is defined by $\Psi^c_a \equiv \left(\begin{array}{cc}0 & i \sigma^2 \\i \sigma^2 & 0\end{array}\right) \Psi^*_a$. Once we impose the conjugation constraint $\Psi^c=\Psi$ we only have the trivial solution $\Psi=0$.
---
abstract: 'Graph link prediction is an important task in cyber-security: relationships between entities within a computer network, such as users interacting with computers, or system libraries and the corresponding processes that use them, can provide key insights into adversary behaviour. Poisson matrix factorisation (PMF) is a popular model for link prediction in large networks, particularly useful for its scalability. In this article, PMF is extended to include scenarios that are commonly encountered in cyber-security applications. Specifically, an extension is proposed to explicitly handle binary adjacency matrices and include known covariates associated with the graph nodes. A seasonal PMF model is also presented to handle dynamic networks. To allow the methods to scale to large graphs, variational methods are discussed for performing fast inference. The results show an improved performance over the standard PMF model and other common link prediction techniques.'
address: |
$\color{blue}\dagger$ – Department of Mathematics, Imperial College London\
$\color{blue}\star$ – Advanced Research in Cyber-Systems, Los Alamos National Laboratory
author:
- '$\color{blue}{}^\dagger$'
- '$\color{blue}{}^\star$'
- '$\color{blue}{}^\dagger$'
bibliography:
- 'biblio.bib'
title: Graph link prediction in computer networks using Poisson matrix factorisation
---
,
,\
\
anomaly detection, dynamic networks, new link prediction, Poisson matrix factorisation, statistical cyber-security, variational inference
Introduction {#intro}
============
In recent years, there has been a significant increase in investment from both government and industry in improving cyber-security using statistical and machine learning techniques on a wide range of data collected from computer networks [@HeardAdams | C. Chamon, R. Jackiw, Y. Nishida, S. Y. Pi and L. Santos, Phys. Rev. B [ * * 81 * * ] { }, 224515 (2010)
[ ^1 ]: In the case of isovector fermion, the corresponding conjugated field is define by $ \Psi^c_a \equiv \left(\begin{array}{cc}0 & i \sigma^2 \\i \sigma^2 & 0\end{array}\right) \Psi^*_a$. Once we levy the conjugation constraint $ \Psi^c=\Psi$ we only own the trivial solution $ \Psi=0$.
---
abstract:' Graph link prediction is an significant task in cyber - security: relationship between entities within a computer net, such as users interacting with computers, or system library and the corresponding summons that use them, can provide cardinal insights into adversary behaviour. Poisson matrix factorisation (PMF) is a democratic model for link prediction in large networks, particularly utilitarian for its scalability. In this article, PMF is extended to admit scenarios that are commonly encountered in cyber - security application. Specifically, an extension is proposed to explicitly handle binary adjacency matrices and include know covariates associated with the graph nodes. A seasonal PMF model is also presented to handle dynamic networks. To allow the methods to scale to bombastic graphs, variational methods are hash out for perform fast inference. The results show an improved operation over the standard PMF model and early common link prediction techniques.'
savoir-faire: |
$ \color{blue}\dagger$ – Department of Mathematics, Imperial College London\
$ \color{blue}\star$ – Advanced Research in Cyber - Systems, Los Alamos National Laboratory
author:
-' $ \color{blue}{}^\dagger$'
-' $ \color{blue}{}^\star$'
-' $ \color{blue}{}^\dagger$'
bibliography:
-' biblio.bib'
title: Graph link prediction in computer networks using Poisson matrix factorization
---
,
, \
\
anomaly detection, dynamic networks, new link prediction, Poisson matrix factorisation, statistical cyber - security, variational inference
Introduction { # presentation }
= = = = = = = = = = = =
In late class, there has been a significant increase in investment from both politics and industry in improving cyber - security using statistical and machine learning techniques on a wide range of datum collected from computer networks [ @HeardAdams | C. Chwmon, R. Jackiw, Y. Nishida, S. Y. Ki and L. Santos, Phis. Eev. B [**81**]{}, 224515 (2010)
[^1]: In fhe case of isovector fermion, the cocrespondibg conjugated field is defined hy $\Psi^c_a \equmv \left(\begin{arrab}{dc}0 & i \slyma^2 \\i \digmc^2 & 0\end{array}\right) \Psi^*_a$. Once we impose the cunlugation constraint $\Psi^c=\Psi$ we only rave thr hrivial solutijn $\Pxy=0$.
---
abanrcct: 'Graph link prediction is an importent task in cybrr-security: relationships bftwefn entities within a computer netrirk, such as jsers integccting with computers, or system libraries xnd tke correspobdung krocesses thet use them, can provide kef insignts into adverxarb beyaviour. Poisson matrie factorisation (PMF) ys a popunax model for link preducrion hn ldrge betdoris, pzrticuparmy useful ror its scaoability. In this aruicjv, PMF is extehded tj ynclude scenarios that are commonly encmunfered in cyber-security qpplications. Specificwlly, an evtension is proposed to explicitly handle binary ddjacxnzy nabricdw wnd include known covariates associated with ege ggaph nodes. A seasjnal PMF mocep of also presenjed to kznsle dynamic networns. To ajlow rhe methoqs tp scale to large graphs, varuational metkodw are discussed fox performing fasj infetence. The results show cn impdoved perfogmance ovsf the standard POF kogel and other common link [redictioi teckniques.'
aadrexs: |
$\color{bluf}\daggcs$ – Department of Mwthemctics, Imperial Follege London\
$\color{blue}\star$ – Advanced Resgarwh pn Cyber-Svstems, Los Alamos Nwtional Laborajory
authjr:
- '$\couor{blue}{}^\dagfer$'
- '$\colmr{blue}{}^\star$'
- '$\solor{blue}{}^\daggat$'
bibliography:
- 'biblio.byb'
tirle: Traph lkvk prediction on computvr netqorks using Poissok mattis factorisation
---
,
,\
\
animaly detection, dyvamyc nxtworhv, new link psediztium, Poirson mauxix wactprisation, statistican cyger-security, variatookal infergnce
Introquction {#intro}
============
On recent years, thtre hav bxen a xiggificant increase in investmenf from bohh novernment anq ineustry in im'roving cyber-security using statistical end machine learning tgchniques on a wide rcnne of data cplleceed from womputer networks [@HeqrdAdams | C. Chamon, R. Jackiw, Y. Nishida, S. and Santos, Phys. B [**81**]{}, 224515 of fermion, the corresponding field is defined $\Psi^c_a \equiv \left(\begin{array}{cc}0 & i \sigma^2 \sigma^2 & 0\end{array}\right) \Psi^*_a$. Once we impose the conjugation constraint $\Psi^c=\Psi$ we only the trivial solution $\Psi=0$. --- abstract: 'Graph link prediction is an important task cyber-security: between within computer network, such as users interacting with computers, or system libraries and the corresponding processes that them, can provide key insights into adversary behaviour. matrix factorisation (PMF) is popular model for link prediction large particularly useful its In article, PMF is to include scenarios that are commonly encountered in cyber-security applications. Specifically, an extension is proposed to explicitly binary adjacency include known associated the nodes. A seasonal is also presented to handle dynamic the methods to scale to large graphs, variational are discussed performing fast inference. The results show improved performance over the standard PMF model and common link prediction techniques.' address: | $\color{blue}\dagger$ – Department of Mathematics, Imperial College London\ $\color{blue}\star$ Research in Cyber-Systems, Los National Laboratory author: '$\color{blue}{}^\dagger$' '$\color{blue}{}^\star$' '$\color{blue}{}^\dagger$' - 'biblio.bib' Graph link prediction in computer networks using Poisson matrix factorisation --- ,\ \ anomaly detection, dynamic networks, new link prediction, Poisson statistical variational inference Introduction ============ In recent years, has a significant increase in both and cyber-security statistical machine learning techniques on wide range of data collected computer networks [@HeardAdams | C. Chamon, R. Jackiw, Y. Nishida, S. Y. PI and L. SantoS, Phys. rev. b [**81**]{}, 224515 (2010)
[^1]: In ThE casE of iSovector fermioN, The cOrresponding conjugated Field Is DEfinED bY $\Psi^c_A \equiv \lEFt(\BEGin{ArRaY}{cc}0 & I \sIGmA^2 \\i \sigMa^2 & 0\eNd{array}\Right) \Psi^*_a$. ONce We Impose the conJUgAtion constRaiNt $\Psi^c=\Psi$ we oNly Have thE tRivIAl solUtiOn $\Psi=0$.
---
AbstraCT: 'Graph Link prediCtIOn is an IMportanT TAsK in cYber-security: relatIOnSHips between entIties wItHIn A COmpUteR network, suCh As useRS interaCTiNG WIth COmputers, or sysTem librarieS And The corReSpoNDing prOcessEs THat Use them, can pRoviDe key insiGhts inTO adversARy behavIour. PoIssOn mAtriX FaCtOriSaTIon (pmF) Is a POpuLar model FoR lInk prEdicTION In laRge NetwOrks, pArticularly usEfuL for ITs sCalabIlity. in thIs ArticLe, PMF iS exteNdEd to include scenArioS that are cOmmOnLy eNcOunteREd in cyBer-SecUrity apPlicatiONs. SPeCIFIcAlly, an extension is pRoPOSeD to expliCitly hANdLe BInary adjAcEncY matRICes anD incLUdE known coVariatES aSsOciated WiTh the gRaPh nOdeS. A seaSOnal pMF modEl is also PreseNTed to handle dynAMic networks. To ALlOW ThE MethOds To scale to laRge gRAphs, VariATiOnaL MethoDs are DiSCuSSed for performing fasT iNferenCe. The Results show an Improved peRFORmance ovEr thE StANdard PMF model aNd othEr common liNK predictIon teChniques.'
Address: |
$\coLOR{blue}\dagGer$ – depArtMenT OF MAthematics, ImpERIal COlLege LonDon\
$\Color{blUe}\sTar$ – advAncEd research iN Cyber-SyStEmS, LOs alaMos NaTIonal LabOrAtoRy
AutHor:
- '$\coLOr{blue}{}^\DaggeR$'
- '$\colOr{BlUE}{}^\stAr$'
- '$\color{BLuE}{}^\DAggeR$'
bIbLiogRapHy:
- 'BibliO.bib'
TItlE: Graph lInk predicTioN In coMpUtEr netwoRks using PoissOn Matrix factOrIsaTion
---
,
,\
\
anOMAly detecTion, dynamic networks, new lINk prediCtiOn, PoiSson Matrix facTorIsatioN, stATisticAl cybeR-secuRiTy, vARIatioNAL iNfeReNce
IntroduCTIon {#Intro}
============
in ReceNt years, There has been a signiFIcaNt increase in iNveStmeNT FrOm bOTh GOveRnMEnt AND industry in imprOving cyber-SeCUrIty using stATisTiCal and mAchine lEarniNG techniQues on a wiDe range of DaTa coLLEctEd from compUter netwOrks [@HeardaDams | C. Chamon, R. Jackiw, Y. N ishida, S. Y. P i a ndL. San tos, Phys. Rev. B [**8 1**]{}, 224515 (2010)
[^1] :I n th e c ase o f isove c to r fer mi on , t he co rresp ond ing con jugated fi eld i s defined by $\ Psi^c_a \e qui v \left(\beg in{ array} {c c}0 & i \ sig ma^2\\i \s i gma^2& 0\end{a rr a y}\rig h t) \Psi ^ * _a $. O nce we impose the co n jugation const raint$\ P si ^ c =\P si$ we only h av e the trivial so l u t ion $\Psi=0$.
-- -
abstract: 'Gr aph li nk pr e dictio n isan imp ortant task incyber-sec urity: relatio n ships b etween en tit iesw it hi n a c o mpu t er ne t wor k, suchas u sersinte r a c t ingwit h co mpute rs, or system li brar i esand t he co rres po nding proce ssesth at use them, ca n pr ovide key in si ght sintoa dversa rybeh aviour. Poisso n ma tr i x fa ctorisation (PMF)is a p opular m odel f o rli n k predic ti onin l a r ge ne twor k s, particu larlyu se fu l for i ts scala bi lit y.In th i s ar ticle, PMF isexten d ed to includes cenarios that ar e co m monl y e ncounteredin c y ber- secu r it y a p plica tions .S pe c ifically, an extens io n is p ropos ed to explici tly handle b i nary adj acen c ym atrices and in clude known cov a riates a ssoci ated wit h the gra p h nodes.A s eas ona l P M F m odel is alsop r esen te d to ha ndl e dynam icnet wor ks. T o allow t he metho ds t osc ale to l a rge grap hs , v ar iat ional method s are dis cu ss e d f or perf o rm i n g fa st i nfer enc e. Theresu l tsshow an improved pe r form an ce over t he standard P MF model and o the r comm o n link pr ediction techniques.'
a d dress:|
$ \col or{blue}\ dag ger$ – De p artmen t of M athem at ics , Imper i a lCol le ge London\ $\col or {blu e}\star $ – Advanced Resea r chin Cyber-Syst ems , Lo s Al amo s N a tio na l La b o ratory
author:
- '$\colo r{ b lu e}{}^\dagg e r$'
- '$\col or{blue }{}^\ s tar$'
- '$\color {blue}{}^ \d agge r $ '
b ibliograph y:
- 'bi blio.bib' title : G raphlin k pred ic tio n incomput e r n etwor ks usi ng Poiss on ma tr ix facto risation
---
,
,\
\
anom aly d ete ction, dy nam i c n etworks,newlink predi cti on, Pois son matri x fa c to ris a tion, sta t istical c y be r-s e c ur ity, variat i o n alinfer enc e
Intr oduc tion {#intro}
=== = ========
In r ecen t yea rs, ther ehas been a sig nif ic a n t increa se in investm ent from b o th go vernme nt and indust r y i n impro ving cy ber-secur ity u s ing sta ti st i cal an d ma ch ine le arning tech n i ques on a wide r angeo f data col lecte dfrom co m pute r networks [@HeardAda ms | C. Chamon, R. Jackiw,_Y. Nishida, S. Y. Pi_and L. Santos, Phys. Rev. B_[**81**]{}, 224515_(2010)
[^1]:_In the_case_of isovector fermion,_the corresponding conjugated_field is defined by_$\Psi^c_a \equiv \left(\begin{array}{cc}0_&_i \sigma^2 \\i \sigma^2 & 0\end{array}\right) \Psi^*_a$. Once we impose the conjugation constraint $\Psi^c=\Psi$_we_only have_the_trivial_solution $\Psi=0$.
---
abstract: 'Graph link_prediction is an important task_in cyber-security:_relationships between entities within a computer network, such_as_users interacting with_computers, or system libraries and the corresponding processes that_use them, can provide key insights_into adversary behaviour._Poisson_matrix_factorisation (PMF) is a_popular model for link prediction in_large networks, particularly useful for its_scalability. In this article, PMF is extended_to include scenarios that are commonly_encountered in cyber-security applications. Specifically,_an extension_is proposed to explicitly handle_binary adjacency matrices_and include_known covariates associated_with the graph nodes. A seasonal_PMF model is_also presented to handle dynamic networks._To_allow the methods_to_scale_to large_graphs, variational methods_are_discussed for_performing_fast inference. The results show an_improved_performance over the standard PMF model and_other common link prediction_techniques.'
address:_|
_ $\color{blue}\dagger$ – Department of_Mathematics, Imperial College London\
_ $\color{blue}\star$_– Advanced_Research in Cyber-Systems, Los Alamos National Laboratory
author:
- '$\color{blue}{}^\dagger$'
- '$\color{blue}{}^\star$'
- '$\color{blue}{}^\dagger$'
bibliography:
-_'biblio.bib'
title: Graph link prediction in computer_networks using Poisson matrix_factorisation
---
,
,\
\
anomaly_detection,_dynamic networks, new_link_prediction, Poisson_matrix factorisation, statistical cyber-security, variational inference
Introduction {#intro}
============
In_recent years,_there has been a significant increase_in investment from both_government_and industry in improving cyber-security using_statistical and machine learning techniques on_a wide range of data_collected_from_computer networks [@HeardAdams |
pi=0$) and highly suppressed.
We observed also the two levels located just above the proton separation energy ($S_p=5513.8$ keV). The first one corresponds to a $4^+$ state with an energy of 5517 keV. This level was also reported in Refs. [@PhysRevC.75.062801; @Komatsubara2014]. The second level at 5929 keV was previously observed in $\beta$-delayed proton emission by Thomas *et al.* [@Thomas2004] and more recently reported in our previous paper describing the present experiment [@Bennett2013]. The results presented here with the same set of data, but with an independent analysis, confirm the evidence for the observation of a $\gamma$ ray emitted from that level in the present experiment.
[c d c d c d c d c d c c]{} & & & & &\
&&& &&\
$J_n^\pi$ & &$J_n^\pi$ & &$J_n^\pi$ & & $J_n^\pi$ && $J_n^\pi$ && $J_n^\pi$ &\
$2_1^+$ & 1797.1(3) & $2_1^+$ & 1795.9(2) & $2_1^+$ & 1797.3(1) & $2_1^+$ & 1797.4(4)&$2_1^+$ &1797.3(1) &$2_1^+$ &1887\
$2_2^+$ & 2786.4(3) & $2_2^+$ & 2783.5(4) & $2_2^+$ & 2786.4(2) & $2_2^+$ & 2786.8(6)&$2_2^+$ &2786.4(2) &$2_2^+$ &2948\
&& && $0_2^+$ & 3336.4(6) & $0_2^+$ & 3335.3(4)&$0_2^+$ & 3336.4(2)&&\
$3_1^+$ & 3756.8(3) & $(3_1^+)$ & 3756(2) & $3_1^+$ & 3756.9(2) & $3_1^+$ & 3756.9(4)& $3_1^+$ & 3757.1(3)&$3 | pi=0 $) and highly suppressed.
We observed also the two levels locate barely above the proton separation energy ($ S_p=5513.8 $ keV). The first one corresponds to a $ 4^+$ country with an energy of 5517 keV. This level was besides reported in Refs. [ @PhysRevC.75.062801; @Komatsubara2014 ]. The second level at 5929 keV was previously observed in $ \beta$-delayed proton discharge by Thomas * et al. * [ @Thomas2004 ] and more recently reported in our previous newspaper describing the present experiment [ @Bennett2013 ]. The results stage here with the same set of data, but with an independent analysis, confirm the evidence for the observation of a $ \gamma$ beam emitted from that level in the present experiment.
[ c d deoxycytidine monophosphate d c d c vitamin d c d c c ] { } & & & & & \
& & & & & \
$ J_n^\pi$ & & $ J_n^\pi$ & & $ J_n^\pi$ & & $ J_n^\pi$ & & $ J_n^\pi$ & & $ J_n^\pi$ & \
$ 2_1^+$ & 1797.1(3) & $ 2_1^+$ & 1795.9(2) & $ 2_1^+$ & 1797.3(1) & $ 2_1^+$ & 1797.4(4)&$2_1^+$ & 1797.3(1) & $ 2_1^+$ & 1887\
$ 2_2^+$ & 2786.4(3) & $ 2_2^+$ & 2783.5(4) & $ 2_2^+$ & 2786.4(2) & $ 2_2^+$ & 2786.8(6)&$2_2^+$ & 2786.4(2) & $ 2_2^+$ & 2948\
& & & & $ 0_2^+$ & 3336.4(6) & $ 0_2^+$ & 3335.3(4)&$0_2^+$ & 3336.4(2)&&\
$ 3_1^+$ & 3756.8(3) & $ (3_1^+)$ & 3756(2) & $ 3_1^+$ & 3756.9(2) & $ 3_1^+$ & 3756.9(4) & $ 3_1^+$ & 3757.1(3)&$3 | pi=0$) wnd highly suppressed.
We ubserved also tkw two nevels located just above the proton separetiob eneegy ($S_p=5513.8$ keV). The first ond corresplnds to q $4^+$ suate with an enerjg of 5517 keY. Thia levzl was also repotted in Refs. [@PhysRevC.75.062801; @Komadsjbcra2014]. The second level at 5929 keV was prevyously pbderved in $\beta$-qelajeq prknok emission by Thomas *et al.* [@Thomaa2004] and mmre recently teported in our previous pwper describing the prfsent experumene [@Bennett2013]. The results pgzsented herg with the same set of data, but dith cn independgur ajdlysis, confmrm thv evidence fov the otservatoon of a $\gamma$ rab emutted from that level in the present exkeriment.
[c g e d c d c d c d c c]{} & & & & &\
&&& &&\
$J_n^\ph$ & &$J_t^\pi$ & &$H_n^\pk$ & & $J_i^\pi$ && $J_n^\pi$ && $J_i^\pi$ &\
$2_1^+$ & 1797.1(3) & $2_1^+$ & 1795.9(2) & $2_1^+$ & 1797.3(1) & $2_1^+$ & 1797.4(4)&$2_1^+$ &1797.3(1) &$2_1^+$ &1887\
$2_2^+$ & 2786.4(3) & $2_2^+$ & 2783.5(4) & $2_2^+$ & 2786.4(2) & $2_2^+$ & 2786.8(6)&$2_2^+$ &2786.4(2) &$2_2^+$ &2948\
&& && $0_2^+$ & 3336.4(6) & $0_2^+$ & 3335.3(4)&$0_2^+$ & 3336.4(2)&&\
$3_1^+$ & 3756.8(3) & $(3_1^+)$ & 3756(2) & $3_1^+$ & 3756.9(2) & $3_1^+$ & 3756.9(4)& $3_1^+$ & 3757.1(3)&$3 | pi=0$) and highly suppressed. We observed also levels just above proton separation energy corresponds a $4^+$ state an energy of keV. This level was also reported Refs. [@PhysRevC.75.062801; @Komatsubara2014]. The second level at 5929 keV was previously observed in proton emission by Thomas *et al.* [@Thomas2004] and more recently reported in our paper the experiment The results presented here with the same set of data, but with an independent analysis, confirm evidence for the observation of a $\gamma$ ray from that level in present experiment. [c d c c c d d c]{} & & & &&& &&\ $J_n^\pi$ & &$J_n^\pi$ & &$J_n^\pi$ & & $J_n^\pi$ && $J_n^\pi$ && $J_n^\pi$ &\ $2_1^+$ & & $2_1^+$ & $2_1^+$ 1797.3(1) $2_1^+$ 1797.4(4)&$2_1^+$ &1797.3(1) &$2_1^+$ & 2786.4(3) & $2_2^+$ & 2783.5(4) 2786.4(2) & $2_2^+$ & 2786.8(6)&$2_2^+$ &2786.4(2) &$2_2^+$ &2948\ && $0_2^+$ 3336.4(6) & $0_2^+$ & 3335.3(4)&$0_2^+$ & $3_1^+$ & 3756.8(3) & $(3_1^+)$ & 3756(2) & & 3756.9(2) & $3_1^+$ & 3756.9(4)& $3_1^+$ & 3757.1(3)&$3 | pi=0$) and highly suppressed.
We obServed also The twO leVelS lOcatEd juSt above the protON sepAration energy ($S_p=5513.8$ keV). The fIrst oNe COrreSPoNds to A $4^+$ state wITh AN EneRgY oF 5517 kev. THIs Level Was Also repOrted in RefS. [@PhYsrevC.75.062801; @KomatsubARa2014]. the second lEveL at 5929 keV was preVioUsly obSeRveD In $\betA$-deLayed Proton EMissioN by Thomas *Et AL.* [@ThomaS2004] And more RECeNtly Reported in our prevIOuS Paper describinG the prEsENt EXPerImeNt [@Bennett2013]. THe ResulTS presenTEd HERE wiTH the same set of Data, but with AN inDependEnT anALysis, cOnfirM tHE evIdence for thE obsErvation oF a $\gammA$ Ray emitTEd from tHat levEl iN thE preSEnT eXpeRiMEnt.
[C D c D c d C D c d C c]{} & & & & &\
&&& &&\
$J_n^\pi$ & &$J_n^\Pi$ & &$j_n^\Pi$ & & $J_n^\pI$ && $J_n^\pI$ && $j_N^\PI$ &\
$2_1^+$ & 1797.1(3) & $2_1^+$ & 1795.9(2) & $2_1^+$ & 1797.3(1) & $2_1^+$ & 1797.4(4)&$2_1^+$ &1797.3(1) &$2_1^+$ &1887\
$2_2^+$ & 2786.4(3) & $2_2^+$ & 2783.5(4) & $2_2^+$ & 2786.4(2) & $2_2^+$ & 2786.8(6)&$2_2^+$ &2786.4(2) &$2_2^+$ &2948\
&& && $0_2^+$ & 3336.4(6) & $0_2^+$ & 3335.3(4)&$0_2^+$ & 3336.4(2)&&\
$3_1^+$ & 3756.8(3) & $(3_1^+)$ & 3756(2) & $3_1^+$ & 3756.9(2) & $3_1^+$ & 3756.9(4)& $3_1^+$ & 3757.1(3)&$3 | pi=0$) and highly suppress ed.
We ob serve d a lso t he t wo l evels locatedj ustabove the proton separ ation e n ergy ($ S_p=5 513.8$k eV ) . Th efi rst o n ecorre spo nds toa $4^+$ st ate w ith an energ y o f 5517 keV . T his level wa s a lso re po rte d in R efs . [@P hysRev C .75.06 2801; @Ko ma t subara 2 014]. T h e s econ d level at 5929 k e Vw as previouslyobserv ed in $ \be ta$ -delayed p ro ton e m issionb yT h o mas *et al.* [@Th omas2004] a n d m ore re ce ntl y repor ted i no urprevious pa perdescribin g thep resente xperime nt [@B enn ett 2013 ] .Th e r es u lts pr ese n ted here wi th t he sa me s e t o f da ta, but with an independe ntanal y sis , con firmtheev idenc e forthe o bs ervation of a $ \gam ma$ ray e mit te d f ro m tha t level in th e prese nt expe r ime nt . [c d c d c d c d c d c c ]{ } & & && &\
& & &&& \
$J_n^\p i$ &&$J_ n ^ \pi$& &$ J _n ^\pi$ && $J_n ^ \p i$ && $J_ n^ \pi$ & &$J_ n^\ pi$ & \
$2_ 1^+$ & 1797.1( 3) &$ 2_1^+$ & 1795. 9 (2) & $2_1^+$ &1 7 97 . 3(1) &$2_1^+$ & 1 797. 4 (4)& $2_1 ^ +$ &1 7 97.3( 1) &$ 2_ 1 ^+ $ &1887\
$2_2^+$ & 2 78 6.4(3) & $2 _2^+$ & 2783. 5(4) & $2_ 2 ^ + $ & 2786 .4(2 ) & $2_2^+$ & 2786 .8(6) &$2_2^+$ & 2 786.4(2) &$2_ 2^+$ &29 48\
&& && $ 0_2^+$ & 33 36. 4(6 ) & $ 0_ 2^+$ & 3335.3 ( 4 )&$0 _2 ^+$ & 3 336 .4(2)&& \
$ 3_1 ^+$ &37 56.8(3) & $(3_1^+ )$ & 3 75 6(2 ) & $ 3 _1^+$ &37 56. 9( 2)& $3_ 1 ^+$ &3756. 9(4) &$3 _ 1^+ $ & 375 7 .1 ( 3 )&$3 | pi=0$) and_highly suppressed.
We_observed also the two_levels located_just_above the_proton_separation energy ($S_p=5513.8$ keV)._The first one_corresponds to a $4^+$_state with an_energy_of 5517 keV. This level was also reported in Refs. [@PhysRevC.75.062801; @Komatsubara2014]. The second level_at_5929 keV was_previously_observed_in $\beta$-delayed proton emission by_Thomas *et al.* [@Thomas2004] and_more recently_reported in our previous paper describing the present_experiment_[@Bennett2013]. The results_presented here with the same set of data, but_with an independent analysis, confirm the_evidence for the_observation_of_a $\gamma$ ray emitted_from that level in the present_experiment.
[c d c d c d_c d c d c c]{} &_& & & &\
&&& &&\
$J_n^\pi$ &_&$J_n^\pi$ & &$J_n^\pi$ & &_$J_n^\pi$ &&_$J_n^\pi$ && $J_n^\pi$ &\
$2_1^+$ &_1797.1(3) & $2_1^+$_& 1795.9(2)_& $2_1^+$ &_1797.3(1) & $2_1^+$ & 1797.4(4)&$2_1^+$ &1797.3(1)_&$2_1^+$ &1887\
$2_2^+$ &_2786.4(3) & $2_2^+$ & 2783.5(4) &_$2_2^+$_& 2786.4(2) &_$2_2^+$_&_2786.8(6)&$2_2^+$ &2786.4(2)_&$2_2^+$ &2948\
&& &&_$0_2^+$_& 3336.4(6)_&_$0_2^+$ & 3335.3(4)&$0_2^+$ & 3336.4(2)&&\
$3_1^+$ &_3756.8(3)_& $(3_1^+)$ & 3756(2) & $3_1^+$ &_3756.9(2) & $3_1^+$ &_3756.9(4)&_$3_1^+$ & 3757.1(3)&$3 |
power allocation policy based energy demand policy in [@akin2017energy] under energy underflow probability constraint. Energy underflow refers to the case the energy level in the battery falls below a certain level.
[^14]: Although our framework is good for data transmission settings with varying data traffic rate, we consider a fixed data transmission rate for mathematical tractability. We refer interested readers to [@pasch1993comparing] for more details on data traffic types. Voice and video traffic can be modeled with a constant data service rate.
[^15]: We do not define a specific function but consider a general definition for the function that sets the transmission rate. In other words, $g(\cdot)$ is a matter of system design and may depend on the statistics of the energy demand policy. However, for instance, one can set $g(\cdot)$ as a linear function of the energy demand or the available energy in the battery.
[^16]: After the battery reaches the steady-state, the probability of the battery being in state $m$ in any time frame becomes $\pi_{m}$, as defined in Section \[sec:energy\_outage\], and it does not change with state transitions in the consecutive time frames.
[^17]: Here, we simulate the energy arrivals and the channel fading gains with respect to their distributions assuming that Gaussian codebooks are employed. However, one can do simulations considering practical modulation techniques and their corresponding instantaneous mutual information values given the channel conditions.
---
abstract: 'Let $p$ be an odd prime, $m\in {{\mathbb N}}$ and set $q=p^m$, ${\boldsymbol{G}}=\PSL_n(q)$. Let $\theta$ be a standard graph automorphism of ${\boldsymbol{G}}$, $d$ be a diagonal automorphism and ${\operatorname{Fr}}_q$ be the Frobenius endomorphism of $\PSL_n(\overline{{{\mathbb F}}_q})$. We show that every $(d\circ \theta)$-conjugacy class of a $(d\circ \theta,p)$-regular element in ${\boldsymbol{G}}$ is represented in some ${\operatorname{Fr}}_q$-stable maximal torus of $\PSL_n(\overline{{{\mathbb F}}_q})$ and that most of them are of type D. We write out the possible exceptions and show that, in particular, if $n\geq5$ is either odd or a multiple of $4$ and $q>7$, then all such classes are of type D. | power allocation policy based energy requirement policy in [ @akin2017energy ] under department of energy underflow probability constraint. Energy underflow refer to the lawsuit the energy level in the barrage falls below a certain level.
[ ^14 ]: Although our model is beneficial for data infection settings with varying datum traffic rate, we consider a fixed data transmission rate for numerical tractability. We refer interested readers to [ @pasch1993comparing ] for more detail on data traffic type. Voice and video recording traffic can be modeled with a constant data service pace.
[ ^15 ]: We do not define a specific function but consider a general definition for the function that sets the transmittance rate. In other words, $ g(\cdot)$ is a matter of system design and may depend on the statistics of the energy demand policy. However, for case, one can set $ g(\cdot)$ as a linear routine of the department of energy need or the available energy in the battery.
[ ^16 ]: After the barrage reaches the regular - state, the probability of the battery organism in state $ m$ in any time frame becomes $ \pi_{m}$, as defined in Section \[sec: energy\_outage\ ], and it does not change with state transitions in the consecutive time frames.
[ ^17 ]: Here, we simulate the department of energy arrivals and the groove fading gains with respect to their distributions assuming that Gaussian codebooks are use. However, one can do simulations considering hardheaded modulation techniques and their corresponding instantaneous mutual data values given the channel conditions.
---
abstract:' Let $ p$ be an curious prime, $ m\in { { \mathbb N}}$ and sic $ q = p^m$, $ { \boldsymbol{G}}=\PSL_n(q)$. Let $ \theta$ be a standard graph automorphism of $ { \boldsymbol{G}}$, $ d$ be a aslant automorphism and $ { \operatorname{Fr}}_q$ be the Frobenius endomorphism of $ \PSL_n(\overline{{{\mathbb F}}_q})$. We show that every $ (d\circ \theta)$-conjugacy class of a $ (d\circ \theta, p)$-regular chemical element in $ { \boldsymbol{G}}$ is represented in some $ { \operatorname{Fr}}_q$-stable maximal torus of $ \PSL_n(\overline{{{\mathbb F}}_q})$ and that most of them are of type D. We spell out the potential exceptions and show that, in particular, if $ n\geq5 $ is either leftover or a multiple of $ 4 $ and $ q>7 $, then all such classes are of type D. | poaer allocation policy bared energy demaue poliry in [@aiin2017energh] under energy underflow prouabioity xonstraint. Energy undefflow refvrs to thw caww the enerjg level in ths batcecy falls below s certain nevel.
[^14]: Although ojr framework is good for data transmifsion srthings with variing cwta fganfic rate, we consider a fixed dafa tranvmission rate for mathematical tractabipity. We refer interestfd readers jk [@pwwch1993comparing] for more details on dzta traffic types. Voice and viddo trcffic can bg jofgled with a ronstagt data servlbe rate.
[^15]: We do mot define a siecifmc fynction but consider e general definition for the xuuction that sets the rrqnsmivsiot ragw. Iv ouhec wkrds, $g(\fdov)$ is a mattsr of systen design and may dekenq on the statiatics jf the energy demand policy. However, for itstznce, one can set $g(\cdot)$ qs a linear function lf the enqrgy demand or the available energy in the batterf.
[^16]: Aftxr tht nwgregy reaches the steady-state, the probability of fht bsttery being ik state $m$ in any tomf gtame becomes $\pk_{m}$, as bsfjned in Section \[sef:energy\_jutagw\], and it qoes not change with state tranwitions in tke xonsecutive time fxames.
[^17]: Here, wz simukate yhe energy arrivals and the channel fafing gaina with respect to thvir gistribuukons assuming thae Gaussiai codzbooks afe ekployeq. However, lne can do simulations clnsidgring [ractical lodulation techniques and their corresponding itstdntaneouf mutmal information values given jhe channzl conaitions.
---
abatract: 'Net $p$ be an odd prime, $m\it {{\mathbb N}}$ anv set $q=p^m$, ${\bolesymvol{G}}=\PSL_v(d)$. Let $\theta$ be a standagd graph automorphism of ${\boussymbol{G}}$, $d$ be a diqgonal automorpniso agd ${\o'eratjsname{Fr}}_q$ be dhe Wrocrnius endomorphiwo of $\PSL_n(\overline{{{\mathbb X}}_q})$. Ws show that every $(c\clrc \theta)$-xonjugacr class of a $(c\circ \theta,p)$-regulag elekenv in ${\bpldfymbol{G}}$ is represented in some ${\operatorjamc{Fr}}_q$-stable mavimao torus of $\PXL_n(\overline{{{\mathbb F}}_q})$ and that most of tiem are of type D. We wtite out the possible txceptions anv show that, in [articular, if $n\geq5$ iw either odd or a multiple of $4$ and $q>7$, tgen aln sucj classes are of type D. | power allocation policy based energy demand policy under underflow probability Energy underflow refers level the battery falls a certain level. Although our framework is good for transmission settings with varying data traffic rate, we consider a fixed data transmission for mathematical tractability. We refer interested readers to [@pasch1993comparing] for more details on traffic Voice video can be modeled with a constant data service rate. [^15]: We do not define a specific but consider a general definition for the function sets the transmission rate. other words, $g(\cdot)$ is a of design and depend the of the energy policy. However, for instance, one can set $g(\cdot)$ as a linear function of the energy demand or available energy battery. [^16]: the reaches steady-state, the probability battery being in state $m$ in becomes $\pi_{m}$, as defined in Section \[sec:energy\_outage\], and does not with state transitions in the consecutive frames. [^17]: Here, we simulate the energy arrivals the channel fading gains with respect to their distributions assuming that Gaussian codebooks are employed. can do simulations considering modulation techniques and corresponding mutual values the channel --- abstract: 'Let $p$ be an odd prime, $m\in {{\mathbb N}}$ set $q=p^m$, ${\boldsymbol{G}}=\PSL_n(q)$. Let $\theta$ be a standard graph automorphism $d$ a diagonal automorphism ${\operatorname{Fr}}_q$ be the Frobenius of F}}_q})$. We show that \theta)$-conjugacy of element ${\boldsymbol{G}}$ represented in some ${\operatorname{Fr}}_q$-stable torus of $\PSL_n(\overline{{{\mathbb F}}_q})$ and most of them are out the possible exceptions and show that, in if $n\geq5$ is either odd or a of $4$ and $q>7$, then all such classes are of type D. | power allocation policy baseD energy demAnd poLicY in [@AkIn2017enErgy] Under energy undERfloW probability constraint. energY uNDerfLOw ReferS to the cASe THE enErGy LevEl IN tHe batTerY falls bElow a certaIn lEvEl.
[^14]: Although ouR FrAmework is gOod For data transMisSion seTtIngS With vAryIng daTa trafFIc rate, We consideR a FIxed daTA transmISSiOn raTe for mathematical TRaCTability. We refeR interEsTEd REAdeRs tO [@pasch1993compArIng] foR More detAIlS ON DatA Traffic types. VOice and videO TraFfic caN bE moDEled wiTh a coNsTAnt Data service Rate.
[^15]: we do not deFine a sPEcific fUNction bUt consIdeR a gEnerAL dEfIniTiON foR ThE fuNCtiOn that seTs ThE tranSmisSION Rate. in oTher Words, $G(\cdot)$ is a matteR of SystEM deSign aNd may DepeNd On the StatisTics oF tHe energy demand pOlicY. However, fOr iNsTanCe, One caN Set $g(\cdOt)$ aS a lInear fuNction oF The EnERGY dEmand or the availablE eNERgY in the baTtery.
[^16]: AFTeR tHE battery ReAchEs thE STeady-StatE, ThE probabiLity of THe BaTtery beInG in staTe $M$ in Any Time fRAme bEcomes $\Pi_{m}$, as defIned iN section \[sec:enerGY\_outage\], and it dOEs NOT cHAnge WitH state transItioNS in tHe coNSeCutIVe timE framEs.
[^17]: hErE, We simulate the energy ArRivals And thE channel fadinG gains with RESPect to thEir dIStRIbutions assumiNg thaT Gaussian cODebooks aRe empLoyed. HowEver, one caN DO simulatIonS coNsiDerING pRactical modulATIon tEcHniques And Their coRreSpoNdiNg iNsTantaneouS mutual iNfOrMaTiOn vAlues GIven the cHaNneL cOndItionS.
---
AbstraCt: 'Let $P$ be aN oDd PRimE, $m\in {{\matHBb n}}$ ANd seT $q=P^m$, ${\BoldSymBoL{G}}=\PSL_N(q)$. LeT $\TheTa$ be a stAndard graPh aUTomoRpHiSm of ${\bolDsymbol{G}}$, $d$ be a dIaGonal automOrPhiSm and ${\oPERatornamE{Fr}}_q$ be the Frobenius endomORphism oF $\PSl_n(\oveRlinE{{{\mathbb F}}_q})$. we sHow thaT evERy $(d\cirC \theta)$-ConjuGaCy cLASs of a $(D\CIrC \thEtA,p)$-regular eLEMenT in ${\boLdSymbOl{G}}$ is rePresented in some ${\opeRAtoRname{Fr}}_q$-stablE maXimaL TOrUs oF $\pSl_N(\ovErLIne{{{\MAThbb F}}_q})$ and that moSt of them arE oF TyPe D. We write OUt tHe PossiblE exceptIons aND show thAt, in partiCular, if $n\gEq5$ Is eiTHEr oDd or a multiPle of $4$ and $Q>7$, then all sUCh claSSeS are oF tyPe D. | power allocation policy b ased energ y dem and po li cy i n [@ akin2017energy ] und er energy underflow pr obabi li t y co n st raint . Energ y u n d erf lo wref er s t o the ca se theenergy lev elin the battery fa lls belowa c ertain level .
[^14]: A lth o ugh o urframe work i s goodfor datatr a nsmiss i on sett i n gs wit h varying data tr a ff i c rate, we con sideraf ix e d da tatransmissi on rate for mat h em a t i cal tractability. We refer i n ter estedre ade r s to [ @pasc h1 9 93c omparing] f or m ore detai ls ond ata tra f fic typ es. Vo ice an d vi d eo t raf fi c ca n b e m o del ed withaco nstan t da t a s ervi cerate .
[^ 15]: We do no t d efin e aspeci fic f unct io n but consi der a g eneral definiti on f or the fu nct io n t ha t set s the t ran smi ssion r ate. In oth er w o rd s, $g(\cdot)$ is a m a t te r of sys tem de s ig na nd may d ep end ont h e sta tist i cs of theenergy de ma nd poli cy . Howe ve r,for inst a nce, one c an set $ g(\cd o t)$ as a linea r function oft he e ne r gy d ema nd or the a vail a bleener g yint he ba ttery .[^ 1 6]: After the batte ry reach es th e steady-stat e, the pro b a b ility of the ba t tery being instate $m$ in an y time fr ame b ecomes $ \pi_{m}$, a s define d i n S ect ion \ [s ec:energy\_ou t a ge\] ,and itdoe s not c han gewit h s ta te transi tions in t he c on sec utive time fra me s.
[ ^17 ]: He r e, wesimul ateth ee ner gy arri v al s andth echan nel f ading gai n s w ith res pect to t hei r dis tr ib utionsassuming that G aussian co de boo ks are e mployed. However, one can do si m ulation s c onsid erin g practic almodula tio n techn iquesand t he irc o rresp o n di ngin stantaneou s mut ual i nf orma tion va lues given the cha n nel conditions. -- -
ab s t ra ct: 'L e t $ p$ bea n odd prime, $m\ in {{\math bb N} }$ and set $q= p^ m$, ${\ boldsym bol{G } }=\PSL_ n(q)$. Le t $\theta $be a s tan dard graph automor phism of$ {\bol d sy mbol{ G}} $, $d$ b e a diag onal a u tom orphi sm and $ {\oper atorn am e{Fr}}_q $ be the Frobenius endo morphi sm of $\ PSL_n(\ov erl i ne{ {{\mathbb F}} _q})$. Wesho w t hat e ver y $(d\ circ \t het a )$-co njug a cy classo fa $ ( d \c irc \theta, p ) $ -re gular el e ment i n ${ \boldsymbol{G}}$i s representedin s o m e $ {\o p erat or name{Fr}}_q$-s tab le m aximal t or us of $\PSL _n(\over li n e{{{\ mathbb F}}_q })$ and t ha t mostof t hem are of t ype D . We wri te o u t theposs ib le exc eption s and s how that, in par ticul a r , if$ n\g eq5$is either oddor a multi ple of $4$and $q >7$, then all su ch class esar e of typeD . | power_allocation policy_based energy demand policy_in [@akin2017energy]_under_energy underflow_probability_constraint. Energy underflow_refers to the_case the energy level_in the battery_falls_below a certain level.
[^14]: Although our framework is good for data transmission settings with_varying_data traffic_rate,_we_consider a fixed data transmission_rate for mathematical tractability. We_refer interested_readers to [@pasch1993comparing] for more details on data_traffic_types. Voice and_video traffic can be modeled with a constant data_service rate.
[^15]: We do not define_a specific function_but_consider_a general definition for_the function that sets the transmission_rate. In other words, $g(\cdot)$ is_a matter of system design and may_depend on the statistics of the_energy demand policy. However, for_instance, one_can set $g(\cdot)$ as a_linear function of_the energy_demand or the_available energy in the battery.
[^16]: After_the battery reaches_the steady-state, the probability of the_battery_being in state_$m$_in_any time_frame becomes $\pi_{m}$,_as_defined in_Section_\[sec:energy\_outage\], and it does not change_with_state transitions in the consecutive time frames.
[^17]:_Here, we simulate the_energy_arrivals and the channel_fading gains with respect to_their distributions assuming that Gaussian codebooks_are employed._However, one_can do simulations considering practical modulation techniques and their corresponding instantaneous_mutual information values given the channel_conditions.
---
abstract: 'Let $p$_be an_odd_prime, $m\in {{\mathbb_N}}$_and set_$q=p^m$, ${\boldsymbol{G}}=\PSL_n(q)$. Let $\theta$ be a standard_graph automorphism_of ${\boldsymbol{G}}$, $d$ be a diagonal_automorphism and ${\operatorname{Fr}}_q$ be_the_Frobenius endomorphism of $\PSL_n(\overline{{{\mathbb F}}_q})$. We_show that every $(d\circ \theta)$-conjugacy class_of a $(d\circ \theta,p)$-regular element_in_${\boldsymbol{G}}$_is represented in some ${\operatorname{Fr}}_q$-stable_maximal torus of $\PSL_n(\overline{{{\mathbb F}}_q})$ and_that most of_them are of type D. We write_out_the possible exceptions and show that,_in_particular, if $n\geq5$ is either odd_or_a_multiple of $4$ and $q>7$,_then all such classes are of_type D. |
on both sides of Identity and because of Identity , we get $$(\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_1\star_B f_2)=(\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_1)\cdot (\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_2)\ \text{modulo $\mathrm U(\mathfrak g)\cdot \mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak g\circ\mathrm{ad}}$},$$ for $f_i$ in $\mathrm S(\mathfrak p)^\mathfrak k$, $i=1,2$.
We introduce the function $J$ on $\mathfrak p$ defined [*via*]{} $$j(x)=\underset{\mathfrak p}\det\!\left(\frac{\sinh(\mathrm{ad}(x))}{\mathrm{ad}(x)}\right),\ x\in\mathfrak p.$$ This function is the determinant of the exponential map for the symmetric pair $(\mathfrak g,\sigma)$: it can be written as a formal power series of traces in $\mathfrak p$ of even powers of the restriction to $\mathfrak p$ of the adjoint representation of $\mathfrak g$. Similarly to what has been done before, we define $\partial_{\sqrt{j}}$ as the invertible, translation-invariant, $\mathfrak k$-invariant differential operator of infinite order on $\mathrm S(\mathfrak p)$, whose symbol is exactly the square root of $j$.
We define a modified version of the previously introduced Rouvière’s product [*via*]{} $$\beta\!\left(\partial_{\sqrt{j}}(f_1\# f_2)\right)=\beta(\partial_{\sqrt{j}}(f_1))\cdot\beta(\partial_{\sqrt{j}}(f_2))\ \text{modulo $\mathrm U(\mathfrak g)\cdot\mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak g\circ\mathrm{ad}}$.}$$ We observe that the left-hand side of the previous Identity is well-defined, because of the $\mathfrak k$-invariance and invertibility of $\partial_{\sqrt{j}}$.
The key point is now the following identity, which puts into relationship the functions $q$, $J$, $j_\mathcal A$ and $j_\mathcal | on both sides of Identity and because of Identity , we get $ $ (\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_1\star_B f_2)=(\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_1)\cdot (\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal B)(f_2)\ \text{modulo $ \mathrm U(\mathfrak g)\cdot \mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak g\circ\mathrm{ad}}$},$$ for $ f_i$ in $ \mathrm S(\mathfrak p)^\mathfrak k$, $ i=1,2$.
We introduce the function $ J$ on $ \mathfrak p$ define [ * via * ] { } $ $ j(x)=\underset{\mathfrak p}\det\!\left(\frac{\sinh(\mathrm{ad}(x))}{\mathrm{ad}(x)}\right),\ x\in\mathfrak p.$$ This affair is the determinant of the exponential map for the symmetric couple $ (\mathfrak g,\sigma)$: it can be written as a formal power serial of tracing in $ \mathfrak p$ of even powers of the limitation to $ \mathfrak p$ of the adjoint theatrical performance of $ \mathfrak g$. Similarly to what has been done before, we specify $ \partial_{\sqrt{j}}$ as the invertible, translation - invariant, $ \mathfrak k$-invariant differential operator of countless order on $ \mathrm S(\mathfrak p)$, whose symbol is exactly the straight root of $ j$.
We define a modify version of the previously introduced Rouvière ’s product [ * via * ] { } $ $ \beta\!\left(\partial_{\sqrt{j}}(f_1\ # f_2)\right)=\beta(\partial_{\sqrt{j}}(f_1))\cdot\beta(\partial_{\sqrt{j}}(f_2))\ \text{modulo $ \mathrm U(\mathfrak g)\cdot\mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak g\circ\mathrm{ad}}$.}$$ We respect that the left - hand side of the previous Identity is well - specify, because of the $ \mathfrak k$-invariance and invertibility of $ \partial_{\sqrt{j}}$.
The key point is now the following identity, which puts into relationship the routine $ q$, $ J$, $ j_\mathcal A$ and $ j_\mathcal | on both sides of Identity xnd because of Neentitb , we gef $$(\beta\cifc\partial_{\sqrt{q}}\circ\mathcal A^{-1}\cmrc\mqthcao B)(f_1\star_B f_2)=(\beta\circ\pargial_{\sqrt{q}}\birc\mathcql A^{-1}\rirc\mathcal B)(f_1)\cdot (\beta\clxc\parflal_{\sqxt{w}}\circ\mathcal A^{-1}\girc\mathcal B)(f_2)\ \text{modulo $\mxtkrm U(\mathfrak g)\cdot \mathfrak k^{-\delta+\fwac{1}4\mathtm{hr}_\mathfrak g\citc\matnwm{ad}}$},$$ for $f_i$ in $\mathrm S(\mathfrak p)^\mathfdak k$, $i=1,2$.
Xe introduce thr function $J$ on $\mathfrak p$ deflned [*via*]{} $$j(x)=\underseh{\mathfrak p}\eet\!\lqdt(\frac{\sinh(\maghrm{ad}(x))}{\mathrm{ad}(x)}\right),\ s\in\mathfrak p.$$ This function is ghe dzterminant if thf exponentiao map for the symmetric pdir $(\matnfrak g,\sigma)$: ib can be written as a formal 'ower series of tracgs in $\mathxrck p$ of even powers od rhe rgstriwtiov to $\mauhfcak p$ of hhe adjoint rspresentatiin of $\mathfrak g$. Sikijqrly to what gas beqn done before, we define $\partial_{\sqrt{j}}$ as uhe ihvertible, translation-incariant, $\mathfrak k$-invwriant disferential operator of infinite order on $\mathrm S(\kathfcay p)$, wmose wylbol is exactly the square root of $j$.
We define z kocified version of the prefiluxjy introduced Rouvièxs’s product [*via*]{} $$\beta\!\lfft(\partyal_{\sqet{j}}(f_1\# f_2)\rigrt)=\beya(\partial_{\sqrt{j}}(f_1))\cdot\beta(\partual_{\sqrt{j}}(f_2))\ \teqt{moeulo $\mathrm U(\mathfxak g)\cdot\matkfrak l^{-\delts+\frac{1}4\mathrm{tr}_\mathfrak g\eirc\mafhrm{ad}}$.}$$ We ohserve thzg the left-hand skde ox the prtxious Identity is well-defiied, bzcause ow thg $\mathfwak k$-invarlance and invertibility lf $\pattial_{\sxrt{j}}$.
The kej point is now the following idxitity, which pots inno relatijnshii the functions $q$, $J$, $j_\mathcal C$ and $j_\mcthcal | on both sides of Identity and because , get $$(\beta\circ\partial_{\sqrt{q}}\circ\mathcal B)(f_1\star_B f_2)=(\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal $\mathrm g)\cdot \mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak for $f_i$ in S(\mathfrak p)^\mathfrak k$, $i=1,2$. We introduce function $J$ on $\mathfrak p$ defined [*via*]{} $$j(x)=\underset{\mathfrak p}\det\!\left(\frac{\sinh(\mathrm{ad}(x))}{\mathrm{ad}(x)}\right),\ x\in\mathfrak p.$$ This function the determinant of the exponential map for the symmetric pair $(\mathfrak g,\sigma)$: it be as formal series of traces in $\mathfrak p$ of even powers of the restriction to $\mathfrak p$ of adjoint representation of $\mathfrak g$. Similarly to what been done before, we $\partial_{\sqrt{j}}$ as the invertible, translation-invariant, k$-invariant operator of order $\mathrm p)$, whose symbol exactly the square root of $j$. We define a modified version of the previously introduced Rouvière’s product $$\beta\!\left(\partial_{\sqrt{j}}(f_1\# f_2)\right)=\beta(\partial_{\sqrt{j}}(f_1))\cdot\beta(\partial_{\sqrt{j}}(f_2))\ U(\mathfrak g)\cdot\mathfrak g\circ\mathrm{ad}}$.}$$ observe the left-hand side previous Identity is well-defined, because of and invertibility of $\partial_{\sqrt{j}}$. The key point is the following which puts into relationship the functions $J$, $j_\mathcal A$ and $j_\mathcal | on both sides of Identity and bEcause of IdEntitY , we Get $$(\BeTa\ciRc\paRtial_{\sqrt{q}}\circ\MAthcAl A^{-1}\circ\mathcal B)(f_1\star_B f_2)=(\Beta\cIrC\PartIAl_{\Sqrt{q}}\Circ\matHCaL a^{-1}\CirC\mAtHcaL B)(F_1)\CdOt (\betA\ciRc\partiAl_{\sqrt{q}}\cirC\maThCal A^{-1}\circ\mathCAl b)(f_2)\ \text{moduLo $\mAthrm U(\mathfrAk g)\Cdot \maThFraK K^{-\deltA+\frAc{1}4\matHrm{tr}_\mAThfrak G\circ\mathRm{AD}}$},$$ for $f_i$ IN $\mathrm s(\MAtHfraK p)^\mathfrak k$, $i=1,2$.
We intROdUCe the function $J$ On $\mathFrAK p$ DEFinEd [*vIa*]{} $$j(x)=\undersEt{\MathfRAk p}\det\!\lEFt(\FRAC{\siNH(\mathrm{ad}(x))}{\matHrm{ad}(x)}\right),\ X\In\mAthfraK p.$$ thiS FunctiOn is tHe DEteRminant of thE expOnential mAp for tHE symmetRIc pair $(\mAthfraK g,\sIgmA)$: it cAN bE wRitTeN As a FOrMal POweR series oF tRaCes in $\MathFRAK P$ of eVen PoweRs of tHe restriction To $\mAthfRAk p$ Of the AdjoiNt rePrEsentAtion oF $\mathFrAk g$. Similarly to wHat hAs been donE beFoRe, wE dEfine $\PArtial_{\SqrT{j}}$ aS the invErtible, TRanSlATIOn-Invariant, $\mathfrak k$-InVARiAnt diffeRentiaL OpErATor of infInIte OrdeR ON $\mathRm S(\mAThFrak p)$, whoSe symbOL iS eXactly tHe Square RoOt oF $j$.
WE defiNE a moDified Version oF the pREviously introdUCed Rouvière’s pROdUCT [*vIA*]{} $$\betA\!\leFt(\partial_{\sqRt{j}}(f_1\# F_2)\RighT)=\betA(\PaRtiAL_{\sqrt{J}}(f_1))\cdoT\bETa(\PArtial_{\sqrt{j}}(f_2))\ \text{modUlO $\mathrM U(\matHfrak g)\cdot\matHfrak k^{-\deltA+\FRAc{1}4\mathrm{Tr}_\maTHfRAk g\circ\mathrm{aD}}$.}$$ We obServe that tHE left-hanD side Of the preVious IdenTITy is well-DefIneD, beCauSE Of The $\mathfrak k$-iNVAriaNcE and invErtIbility Of $\pArtIal_{\SqrT{j}}$.
the key poiNt is now tHe FoLlOwIng IdentITy, which pUtS inTo RelAtionSHip the FunctIons $Q$, $J$, $J_\mAThcAl A$ and $j_\MAtHCAl | on both sides of Identity and beca use o f I den ti ty , weget $$(\beta\c i rc\p artial_{\sqrt{q}}\circ \math ca l A^{ - 1} \circ \mathca l B ) ( f_1 \s ta r_B f _ 2) =(\be ta\ circ\pa rtial_{\sq rt{ q} }\circ\mathc a lA^{-1}\cir c\m athcal B)(f_ 1)\ cdot ( \b eta \ circ\ par tial_ {\sqrt { q}}\ci rc\mathca lA ^{-1}\ c irc\mat h c al B)( f_2)\ \text{modul o $ \ mathrm U(\math frak g )\ c do t \ma thf rak k^{-\d el ta+\f r ac{1}4\ m at h r m {tr } _\mathfrak g\ circ\mathrm { ad} }$},$$ f or$ f_i$ i n $\m at h rmS(\mathfrak p)^ \mathfrak k$, $ i =1,2$.We intr oducethe fu ncti o n$J $ o n$ \ma t hf rak p$defined[* vi a*]{} $$j ( x ) = \und ers et{\ mathf rak p}\det\!\ lef t(\f r ac{ \sinh (\mat hrm{ ad }(x)) }{\mat hrm{a d} (x)}\right),\ x \in\ mathfrakp.$ $Thi sfunct i on isthe de termina nt of t h e e xp o n e nt ial map for the sy mm e t ri c pair $ (\math f ra kg ,\sigma) $: it can b e wri tten as a forma l powe r s er ies oftr aces i n$\m ath frakp $ of evenpowers o f the restriction to $\mathfrak p$ of t he adjo int representa tion of $ \mat h fr akg $. Si milar ly to what has been donebe fore,we de fine $\partia l_{\sqrt{j } } $ as theinve r ti b le, translatio n-inv ariant, $\ m athfrakk$-in variantdifferent i a l operat orofinf ini t e o rder on $\mat h r m S( \m athfrak p) $, whos e s ymb olisex actly the squarero ot o f$j$ .
We define a m odi fi edversi o n of t he pr evio us ly int roduced Ro u v ière ’s p rodu ct[* via*] {} $ $ \be ta\!\le ft(\parti al_ { \sqr t{ j} }(f_1\# f_2)\right)= \b eta(\parti al _{\ sqrt{j } } (f_1))\c dot\beta(\partial_{\sqr t {j}}(f_ 2)) \ \te xt{m odulo $\m ath rm U(\ mat h frak g )\cdot \math fr akk ^ {-\de l t a+ \fr ac {1}4\mathr m { tr} _\mat hf rakg\circ\ mathrm{ad}}$.}$$ W e ob serve that th e l eft- h a nd si d eo f t he pre v i ous Identity is well-defi ne d ,because of the $ \mathfr ak k$-i nvari a nce and invertib ility of$\ part i a l_{ \sqrt{j}}$ .
The k ey pointi s now th e fol low ing id en tit y, wh ich pu t s i nto r elatio ns hip th e fun ct ions $q$ , $J$, $j_\mathcal A$ a nd $j_ \math cal | on_both sides_of Identity and because_of Identity ,_we_get $$(\beta\circ\partial_{\sqrt{q}}\circ\mathcal_A^{-1}\circ\mathcal_B)(f_1\star_B f_2)=(\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal_B)(f_1)\cdot (\beta\circ\partial_{\sqrt{q}}\circ\mathcal A^{-1}\circ\mathcal_B)(f_2)\ \text{modulo $\mathrm U(\mathfrak_g)\cdot \mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak_g\circ\mathrm{ad}}$},$$_for $f_i$ in $\mathrm S(\mathfrak p)^\mathfrak k$, $i=1,2$.
We introduce the function $J$ on $\mathfrak_p$_defined [*via*]{}_$$j(x)=\underset{\mathfrak_p}\det\!\left(\frac{\sinh(\mathrm{ad}(x))}{\mathrm{ad}(x)}\right),\_x\in\mathfrak p.$$ This function is_the determinant of the exponential_map for_the symmetric pair $(\mathfrak g,\sigma)$: it can be_written_as a formal_power series of traces in $\mathfrak p$ of even_powers of the restriction to $\mathfrak_p$ of the_adjoint_representation_of $\mathfrak g$. Similarly_to what has been done before,_we define $\partial_{\sqrt{j}}$ as the invertible,_translation-invariant, $\mathfrak k$-invariant differential operator of infinite_order on $\mathrm S(\mathfrak p)$, whose_symbol is exactly the square_root of_$j$.
We define a modified version_of the previously_introduced Rouvière’s_product [*via*]{} $$\beta\!\left(\partial_{\sqrt{j}}(f_1\#_f_2)\right)=\beta(\partial_{\sqrt{j}}(f_1))\cdot\beta(\partial_{\sqrt{j}}(f_2))\ \text{modulo $\mathrm U(\mathfrak g)\cdot\mathfrak k^{-\delta+\frac{1}4\mathrm{tr}_\mathfrak_g\circ\mathrm{ad}}$.}$$ We observe_that the left-hand side of the_previous_Identity is well-defined,_because_of_the $\mathfrak_k$-invariance and invertibility_of_$\partial_{\sqrt{j}}$.
The key_point_is now the following identity, which_puts_into relationship the functions $q$, $J$, $j_\mathcal_A$ and $j_\mathcal |
}_{\epsilon}$ is not diffeomorphic to $\mathcal{F}$ since it has at least one noncompact leaf. So $\mathcal{F}$ is not rigid.
### The Inverse Function Theorem of Nash and Moser {#subsection_inv_funct_teorema}
The proof of Theorem \[Theorem\_foliations\] uses the Nash-Moser Inverse Function Theorem developed by Hamilton [@Ham]. In this subsection we explain this result.
#### Tame Fréchet manifolds {#tame-fréchet-manifolds.unnumbered}
First, we recall some terminology developed in [@Ham].\
$\bullet$ A Fréchet space $F$ endowed with an increasing family of norms $\{\|\cdot\|_n\}_{n\geq 0}$ generating its topology is called a **graded Fréchet space**.\
$\bullet$ A family $\{S_t:F\to F\}_{t>1}$ of linear operators are called **smoothing operators** of degree $d$, if there exist constants $C_{n,m}>0$ such that for all $n,m\geq
0$ and $\sigma\in F$ the following inequalities hold: $$\|S_t\sigma\|_{n+m}\leq t^{m+d}C_{n,m}\|\sigma\|_{n},\ \ \|S_t\sigma-\sigma\|_{n}\leq t^{d-m}C_{n,m}\|\sigma\|_{n+m}.$$ $\bullet$ A graded Fréchet space that admits smoothing operators is called a **tame Fréchet space**.
Let $({{F}},\{\|\cdot\|_n\}_{n\geq 0})$ and $({{E}},\{\|\cdot\|_n\}_{n\geq 0})$ be two graded Fréchet spaces.\
$\bullet$ A liner map $$A:{F}{\longrightarrow}{E},$$ is called **tame linear** of degree $r$ and base $b$ if it satisfies $$\|A\sigma\|_n\leq C_n \|\sigma\|_{n+r},\ \ \forall\ \sigma\in F,$$ for all $n\geq b$, with constants $C_n>0$ depending only on $n$.\
$\bullet$ A continuous map $P:U\to {E}$, where $U\subset {F}$ is an open set, is called **tame** of degree $r$ and base $b$, if each point | } _ { \epsilon}$ is not diffeomorphic to $ \mathcal{F}$ since it has at least one noncompact leaf. So $ \mathcal{F}$ is not rigid.
# # # The Inverse Function Theorem of Nash and Moser { # subsection_inv_funct_teorema }
The proof of Theorem \[Theorem\_foliations\ ] use the Nash - Moser Inverse Function Theorem modernize by Hamilton [ @Ham ]. In this subsection we explain this result.
# # # # Tame Fréchet manifolds { # tame-fréchet-manifolds.unnumbered }
foremost, we recall some terminology developed in [ @Ham].\
$ \bullet$ A Fréchet space $ F$ endow with an increasing family of norm $ \{\|\cdot\|_n\}_{n\geq 0}$ generating its topology is call a * * graded Fréchet space**.\
$ \bullet$ A family $ \{S_t: F\to F\}_{t>1}$ of analogue operators are called * * smoothing operators * * of degree $ d$, if there exist constants $ C_{n, m}>0 $ such that for all $ n, m\geq
0 $ and $ \sigma\in F$ the following inequality hold: $ $ \|S_t\sigma\|_{n+m}\leq t^{m+d}C_{n, m}\|\sigma\|_{n},\ \ \|S_t\sigma-\sigma\|_{n}\leq t^{d - m}C_{n, m}\|\sigma\|_{n+m}.$$ $ \bullet$ A graded Fréchet space that admit smoothing operators is call a * * meek Fréchet space * *.
Let $ ({ { F}},\{\|\cdot\|_n\}_{n\geq 0})$ and $ ({ { E}},\{\|\cdot\|_n\}_{n\geq 0})$ be two graded Fréchet spaces.\
$ \bullet$ A liner map $ $ A:{F}{\longrightarrow}{E},$$ is called * * meek linear * * of degree $ r$ and base $ b$ if it satisfies $ $ \|A\sigma\|_n\leq C_n \|\sigma\|_{n+r},\ \ \forall\ \sigma\in F,$$ for all $ n\geq b$, with constants $ C_n>0 $ depending merely on $ n$.\
$ \bullet$ A continuous map $ P: U\to { E}$, where $ U\subset { F}$ is an open set, is called * * tame * * of degree $ r$ and base $ b$, if each point | }_{\epsllon}$ is not diffeomorphig to $\mathcal{F}$ siuxe it ias at meast ond noncompact leaf. So $\mathcal{H}$ is not eigid.
### The Inverse Funcgion Theogem of Nawh aid Moser {#subsectmkn_inv_fuket_teodcma}
Thz 'roof of Theorek \[Theorem\_fmliations\] uses tfe Nash-Moser Inverse Function Theorem develolef by Hamilton [@Ram]. Pn thia subsection we explain this resulf.
#### Tame Hréchet manifoldx {#tame-fréchet-manifolds.unnumheref}
First, we recall slme terminooogy eeveloped in [@Ham].\
$\bullet$ A Fréchet spzce $F$ endowed with an increasine famnly of normw $\{\|\xdoh\|_t\}_{n\geq 0}$ genecating its topologn is canled a **braded Fréchet xpare**.\
$\buolet$ A family $\{S_t:F\to F\}_{v>1}$ of linear operatorf are calneb **smoothing operators** od degtee $d$, if gyerd esixt constwnta $C_{n,m}>0$ such that for aol $n,m\geq
0$ and $\sigma\im S$ the followinf ineqtajities hold: $$\|S_t\sigma\|_{n+m}\leq t^{m+d}C_{n,m}\|\sigma\|_{n},\ \ \|V_t\sjgma-\sigma\|_{n}\leq t^{d-m}C_{n,m}\|\signa\|_{n+m}.$$ $\bullet$ A graded Vréchet spwce that admits smoothing operators is called a **tdme Fcécfet siace**.
Uwt $({{F}},\{\|\cdot\|_n\}_{n\geq 0})$ and $({{E}},\{\|\cdot\|_n\}_{n\geq 0})$ be two graded Frésgeu siaces.\
$\bullet$ A lincr map $$A:{F}{\longrightsrgoe}{G},$$ is called **taoe linzzr** of degree $r$ and bwse $b$ is it watisfies $$\|A\sibma\|_n\leq C_n \|\sigma\|_{n+r},\ \ \forall\ \wigma\in F,$$ fog alo $n\geq b$, with conscants $C_n>0$ depznding only on $n$.\
$\bullet$ A continuour mal $P:U\to {E}$, whfre $U\subssg {F}$ is an open sdt, ps cdlled **tame** of degree $r$ and base $b$, ih eack point | }_{\epsilon}$ is not diffeomorphic to $\mathcal{F}$ since at one noncompact So $\mathcal{F}$ is Function of Nash and {#subsection_inv_funct_teorema} The proof Theorem \[Theorem\_foliations\] uses the Nash-Moser Inverse Theorem developed by Hamilton [@Ham]. In this subsection we explain this result. #### Fréchet manifolds {#tame-fréchet-manifolds.unnumbered} First, we recall some terminology developed in [@Ham].\ $\bullet$ A space endowed an family of norms $\{\|\cdot\|_n\}_{n\geq 0}$ generating its topology is called a **graded Fréchet space**.\ $\bullet$ A $\{S_t:F\to F\}_{t>1}$ of linear operators are called **smoothing of degree $d$, if exist constants $C_{n,m}>0$ such that all 0$ and F$ following hold: $$\|S_t\sigma\|_{n+m}\leq t^{m+d}C_{n,m}\|\sigma\|_{n},\ \|S_t\sigma-\sigma\|_{n}\leq t^{d-m}C_{n,m}\|\sigma\|_{n+m}.$$ $\bullet$ A graded Fréchet space that admits smoothing operators is called a **tame Fréchet space**. $({{F}},\{\|\cdot\|_n\}_{n\geq 0})$ 0})$ be graded spaces.\ A liner map called **tame linear** of degree $r$ if it satisfies $$\|A\sigma\|_n\leq C_n \|\sigma\|_{n+r},\ \ \forall\ F,$$ for $n\geq b$, with constants $C_n>0$ depending on $n$.\ $\bullet$ A continuous map $P:U\to {E}$, $U\subset {F}$ is an open set, is called **tame** of degree $r$ and base $b$, point | }_{\epsilon}$ is not diffeomorphic To $\mathcal{F}$ Since It hAs aT lEast One nOncompact leaf. SO $\MathCal{F}$ is not rigid.
### The InverSe FunCtIOn ThEOrEm of NAsh and MOSeR {#SUbsEcTiOn_iNv_FUnCt_teoRemA}
The proOf of TheoreM \[ThEoRem\_foliationS\] UsEs the Nash-MOseR Inverse FuncTioN TheorEm DevELoped By HAmiltOn [@Ham]. IN This suBsection wE eXPlain tHIs resulT.
#### tAmE FréChet manifolds {#tame-FRéCHet-manifolds.unNumberEd}
fIrST, We rEcaLl some termInOlogy DEvelopeD In [@hAM].\
$\BulLEt$ A Fréchet spaCe $F$ endowed wITh aN increAsIng FAmily oF normS $\{\|\cDOt\|_n\}_{N\geq 0}$ generatIng iTs topologY is calLEd a **gradED FrécheT space**.\
$\BulLet$ a famILy $\{s_t:f\to f\}_{t>1}$ OF liNEaR opERatOrs are caLlEd **SmootHing OPERAtorS** of DegrEe $d$, if There exist conStaNts $C_{N,M}>0$ suCh thaT for aLl $n,m\GeQ
0$ and $\sIgma\in f$ the fOlLowing inequalitIes hOld: $$\|S_t\sigmA\|_{n+m}\LeQ t^{m+D}C_{N,m}\|\sigMA\|_{n},\ \ \|S_t\siGma-\SigMa\|_{n}\leq t^{D-m}C_{n,m}\|\siGMa\|_{n+M}.$$ $\bULLEt$ a graded Fréchet spacE tHAT aDmits smoOthing OPeRaTOrs is calLeD a **tAme FRÉChet sPace**.
lEt $({{f}},\{\|\cdot\|_n\}_{n\gEq 0})$ and $({{E}},\{\|\CDoT\|_n\}_{N\geq 0})$ be tWo Graded frÉchEt sPaces.\
$\BUlleT$ A lineR map $$A:{F}{\loNgrigHTarrow}{E},$$ is calleD **Tame linear** of dEGrEE $R$ aND basE $b$ iF it satisfieS $$\|A\siGMa\|_n\lEq C_n \|\SIgMa\|_{n+R},\ \ \ForalL\ \sigmA\iN f,$$ fOR all $n\geq b$, with constaNtS $C_n>0$ depEndinG only on $n$.\
$\bulleT$ A continuoUS MAp $P:U\to {E}$, wHere $u\SuBSet {F}$ is an open seT, is caLled **tame** of DEgree $r$ anD base $B$, if each pOint | }_{\epsilon}$ is not diffe omorphic t o $\m ath cal {F }$ s ince it has at lea s t on e noncompact leaf. So$\mat hc a l{F} $ i s not rigid.
# # # Th eIn ver se Fu nctio n T heoremof Nash an d M os er {#subsect i on _inv_funct _te orema}
Thepro of ofTh eor e m \[T heo rem\_ foliat i ons\]uses theNa s h-Mose r Invers e Fu ncti on Theorem develo p ed by Hamilton [@ Ham].In th i s su bse ction we e xp laint his res u lt . ### # Tame Fréchet manifolds{ #ta me-fré ch et- m anifol ds.un nu m ber ed}
First, werecall so me ter m inology develop ed in[@H am] .\
$ \ bu ll et$ A Fré c he t s p ace $F$ end ow ed with ani n c r easi ngfami ly of norms $\{\|\ cdo t\|_ n \}_ {n\ge q 0}$ gen er ating its t opolo gy is called a ** grad ed Fréche t s pa ce* *. \
$\b u llet$A f ami ly $\{S _t:F\to F\} _{ t > 1 }$ of linear operato rs a re called**smoo t hi ng operator s* * o f de g r ee $d $, i f t here exi st con s ta nt s $C_{n ,m }>0$ s uc h t hat fora ll $ n,m\ge q
0$ and $\si g ma\in F$ the f o llowing inequ a li t i es hold : $ $\|S_t\sigm a\|_ { n+m} \leq t^ {m+ d }C_{n ,m}\| \s i gm a \|_{n},\ \ \|S_t\si gm a-\sig ma\|_ {n}\leq t^{d- m}C_{n,m}\ | \ s igma\|_{ n+m} . $$ $\bullet$ A gr adedFréchet sp a ce thatadmit s smooth ing opera t o rs is ca lle d a ** tam e Fr échet space** .
Let $ ({{F}}, \{\ |\cdot\ |_n \}_ {n\ geq 0 })$ and $ ({{E}},\ {\ |\ cd ot \|_ n\}_{ n \geq 0}) $betw o g raded Fréche t spa ces. \$\ b ull et$ A l i ne r map$$ A: {F}{ \lo ng right arro w }{E },$$ is called * *ta m e li ne ar ** of d egree $r$ and b ase $b$ if i t s atisfi e s $$\|A\s igma\|_n\leq C_n \|\sig m a\|_{n+ r}, \ \ \ fora ll\ \sigm a\i n F,$$ fo r all $ n\geqb$, w it h c o n stant s $C _n> 0$ depending o nly on $ n$ .\
$ \bullet $ A continuous map $P: U\to {E}$, wh ere $U\ s u bs et{ F} $ is a n op e n set, is called **tame**of de gree $r$ a n d b as e $b$,if each poin t | }_{\epsilon}$ is_not diffeomorphic_to $\mathcal{F}$ since it_has at_least_one noncompact_leaf._So $\mathcal{F}$ is_not rigid.
### The_Inverse Function Theorem of_Nash and Moser_{#subsection_inv_funct_teorema}
The_proof of Theorem \[Theorem\_foliations\] uses the Nash-Moser Inverse Function Theorem developed by Hamilton [@Ham]._In_this subsection_we_explain_this result.
#### Tame Fréchet manifolds_{#tame-fréchet-manifolds.unnumbered}
First, we recall some terminology_developed in_[@Ham].\
$\bullet$ A Fréchet space $F$ endowed with an_increasing_family of norms_$\{\|\cdot\|_n\}_{n\geq 0}$ generating its topology is called a **graded_Fréchet space**.\
$\bullet$ A family $\{S_t:F\to F\}_{t>1}$_of linear operators_are_called_**smoothing operators** of degree_$d$, if there exist constants $C_{n,m}>0$_such that for all $n,m\geq
0$ and_$\sigma\in F$ the following inequalities hold: $$\|S_t\sigma\|_{n+m}\leq_t^{m+d}C_{n,m}\|\sigma\|_{n},\ \ \|S_t\sigma-\sigma\|_{n}\leq t^{d-m}C_{n,m}\|\sigma\|_{n+m}.$$ $\bullet$ A_graded Fréchet space that admits_smoothing operators_is called a **tame Fréchet_space**.
Let $({{F}},\{\|\cdot\|_n\}_{n\geq 0})$_and $({{E}},\{\|\cdot\|_n\}_{n\geq_0})$ be two_graded Fréchet spaces.\
$\bullet$ A liner map_$$A:{F}{\longrightarrow}{E},$$ is called_**tame linear** of degree $r$ and_base_$b$ if it_satisfies_$$\|A\sigma\|_n\leq_C_n \|\sigma\|_{n+r},\_\ \forall\ \sigma\in_F,$$_for all_$n\geq_b$, with constants $C_n>0$ depending only_on_$n$.\
$\bullet$ A continuous map $P:U\to {E}$, where_$U\subset {F}$ is an_open_set, is called **tame**_of degree $r$ and base_$b$, if each point |
the case of gravity, we have $\delta_\xi \Phi= \mathsterling_\xi \Phi$ and $k$ can be written as [@Lee:1990nz; @Wald:1993nt; @Iyer:1994ys] (for more details see [@Azeyanagi:2009wf; @Detournay:2016gao]) $$\label{k}
\int_\Sigma k_\chi[\mathsterling_\xi \Phi; \Phi ]
=\int_{\Sigma} \left (\delta_\xi Q_\chi(\Phi)+i_\chi \Theta(\mathsterling_\xi \Phi, \Phi)
\right ) \,.$$ In the literature, another definition for the $n-2$-form is made by Barnich-Brandt-Compère [@Barnich:2001jy; @Barnich:2007bf]. Those two differ by the so-called $E-$term [@Azeyanagi:2009wf] $$E[\mathsterling_{\xi_1}g;\mathsterling_{\xi_2}g]=\frac12\left(-\frac32 Z^{abcd}\mathsterling_{\xi_1}g_c^e\wedge \mathsterling_{\xi_2} g_{ed}+2Z^{acde}\mathsterling_{\xi_1}g_{cd}\wedge \mathsterling_{\xi_2} g^b_e \right)\epsilon_{abc} dx^c\,.$$ It turns out that in the considered cases, this term always vanishes and so the two definitions are equivalent.
BTZ Black Holes
===============
Symmetries
----------
BTZ black holes are locally AdS$_3$. The latter is a maximally symmetric space. One main property of these spacetimes is that all tensors made out of the curvature and its covariant derivatives, like $R_{\mu\nu},\nabla_\mu R_{\nu\rho},...$ can be expressed in terms of the metric tensor. For example, $$R_{\mu\nu}=-\frac{2}{\ell^2} g_{\mu\nu}$$ where $\ell$ is the AdS radius. Note that there is no way to construct a 3-tensor just from products of the metric. Another consequence of the symmetries is that the left hand side of the equations of motion $E_{\mu\nu}$ evaluated on BTZ black holes, can always be rewritten as a constant times the metric | the case of gravity, we have $ \delta_\xi \Phi= \mathsterling_\xi \Phi$ and $ k$ can be written as [ @Lee:1990nz; @Wald:1993nt; @Iyer:1994ys ] (for more details visualize [ @Azeyanagi:2009wf; @Detournay:2016gao ]) $ $ \label{k }
\int_\Sigma k_\chi[\mathsterling_\xi \Phi; \Phi ]
= \int_{\Sigma } \left (\delta_\xi Q_\chi(\Phi)+i_\chi \Theta(\mathsterling_\xi \Phi, \Phi)
\right) \,.$$ In the literature, another definition for the $ n-2$-form is do by Barnich - Brandt - Compère [ @Barnich:2001jy; @Barnich:2007bf ]. Those two differ by the so - called $ E-$term [ @Azeyanagi:2009wf ] $ $ E[\mathsterling_{\xi_1}g;\mathsterling_{\xi_2}g]=\frac12\left(-\frac32 Z^{abcd}\mathsterling_{\xi_1}g_c^e\wedge \mathsterling_{\xi_2 } g_{ed}+2Z^{acde}\mathsterling_{\xi_1}g_{cd}\wedge \mathsterling_{\xi_2 } g^b_e \right)\epsilon_{abc } dx^c\,.$$ It change state out that in the considered cases, this terminus constantly vanishes and thus the two definitions are equivalent.
BTZ Black Holes
= = = = = = = = = = = = = = =
Symmetries
----------
BTZ bootleg holes are locally AdS$_3$. The latter is a maximally symmetric space. One main place of these spacetimes is that all tensors made out of the curvature and its covariant derivative instrument, like $ R_{\mu\nu},\nabla_\mu R_{\nu\rho},... $ can be expressed in terms of the measured tensor. For example, $ $ R_{\mu\nu}=-\frac{2}{\ell^2 } g_{\mu\nu}$$ where $ \ell$ is the AdS radius. Note that there be no way to construct a 3 - tensor merely from products of the metric function. Another consequence of the symmetries is that the left hand side of the equations of apparent motion $ E_{\mu\nu}$ evaluated on BTZ black holes, can always be rewritten as a constant times the metric | thf case of gravity, we havt $\delta_\xi \Phi= \matkwterliig_\xi \Phj$ and $k$ zan be written as [@Lee:1990nz; @Wald:1993it; @Iter:1994ys] (for more details see [@Xzeyanagi:2009af; @Detouenay:2016jao]) $$\label{k}
\int_\Sigma k_\chi[\mathstedping_\ri \Phi; \Phi ]
=\int_{\Sinma} \left (\denta_\xi Q_\chi(\Phi)+i_\whk \Cheta(\mathsterling_\xi \Phi, \Phi)
\right ) \,.$$ Ig the lotfrature, anothet defpnytioh for the $n-2$-form is made by Barnich-Grandt-Cmmpère [@Barnich:2001ky; @Barnich:2007bf]. Those two difver hy the so-called $E-$tfrm [@Azeyanati:2009wf] $$W[\mathsterline_{\xi_1}g;\mathsttrning_{\xi_2}g]=\frad12\left(-\frac32 Z^{abcd}\mathsterling_{\xi_1}g_c^d\wedgz \mathsterlunt_{\xi_2} c_{ed}+2Z^{acde}\matisterlpng_{\xi_1}g_{cd}\wedge \mathstesling_{\xi_2} g^b_e \right)\epsikon_{ebc} ex^c\,.$$ It turns out that in the considered cases, thhs term always vaniwhws ang so the rwo deriiitjons age xquivalent.
BFZ Black Hooes
===============
Symmetries
----------
BTZ blsch holes are lodally WdF$_3$. The latter is a maximally symmetric skace. Kne main property of thwse spacetimes is thaj all tensjrs made out of the curvature and its covariant darivavixes, llke $F_{\nu\ju},\nabla_\mu R_{\nu\rho},...$ can be expressed in terms of fht mvtric tensor. For cxample, $$R_{\mu\nu}=-\frac{2}{\ekl^2} g_{\ko\nu}$$ where $\ell$ ks the AdA radius. Note that there ys no way to cjnsttuct a 3-tensor just from proeucts of the netric. Another condequence of the symmrtries is that the left hans side of tje equatikvs of motion $E_{\mu\vu}$ vvalgated on BTZ black holes, cwn always be xewrittev as a conftant timed the metric | the case of gravity, we have $\delta_\xi \Phi$ $k$ can written as [@Lee:1990nz; see @Detournay:2016gao]) $$\label{k} \int_\Sigma \Phi; \Phi ] \left (\delta_\xi Q_\chi(\Phi)+i_\chi \Theta(\mathsterling_\xi \Phi, \Phi) ) \,.$$ In the literature, another definition for the $n-2$-form is made by [@Barnich:2001jy; @Barnich:2007bf]. Those two differ by the so-called $E-$term [@Azeyanagi:2009wf] $$E[\mathsterling_{\xi_1}g;\mathsterling_{\xi_2}g]=\frac12\left(-\frac32 Z^{abcd}\mathsterling_{\xi_1}g_c^e\wedge \mathsterling_{\xi_2} \mathsterling_{\xi_2} \right)\epsilon_{abc} It out that in the considered cases, this term always vanishes and so the two definitions are BTZ Black Holes =============== Symmetries ---------- BTZ black are locally AdS$_3$. The is a maximally symmetric space. main of these is all made out of curvature and its covariant derivatives, like $R_{\mu\nu},\nabla_\mu R_{\nu\rho},...$ can be expressed in terms of the metric tensor. example, $$R_{\mu\nu}=-\frac{2}{\ell^2} $\ell$ is AdS Note there is no construct a 3-tensor just from products Another consequence of the symmetries is that the hand side the equations of motion $E_{\mu\nu}$ evaluated BTZ black holes, can always be rewritten as constant times the metric | the case of gravity, we have $\delTa_\xi \Phi= \matHsterLinG_\xi \phI$ and $K$ can Be written as [@Lee:1990NZ; @WalD:1993nt; @Iyer:1994ys] (for more detailS see [@AZeYAnagI:2009Wf; @detouRnay:2016gao]) $$\LAbEL{K}
\inT_\SIgMa k_\ChI[\MaThsteRliNg_\xi \Phi; \phi ]
=\int_{\SigmA} \leFt (\Delta_\xi Q_\chi(\PHI)+i_\Chi \Theta(\maThsTerling_\xi \Phi, \phi)
\Right ) \,.$$ IN tHe lITeratUre, AnothEr defiNItion fOr the $n-2$-forM iS Made by bArnich-BRANdT-ComPère [@Barnich:2001jy; @BarnICh:2007BF]. Those two diffeR by the So-CAlLED $E-$tErm [@azeyanagi:2009wF] $$E[\MathsTErling_{\xI_1}G;\mATHSteRLing_{\xi_2}g]=\frac12\leFt(-\frac32 Z^{abcd}\MAthSterliNg_{\Xi_1}g_C^E\wedge \MathsTeRLinG_{\xi_2} g_{ed}+2Z^{acde}\MathSterling_{\xI_1}g_{cd}\weDGe \mathsTErling_{\xI_2} g^b_e \riGht)\EpsIlon_{ABc} Dx^C\,.$$ It TuRNs oUT tHat IN thE consideReD cAses, tHis tERM ALwayS vaNishEs and So the two definItiOns aRE eqUivalEnt.
BTz BlaCk holes
===============
symmetRies
----------
BtZ Black holes are loCallY AdS$_3$. The laTteR iS a mAxImallY SymmetRic SpaCe. One maIn propeRTy oF tHESE sPacetimes is that all TeNSOrS made out Of the cURvAtURe and its CoVarIant DERivatIves, LIkE $R_{\mu\nu},\naBla_\mu R_{\NU\rHo},...$ Can be exPrEssed iN tErmS of The meTRic tEnsor. FOr examplE, $$R_{\mu\nU}=-\Frac{2}{\ell^2} g_{\mu\nu}$$ whERe $\ell$ is the AdS RAdIUS. NOTe thAt tHere is no way To coNStruCt a 3-tENsOr jUSt froM prodUcTS oF The metric. Another conSeQuence Of the Symmetries is tHat the left HAND side of tHe eqUAtIOns of motion $E_{\mu\Nu}$ evaLuated on BTz Black holEs, can Always be Rewritten AS A constanT tiMes The MetRIC | the case of gravity, we h ave $\delt a_\xi \P hi= \ math ster ling_\xi \Phi$ and$k$ can be written as[@Lee :1 9 90nz ; @ Wald: 1993nt; @I y e r:1 99 4y s](f o rmoredet ails se e [@Azeyan agi :2 009wf; @Deto u rn ay:2016gao ])$$\label{k}\in t_\Sig ma k_ \ chi[\ mat hster ling_\ x i \Phi ; \Phi ]=\ i nt_{\S i gma} \l e f t(\de lta_\xi Q_\chi(\P h i) + i_\chi \Theta( \maths te r li n g _\x i \ Phi, \Phi)
\ right ) \,.$$ In t h e l i terature, ano ther defini t ion for t he $n - 2$-for m isma d e b y Barnich-B rand t-Compère [@Bar n ich:200 1 jy; @Ba rnich: 200 7bf ]. T h os etwo d i ffe r b y t h e s o-called $ E- $term [@A z e y a nagi :20 09wf ] $$E [\mathsterlin g_{ \xi_ 1 }g; \math sterl ing_ {\ xi_2} g]=\fr ac12\ le ft(-\frac32 Z^{ abcd }\mathste rli ng _{\ xi _1}g_ c ^e\wed ge\ma thsterl ing_{\x i _2} g _ { e d} +2Z^{acde}\mathste rl i n g_ {\xi_1}g _{cd}\ w ed ge \mathste rl ing _{\x i _ 2} g^ b_e\ ri ght)\eps ilon_{ a bc }dx^c\,. $$ It tu rn s o utthati n th e cons idered c ases, this term alwa y s vanishes an d s o th e two de finitions a re e q uiva lent .
BTZ Black Hole s= == = ===========
Symmet ri es
--- ----- --
BTZ black holes are l o cally Ad S$_3 $ .T he latter is a maxi mally symm e tric spa ce. O ne mainpropertyo f these s pac eti mes is t ha t all tensors m adeou t of th e c urvatur e a ndits co va riant der ivatives ,li ke $ R_{ \mu\n u },\nabla _\ muR_ {\n u\rho } ,...$can b e ex pr es s edin term s o f theme tr ic t ens or . For exa m ple , $$R_{ \mu\nu}=- \fr a c{2} {\ el l^2} g_ {\mu\nu}$$ wh er e $\ell$ i sthe AdS r a d ius. Not e that there is no wayt o const ruc t a 3 -ten sor justfro m prod uct s of th e metr ic. A no the r conse q u en ceof the symme t r ies is t ha t th e lefthand side of the e q uat ions of motio n $ E_{\ m u \n u}$ ev a lua te d on B TZ black holes, can alway sb erewrittena s a c onstant timesthe m e tric | the_case of_gravity, we have $\delta_\xi_\Phi= \mathsterling_\xi_\Phi$_and $k$_can_be written as_[@Lee:1990nz; @Wald:1993nt; @Iyer:1994ys]_(for more details see_[@Azeyanagi:2009wf; @Detournay:2016gao]) $$\label{k}
\int_\Sigma_k_\chi[\mathsterling_\xi_\Phi; \Phi ]
=\int_{\Sigma} \left (\delta_\xi Q_\chi(\Phi)+i_\chi \Theta(\mathsterling_\xi \Phi, \Phi)
\right ) \,.$$ In the literature,_another_definition for_the_$n-2$-form_is made by Barnich-Brandt-Compère [@Barnich:2001jy;_@Barnich:2007bf]. Those two differ by_the so-called_$E-$term [@Azeyanagi:2009wf] $$E[\mathsterling_{\xi_1}g;\mathsterling_{\xi_2}g]=\frac12\left(-\frac32 Z^{abcd}\mathsterling_{\xi_1}g_c^e\wedge \mathsterling_{\xi_2} g_{ed}+2Z^{acde}\mathsterling_{\xi_1}g_{cd}\wedge \mathsterling_{\xi_2} g^b_e_\right)\epsilon_{abc}_dx^c\,.$$ It turns_out that in the considered cases, this term always_vanishes and so the two definitions_are equivalent.
BTZ Black_Holes
===============
Symmetries
----------
BTZ_black_holes are locally AdS$_3$._The latter is a maximally symmetric_space. One main property of these_spacetimes is that all tensors made out_of the curvature and its covariant_derivatives, like $R_{\mu\nu},\nabla_\mu R_{\nu\rho},...$ can_be expressed_in terms of the metric_tensor. For example,_$$R_{\mu\nu}=-\frac{2}{\ell^2} g_{\mu\nu}$$_where $\ell$ is_the AdS radius. Note that there_is no way_to construct a 3-tensor just from_products_of the metric._Another_consequence_of the_symmetries is that_the_left hand_side_of the equations of motion $E_{\mu\nu}$_evaluated_on BTZ black holes, can always be_rewritten as a constant_times_the metric |
${{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}$ are monotonically decreasing in $\alpha$.
Furthermore, the entropies are additive, e.g. evaluation for an i.i.d. state ${\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}^{{\otimes}n}$ relative to another i.i.d. state ${\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}^{{\otimes}n}$ results in $$\label{eqn:alpha-add}
{{\ensuremath{H_{\alpha}^{}(\textnormal{A$^n$}|\textnormal{B$^n$})_{\rho^{{\otimes}n}|\sigma^{{\otimes}n}}}}} = n {{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}\,.$$ The $\alpha$-entropies are independent of the Hilbert spaces used to represent the density operators locally:
\[lemma:alpha-iso\] Let ${\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}\in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny AB}}}})}}$, ${{\sigma}_{\textnormal{\tiny B}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny B}}}})}}$, ${{\tau}_{\textnormal{\tiny CD}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny CD}}}})}}$, ${{\omega}_{\textnormal{\tiny D}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny D}}}})}}$ and $U$, $V$ isometries s.t. ${{\tau}_{\textnormal{\tiny CD}}} = {(U\!{\otimes}\!V)} {\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}(U^\dagger\!{\otimes}\!V^\dagger)$ and ${{\omega}_{\textnormal{\tiny D}}} = V {\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}V^\dagger$. Then, $$\label{eqn:alpha-iso}
{{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}} = {{\ensuremath{H_{\alpha}^{}(\textnormal{C}|\textnormal{D})_{\tau|\omega}}}}\,.$$
Conditional entropies are measures of the uncertainty about $A$ given $B$, hence we expect them to satisfy a data processing inequality, i.e. local processing by a trace-preserving completely positive map (TP | $ { { \ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}$ are monotonically decreasing in $ \alpha$.
Furthermore, the entropies are additive, e.g. evaluation for an i.i.d. department of state $ { \ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}^{{\otimes}n}$ proportional to another i.i.d. state $ { \ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}^{{\otimes}n}$ results in $ $ \label{eqn: alpha - attention deficit disorder }
{ { \ensuremath{H_{\alpha}^{}(\textnormal{A$^n$}|\textnormal{B$^n$})_{\rho^{{\otimes}n}|\sigma^{{\otimes}n } } } } } = n { { \ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}\,.$$ The $ \alpha$-entropies are independent of the Hilbert space used to represent the concentration hustler locally:
\[lemma: alpha - iso\ ] Let $ { \ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}\in { \ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny AB}}}})}}$, $ { { \sigma}_{\textnormal{\tiny B } } } \in { \ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny B}}}})}}$, $ { { \tau}_{\textnormal{\tiny CD } } } \in { \ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny CD}}}})}}$, $ { { \omega}_{\textnormal{\tiny D } } } \in { \ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny D}}}})}}$ and $ U$, $ V$ isometries s.t. $ { { \tau}_{\textnormal{\tiny CD } } } = { (U\!{\otimes}\!V) } { \ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}(U^\dagger\!{\otimes}\!V^\dagger)$ and $ { { \omega}_{\textnormal{\tiny D } } } = V { \ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}V^\dagger$. Then, $ $ \label{eqn: alpha - iso }
{ { \ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma } } } } = { { \ensuremath{H_{\alpha}^{}(\textnormal{C}|\textnormal{D})_{\tau|\omega}}}}\,.$$
Conditional information are measures of the uncertainty about $ A$ give $ B$, hence we expect them to meet a data processing inequality, i.e. local processing by a trace - preserving completely positive map (TP | ${{\enduremath{H_{\alpha}^{}(\textnormal{X}|\textnormal{B})_{\rho|\sigma}}}}$ ace monofonicallh decreasing in $\alpha$.
Furtherloee, tht entropies are addktive, e.g. enaluation for qn i.i.d. statx ${\ensuremath{{{\rhk}_{\bextnmcmal{\tiny AB}}}}}^{{\otimgs}n}$ relative to another i.i.g. sgace ${\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}^{{\otymes}n}$ rrsklts in $$\label{ezn:alkha-wdd}
{{\ehsuremath{H_{\alpha}^{}(\textnormal{A$^n$}|\textnorjal{B$^n$})_{\rhm^{{\otimes}n}|\sigma^{{\ptimes}n}}}}} = n {{\ensuremath{H_{\alphw}^{}(\texhnormal{A}|\textnormal{H})_{\rho|\sigma}}}}\,.$$ Tye $\ajpha$-entropies are indeptnbent of the Hilbert spaces used to represevt thz density okzeatlts locally:
\[lenma:aliha-iso\] Let ${\ensuremath{{{\sho}_{\textmormal{\tiny AB}}}}}\ik {\ensnremqth{\mathcal{S}({\ensuremath{\kathcal{H}_{\textnormal{\jiny AB}}}})}}$, ${{\sicmc}_{\textnormal{\tiny B}}} \in {\wnwuremdth{\mdthcxo{S}({\evsudekafh{\mathfal{I}_{\textnormal{\finy B}}}})}}$, ${{\tau}_{\twxtnormal{\tiny CD}}} \in {\egwuremath{\mathczl{S}({\enstrqmath{\mathcal{H}_{\textnormal{\tiny CD}}}})}}$, ${{\omega}_{\texunormzl{\tiny D}}} \in {\ensuremath{\mqthcal{S}({\ensuremath{\mathfal{H}_{\textnjrmal{\tiny D}}}})}}$ and $U$, $V$ isometries s.t. ${{\tau}_{\textnormal{\tinf CD}}} = {(J\!{\otnnes}\!V)} {\dbskremath{{{\rho}_{\textnormal{\tiny AB}}}}}(U^\dagger\!{\otimes}\!V^\daggqd)$ snc ${{\omega}_{\textnorial{\tiny D}}} = V {\emfuremath{{{\sigma}_{\jextnormal{\finy B}}}}}V^\dagger$. Then, $$\label{ezn:alpya-iso}
{{\ensuwemayh{H_{\alpha}^{}(\textnormal{A}|\textnormql{B})_{\rho|\sigma}}}} = {{\wnsuremath{H_{\alpha}^{}(\tertnormal{C}|\texcnormak{D})_{\tau|\pmega}}}}\,.$$
Conditional entropnes ars measures lf the unddrtainty about $A$ ginen $T$, hence we expect them to fatisfy a datc procesring inequwlity, i.e. lofal pvmcessing by a tracf-presgrving completelj positive map (TP | ${{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}$ are monotonically decreasing in $\alpha$. Furthermore, are e.g. evaluation an i.i.d. state i.i.d. ${\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}^{{\otimes}n}$ results $$\label{eqn:alpha-add} {{\ensuremath{H_{\alpha}^{}(\textnormal{A$^n$}|\textnormal{B$^n$})_{\rho^{{\otimes}n}|\sigma^{{\otimes}n}}}}} = {{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}\,.$$ The $\alpha$-entropies are independent of Hilbert spaces used to represent the density operators locally: \[lemma:alpha-iso\] Let ${\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}\in AB}}}})}}$, ${{\sigma}_{\textnormal{\tiny B}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny B}}}})}}$, ${{\tau}_{\textnormal{\tiny CD}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny CD}}}})}}$, ${{\omega}_{\textnormal{\tiny D}}} {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny and $V$ s.t. ${{\tau}_{\textnormal{\tiny CD}}} = {(U\!{\otimes}\!V)} {\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}(U^\dagger\!{\otimes}\!V^\dagger)$ and ${{\omega}_{\textnormal{\tiny D}}} = V {\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}V^\dagger$. Then, $$\label{eqn:alpha-iso} = {{\ensuremath{H_{\alpha}^{}(\textnormal{C}|\textnormal{D})_{\tau|\omega}}}}\,.$$ Conditional entropies are measures of the about $A$ given $B$, we expect them to satisfy data inequality, i.e. processing a completely positive map | ${{\ensuremath{H_{\alpha}^{}(\textnormaL{A}|\textnormAl{B})_{\rhO|\siGma}}}}$ ArE monOtonIcally decreasiNG in $\aLpha$.
Furthermore, the entrOpies ArE AddiTIvE, e.g. evAluatioN FoR AN i.i.D. sTaTe ${\eNsUReMath{{{\rHo}_{\tExtnormAl{\tiny AB}}}}}^{{\otImeS}n}$ Relative to anOThEr i.i.d. state ${\EnsUremath{{{\sigma}_{\TexTnormaL{\tIny b}}}}}^{{\OtimeS}n}$ rEsultS in $$\labEL{eqn:alPha-add}
{{\ensUrEMath{H_{\aLPha}^{}(\textNORmAl{A$^n$}|\Textnormal{B$^n$})_{\rho^{{\otIMeS}N}|\sigma^{{\otimes}n}}}}} = n {{\EnsureMaTH{H_{\ALPha}^{}(\TexTnormal{A}|\teXtNormaL{b})_{\rho|\sigMA}}}}\,.$$ THE $\ALphA$-Entropies are iNdependent oF The hilberT sPacES used tO reprEsENt tHe density opEratOrs locallY:
\[lemma:ALpha-iso\] lEt ${\ensurEmath{{{\rHo}_{\tExtNormAL{\tInY AB}}}}}\In {\ENsuREmAth{\MAthCal{S}({\ensuReMaTh{\matHcal{h}_{\TEXTnorMal{\Tiny aB}}}})}}$, ${{\sigMa}_{\textnormal{\tIny b}}} \in {\eNSurEmath{\MathcAl{S}({\eNsUremaTh{\mathCal{H}_{\tExTnormal{\tiny B}}}})}}$, ${{\tau}_{\TextNormal{\tinY CD}}} \In {\EnsUrEmath{\MAthcal{s}({\enSurEmath{\maThcal{H}_{\tEXtnOrMAL{\TiNy CD}}}})}}$, ${{\omega}_{\textnormaL{\tINY D}}} \In {\ensureMath{\maTHcAl{s}({\EnsuremaTh{\MatHcal{h}_{\TExtnoRmal{\TInY D}}}})}}$ and $U$, $V$ iSometrIEs S.t. ${{\Tau}_{\textNoRmal{\tiNy cD}}} = {(U\!{\OtiMes}\!V)} {\eNSureMath{{{\rhO}_{\textnorMal{\tiNY AB}}}}}(U^\dagger\!{\otimES}\!V^\dagger)$ and ${{\omEGa}_{\TEXtNOrmaL{\tiNy D}}} = V {\ensuremAth{{{\sIGma}_{\tExtnORmAl{\tINy B}}}}}V^\dAgger$. thEN, $$\lABel{eqn:alpha-iso}
{{\ensurEmAth{H_{\alPha}^{}(\teXtnormal{A}|\textNormal{B})_{\rho|\SIGMa}}}} = {{\ensureMath{h_{\AlPHa}^{}(\textnormal{C}|\tExtnoRmal{D})_{\tau|\omEGa}}}}\,.$$
ConditIonal EntropieS are measuRES of the unCerTaiNty AboUT $a$ gIven $B$, hence we eXPEct tHeM to satiSfy A data prOceSsiNg iNeqUaLity, i.e. locAl procesSiNg By A tRacE-presERving comPlEteLy PosItive MAp (TP | ${{\ensuremath{H_{\alpha} ^{}(\textn ormal {A} |\t ex tnor mal{ B})_{\rho|\sig m a}}} }$ are monotonically d ecrea si n g in $\ alpha $.
Fur t he r m ore ,th e e nt r op ies a readditiv e, e.g. ev alu at ion for an i . i. d. state $ {\e nsuremath{{{ \rh o}_{\t ex tno r mal{\ tin y AB} }}}}^{ { \otime s}n}$ rel at i ve toa notheri . i. d. s tate ${\ensuremat h {{ { \sigma}_{\text normal {\ t in y B}} }}} ^{{\otimes }n }$ re s ults in $$ \ l a bel { eqn:alpha-add }
{{\ensure m ath {H_{\a lp ha} ^ {}(\te xtnor ma l {A$ ^n$}|\textn orma l{B$^n$}) _{\rho ^ {{\otim e s}n}|\s igma^{ {\o tim es}n } }} }} =n{ {\e n su rem a th{ H_{\alph a} ^{ }(\te xtno r m a l {A}| \te xtno rmal{ B})_{\rho|\si gma }}}} \ ,.$ $ The $\al pha$ -e ntrop ies ar e ind ep endent of the H ilbe rt spaces us ed to r epres e nt the de nsi ty oper ators l o cal ly : \[ lemma:alpha-iso\]Le t ${ \ensurem ath{{{ \ rh o} _ {\textno rm al{ \tin y AB}}} }}\i n { \ensurem ath{\m a th ca l{S}({\ en surema th {\m ath cal{H } _{\t extnor mal{\tin y AB} } }})}}$, ${{\si g ma}_{\textnor m al { \ ti n y B} }}\in {\ensur emat h {\ma thca l {S }({ \ ensur emath {\ m at h cal{H}_{\textnormal {\ tiny B }}}}) }}$, ${{\tau} _{\textnor m a l {\tiny C D}}} \i n {\ensuremath{ \math cal{S}({\e n suremath {\mat hcal{H}_ {\textnor m a l{\tinyCD} }}} )}} $,$ { {\ omega}_{\text n o rmal {\ tiny D} }}\in {\e nsu rem ath {\m at hcal{S}({ \ensurem at h{ \m at hca l{H}_ { \textnor ma l{\ ti nyD}}}} ) }}$ an d $U$ , $V $is o met ries s. t .$ { {\ta u} _{ \tex tno rm al{\t inyC D}} } = {(U \!{\otime s}\ ! V)}{\ en suremat h{{{\rho}_{\t ex tnormal{\t in y A B}}}}} ( U ^\dagger \!{\otimes}\!V^\dagger) $ and ${ {\o mega} _{\t extnormal {\t iny D} }}= V {\ ensure math{ {{ \si g m a}_{\ t e xt nor ma l{\tiny B} } } }}V ^\dag ge r$.Then, $ $\label{eqn:alpha- i so}
{{\ensuremat h{H _{\a l p ha }^{ } (\ t ext no r mal { A }|\textnormal{B })_{\rho|\ si g ma }}}} = {{\ e nsu re math{H_ {\alpha }^{}( \ textnor mal{C}|\t extnormal {D })_{ \ t au| \omega}}}} \,.$$
C onditiona l entr o pi es ar e m easure softhe u ncerta i nty abou t $A$gi ven $B $, he nc e we exp ect them to satisfy a d ata pr ocess ing inequali ty, i.e . local p roce ssing by a tr ace -pres erv i ng co mple t el y p o sitiv e ma p (TP | ${{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}$_are monotonically_decreasing in $\alpha$.
Furthermore, the_entropies are_additive,_e.g. evaluation for_an_i.i.d. state ${\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}^{{\otimes}n}$_relative to another_i.i.d. state ${\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}^{{\otimes}n}$ results_in $$\label{eqn:alpha-add}
{{\ensuremath{H_{\alpha}^{}(\textnormal{A$^n$}|\textnormal{B$^n$})_{\rho^{{\otimes}n}|\sigma^{{\otimes}n}}}}} =_n_{{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}}\,.$$ The $\alpha$-entropies are independent of the Hilbert spaces used to represent the density_operators_locally:
\[lemma:alpha-iso\] Let_${\ensuremath{{{\rho}_{\textnormal{\tiny_AB}}}}}\in_{\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny AB}}}})}}$, ${{\sigma}_{\textnormal{\tiny B}}} \in_{\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny B}}}})}}$, ${{\tau}_{\textnormal{\tiny CD}}} \in_{\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny CD}}}})}}$,_${{\omega}_{\textnormal{\tiny D}}} \in {\ensuremath{\mathcal{S}({\ensuremath{\mathcal{H}_{\textnormal{\tiny D}}}})}}$ and $U$, $V$_isometries_s.t. ${{\tau}_{\textnormal{\tiny CD}}} =_{(U\!{\otimes}\!V)} {\ensuremath{{{\rho}_{\textnormal{\tiny AB}}}}}(U^\dagger\!{\otimes}\!V^\dagger)$ and ${{\omega}_{\textnormal{\tiny D}}} = V_{\ensuremath{{{\sigma}_{\textnormal{\tiny B}}}}}V^\dagger$. Then, $$\label{eqn:alpha-iso}
{{\ensuremath{H_{\alpha}^{}(\textnormal{A}|\textnormal{B})_{\rho|\sigma}}}} = {{\ensuremath{H_{\alpha}^{}(\textnormal{C}|\textnormal{D})_{\tau|\omega}}}}\,.$$
Conditional_entropies are measures_of_the_uncertainty about $A$ given_$B$, hence we expect them to_satisfy a data processing inequality, i.e. local_processing by a trace-preserving completely positive map_(TP |
x},\hat{y}})=u$ and $\cos(\widehat{\hat{y}_0,\hat{z}})=v$, and where: $$\displaystyle t_{\alpha,\alpha'} =
(-)^{t_1+t_2-\tau_x+t_2+T_3'-T}
\cdot {A \left\{\begin{array}{ccc} t_2&\tau_x'&T_3\\#4&T&T_3' \end{array}\right\}}
\cdot {A \left\{\begin{array}{ccc} t_2&t_1&\tau_x\\#4&T_3&\tau_x' \end{array}\right\}}$$
### $\alpha'$ of H type
We define again $x^g_{\alpha \alpha'}$, $y^g_{\alpha \alpha'}$, $z^g_{\alpha \alpha'}$ and an intermediate coordinate $\vec{y}_0$ in the following way: $$\left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}^g_{\alpha \alpha'} \\ \vec{z}^g_{\alpha \alpha'} \end{array} \right) =
\left( \begin{array}{ccc}
1&0&0\\
0& -\frac{1}{\sqrt{3}} & \frac{\sqrt{2}}{\sqrt{3}} \\
0& \frac{\sqrt{2}}{\sqrt{3}} & \frac{1}{\sqrt{3}}
\end{array} \right)
\left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}_0 \\ \vec{z} \end{array} \right)
\; \mbox{and} \;
\left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}_0 \\ \vec{z} \end{array} \right) =
\left( \begin{array}{ccc}
\frac{1}{2} & \frac{\sqrt{3}}{2} & 0\\
\frac{\sqrt{3}}{2} & -\frac{1}{2} & 0\\
0&0&1
\end{array} | x},\hat{y}})=u$ and $ \cos(\widehat{\hat{y}_0,\hat{z}})=v$, and where: $ $ \displaystyle t_{\alpha,\alpha' } =
(-)^{t_1+t_2-\tau_x+t_2+T_3'-T }
\cdot { A \left\{\begin{array}{ccc } t_2&\tau_x'&T_3\\#4&T&T_3' \end{array}\right\ } }
\cdot { A \left\{\begin{array}{ccc } t_2&t_1&\tau_x\\#4&T_3&\tau_x' \end{array}\right\}}$$
# # # $ \alpha'$ of H type
We define again $ x^g_{\alpha \alpha'}$, $ y^g_{\alpha \alpha'}$, $ z^g_{\alpha \alpha'}$ and an intermediate coordinate $ \vec{y}_0 $ in the following way: $ $ \left (\begin{array}{c } \vec{x}^g_{\alpha \alpha' } \\ \vec{y}^g_{\alpha \alpha' } \\ \vec{z}^g_{\alpha \alpha' } \end{array } \right) =
\left (\begin{array}{ccc }
1&0&0\\
0 & -\frac{1}{\sqrt{3 } } & \frac{\sqrt{2}}{\sqrt{3 } } \\
0 & \frac{\sqrt{2}}{\sqrt{3 } } & \frac{1}{\sqrt{3 } }
\end{array } \right)
\left (\begin{array}{c } \vec{x}^g_{\alpha \alpha' } \\ \vec{y}_0 \\ \vec{z } \end{array } \right)
\; \mbox{and } \;
\left (\begin{array}{c } \vec{x}^g_{\alpha \alpha' } \\ \vec{y}_0 \\ \vec{z } \end{array } \right) =
\left (\begin{array}{ccc }
\frac{1}{2 } & \frac{\sqrt{3}}{2 } & 0\\
\frac{\sqrt{3}}{2 } & -\frac{1}{2 } & 0\\
0&0&1
\end{array } | x},\hah{y}})=u$ and $\cos(\widehat{\hat{y}_0,\hau{z}})=v$, and where: $$\disklqystylx t_{\alphz,\alpha'} =
(-)^{t_1+t_2-\tau_x+t_2+T_3'-T}
\cdot {A \left\{\begin{acray}{xcc} t_2&\ucu_x'&T_3\\#4&T&T_3' \end{array}\right\}}
\cdot {A \lvft\{\begin{aeray}{rcc} t_2&t_1&\tau_x\\#4&T_3&\tau_x' \xhd{array}\vnght\}}$$
### $\zppha'$ if H type
We denine again $f^g_{\alpha \alpha'}$, $f^g_{\xl'ha \alpha'}$, $z^g_{\alpha \alpha'}$ and an interiediate clordinate $\vec{y}_0$ in uhe folmowing way: $$\left( \begin{array}{c} \vec{x}^g_{\ampha \alkha'} \\ \vec{y}^g_{\alpha \allha'} \\ \vec{z}^g_{\alpha \alpha'} \end{agray} \right) =
\left( \begin{wrray}{ccc}
1&0&0\\
0& -\ftzc{1}{\szet{3}} & \frac{\sqrt{2}}{\rqrt{3}} \\
0& \frab{\vqrt{2}}{\sqrt{3}} & \rrac{1}{\sqrt{3}}
\end{array} \right)
\left( \begkn{arrcy}{c} \vec{x}^g_{\alkkq \apkha'} \\ \vec{y}_0 \\ \vec{z} \end{array} \rlbht)
\; \mbmx{and} \;
\lrft( \begin{array}{g} \vec{e}^g_{\alpha \alpha'} \\ \vec{y}_0 \\ \vec{z} \end{array} \right) =
\lgft( \begin{asrcy}{ccc}
\frac{1}{2} & \frac{\sqrt{3}}{2} & 0\\
\frac{\vqrt{3}}{2} & -\ffqc{1}{2} & 0\\
0&0&1
\ens{acrag} | x},\hat{y}})=u$ and $\cos(\widehat{\hat{y}_0,\hat{z}})=v$, and where: $$\displaystyle t_{\alpha,\alpha'} \cdot \left\{\begin{array}{ccc} t_2&\tau_x'&T_3\\#4&T&T_3' \cdot {A \left\{\begin{array}{ccc} H We define again \alpha'}$, $y^g_{\alpha \alpha'}$, \alpha'}$ and an intermediate coordinate $\vec{y}_0$ the following way: $$\left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}^g_{\alpha \alpha'} \\ \vec{z}^g_{\alpha \alpha'} \right) = \left( \begin{array}{ccc} 1&0&0\\ 0& -\frac{1}{\sqrt{3}} & \frac{\sqrt{2}}{\sqrt{3}} \\ 0& \frac{\sqrt{2}}{\sqrt{3}} & \end{array} \left( \vec{x}^g_{\alpha \\ \vec{y}_0 \\ \vec{z} \end{array} \right) \; \mbox{and} \; \left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}_0 \\ \end{array} \right) = \left( \begin{array}{ccc} \frac{1}{2} & \frac{\sqrt{3}}{2} 0\\ \frac{\sqrt{3}}{2} & -\frac{1}{2} 0\\ 0&0&1 \end{array} | x},\hat{y}})=u$ and $\cos(\widehat{\hat{y}_0,\haT{z}})=v$, and wherE: $$\dispLayStyLe T_{\alpHa,\alPha'} =
(-)^{t_1+t_2-\tau_x+t_2+T_3'-T}
\cdOT {A \leFt\{\begin{array}{ccc} t_2&\tau_x'&T_3\\#4&T&t_3' \end{aRrAY}\rigHT\}}
\cDot {A \lEft\{\begiN{ArRAY}{ccC} t_2&T_1&\tAu_x\\#4&t_3&\tAU_x' \End{arRay}\Right\}}$$
### $\alPha'$ of H type
we dEfIne again $x^g_{\alPHa \Alpha'}$, $y^g_{\alpHa \aLpha'}$, $z^g_{\alpha \aLphA'}$ and an InTerMEdiatE coOrdinAte $\vec{Y}_0$ In the fOllowing wAy: $$\LEft( \begIN{array}{c} \VEC{x}^G_{\alpHa \alpha'} \\ \vec{y}^g_{\alpha \ALpHA'} \\ \vec{z}^g_{\alpha \alpHa'} \end{aRrAY} \rIGHt) =
\lEft( \Begin{array}{CcC}
1&0&0\\
0& -\frac{1}{\SQrt{3}} & \frac{\SQrT{2}}{\SQRt{3}} \\
0& \fRAc{\sqrt{2}}{\sqrt{3}} & \fraC{1}{\sqrt{3}}
\end{arrAY} \riGht)
\lefT( \bEgiN{Array}{c} \Vec{x}^g_{\AlPHa \aLpha'} \\ \vec{y}_0 \\ \vec{Z} \end{Array} \righT)
\; \mbox{aND} \;
\left( \beGIn{array}{C} \vec{x}^g_{\AlpHa \aLpha'} \\ \VEc{Y}_0 \\ \vEc{z} \EnD{ArrAY} \rIghT) =
\LefT( \begin{arRaY}{cCc}
\fraC{1}{2} & \fraC{\SQRT{3}}{2} & 0\\
\fraC{\sqRt{3}}{2} & -\frAc{1}{2} & 0\\
0&0&1
\end{Array} | x},\hat{y}})=u$ and $\cos( \widehat{\ hat{y }_0 ,\h at {z}} )=v$ , and where: $ $ \dis playstyle t_{\alpha,\a lpha' }=
( - )^ {t_1+ t_2-\ta u _x + t _2+ T_ 3' -T}
\ cd ot {A \l eft\{\b egin{array }{c cc } t_2&\tau_x ' &T _3\\#4&T&T _3' \end{array} \ri ght\}}
\cd o t {A\le ft\{\ begin{ a rray}{ ccc} t_2& t_ 1 &\tau_ x \\#4&T_ 3 & \t au_x ' \end{array}\rig h t\ } }$$
### $\alp ha'$ o fH t y p e
Wedefine aga in $x^g _ {\alpha \a l p h a'} $ , $y^g_{\alph a \alpha'}$ , $z ^g_{\a lp ha\ alpha' }$ an da n i ntermediate coo rdinate $ \vec{y } _0$ int he foll owingway : $ $\le f t( \ beg in { arr a y} {c} \ve c{x}^g_{ \a lp ha \a lpha ' } \ \ \v ec{ y}^g _{\al pha \alpha'}\\\vec { z}^ g_{\a lpha\alp ha '} \e nd{arr ay} \ ri ght) =
\left(\beg in{array} {cc c}
1& 0& 0\\
0 & -\fr ac{ 1}{ \sqrt{3 }} & \f r ac{ \s q r t {2 }}{\sqrt{3}} \\
0 &\ f ra c{\sqrt{ 2}}{\s q rt {3 } } & \fra c{ 1}{ \sqr t { 3}} \end { ar ray} \ri ght)
\ l ef t( \begin {a rray}{ c} \v ec{ x}^g_ { \alp ha \al pha'} \ \ \ve c {y}_0 \\ \vec { z} \end{arra y }\ r ig h t)
\ ; \ mbox{and} \ ;
\l e ft(\beg i n{ arr a y}{c} \vec {x } ^g _ {\alpha \alpha'} \\ \ vec{y} _0 \\ \vec{z} \end {array} \r i g h t) =
\l eft( \b e gin{array}{ccc }
\f rac{1}{2}& \frac{ \sqrt {3}}{2}& 0\\
\f r a c{\sqrt{ 3}} {2} & -\ f r ac {1}{2} & 0\\0 & 0&1\e nd{arra y} | x},\hat{y}})=u$ and_$\cos(\widehat{\hat{y}_0,\hat{z}})=v$, and_where: $$\displaystyle t_{\alpha,\alpha'} _=
(-)^{t_1+t_2-\tau_x+t_2+T_3'-T}
_\cdot_{A \left\{\begin{array}{ccc}_t_2&\tau_x'&T_3\\#4&T&T_3'_\end{array}\right\}}
\cdot {A_\left\{\begin{array}{ccc} t_2&t_1&\tau_x\\#4&T_3&\tau_x' \end{array}\right\}}$$
###_$\alpha'$ of H type
We_define again $x^g_{\alpha_\alpha'}$,_$y^g_{\alpha \alpha'}$, $z^g_{\alpha \alpha'}$ and an intermediate coordinate $\vec{y}_0$ in the following way: $$\left(_\begin{array}{c}_\vec{x}^g_{\alpha \alpha'}_\\_\vec{y}^g_{\alpha_\alpha'} \\ \vec{z}^g_{\alpha \alpha'} \end{array}_\right) =
\left( \begin{array}{ccc}
1&0&0\\
0&_-\frac{1}{\sqrt{3}} &_\frac{\sqrt{2}}{\sqrt{3}} \\
0& \frac{\sqrt{2}}{\sqrt{3}} & \frac{1}{\sqrt{3}}
\end{array} \right)
\left(_\begin{array}{c}_\vec{x}^g_{\alpha \alpha'} _\\ \vec{y}_0 \\ \vec{z} \end{array} \right)
\; \mbox{and}_\;
\left( \begin{array}{c} \vec{x}^g_{\alpha \alpha'} \\ \vec{y}_0_\\ \vec{z} \end{array}_\right)_=_
\left( \begin{array}{ccc}
\frac{1}{2} &_ \frac{\sqrt{3}}{2} & 0\\
\frac{\sqrt{3}}{2} &_ -\frac{1}{2} & 0\\
0&0&1
\end{array} |
\sum_{n=1}^{\infty }\frac{\left( -1\right)
^{n}q^{n}}{n^{s}}\text{.} \label{equation 120}$$As $q\rightarrow 1$ in (\[equation 120\]), turns into$$\lim_{q\rightarrow 1}\zeta \left( s,x:q\right) =\zeta \left( s,x:1\right)
:=\zeta \left( s,x\right) =2\sum_{n=1}^{\infty }\frac{\left( -1\right) ^{n}}{%
n^{s}}$$
which is well-known as Euler-Zeta function (see [@KIM]). By ([equation 118]{}) and (\[equation 120\]), we get $$\zeta \left( -n,x:q\right) =\frac{G_{n+1,q}\left( x\right) }{n+1}\text{.}$$
[99]{} S. Araci, D. Erdal, and J. J. Seo, A study on the fermionic $p$-adic $q$-integral representation on $%
%TCIMACRO{\U{2124} }%
%BeginExpansion
\mathbb{Z}
%EndExpansion
_{p}$ associated with weighted $q$-Bernstein and $q$-Genocchi polynomials, *Abstract and Applied Analysis*, Volume **2011**, Article ID 649248, 10 pages.
S. Araci, J. J. Seo, and D. Erdal, New Construction weighted ($h,q$)-Genocchi numbers and polynomials related to Zeta type function, *Discrete Dynamics in Nature and Society*, Volume **2011**, Article ID 487490, 7 pages.
S. Araci, M. Acikgoz, and Feng Qi, On the $q$-Genocchi numbers and polynomials with weight zero and their applications, Available at http://arxiv.org/abs/1202.2643
H. Jolany and H. Sharifi, Some results for the Apostol-Genocchi polynomials of higher order, *Bulletin of Malaysian Mathematical Sciences Society* (Article in press).
Yuan He and Chunping Wang, Some Formulae of Products of the Apostol-Bernoulli and Apostol-Euler Polynomials, *Discrete Dynamics in Nature and Society*, vol. **2012 | \sum_{n=1}^{\infty } \frac{\left (-1\right)
^{n}q^{n}}{n^{s}}\text {. } \label{equation 120}$$As $ q\rightarrow 1 $ in (\[equation 120\ ]), turns into$$\lim_{q\rightarrow 1}\zeta \left (s, x: q\right) = \zeta \left (s, x:1\right)
: = \zeta \left (s, x\right) = 2\sum_{n=1}^{\infty } \frac{\left (-1\right) ^{n}}{%
n^{s}}$$
which is well - known as Euler - Zeta function (see [ @KIM ]). By ([ equality 118 ] { }) and (\[equation 120\ ]), we make $ $ \zeta \left (-n, x: q\right) = \frac{G_{n+1,q}\left (x\right) } { n+1}\text{.}$$
[ 99 ] { } S. Araci, D. Erdal, and J. J. Seo, A study on the fermionic $ p$-adic $ q$-integral representation on $%
% TCIMACRO{\U{2124 } }%
% BeginExpansion
\mathbb{Z }
% EndExpansion
_ { p}$ consociate with slant $ q$-Bernstein and $ q$-Genocchi polynomials, * Abstract and Applied Analysis *, Volume * * 2011 * *, Article ID 649248, 10 page.
S. Araci, J. J. Seo, and D. Erdal, New Construction slant ($ h, q$)-Genocchi number and polynomials related to Zeta character function, * Discrete Dynamics in Nature and Society *, Volume * * 2011 * *, Article ID 487490, 7 pages.
S. Araci, M. Acikgoz, and Feng Qi, On the $ q$-Genocchi numbers and polynomial with weight zero and their applications, Available at http://arxiv.org/abs/1202.2643
H. Jolany and H. Sharifi, Some result for the Apostol - Genocchi polynomials of higher ordering, * Bulletin of Malaysian Mathematical Sciences Society * (Article in press).
Yuan He and Chunping Wang, Some Formulae of Products of the Apostol - Bernoulli and Apostol - Euler Polynomials, * Discrete Dynamics in Nature and Society *, vol. * * 2012 | \sum_{j=1}^{\infty }\frac{\left( -1\right)
^{n}q^{n}}{k^{s}}\text{.} \label{equcrion 120}$$Av $q\riggtarrow 1$ in (\[equation 120\]), turns into$$\lim_{q\cighrarroq 1}\zeta \left( s,x:q\right) =\zdta \left( d,x:1\right)
:=\zwta \owft( s,x\righv) =2\sum_{n=1}^{\inncy }\frzg{\left( -1\cight) ^{n}}{%
n^{s}}$$
which ls well-knowt as Euler-Zeta fjnetion (see [@KIM]). By ([equation 118]{}) and (\[equatyon 120\]), we gft $$\zeta \left( -n,v:q\ribrt) =\fdac{G_{n+1,q}\left( x\right) }{n+1}\text{.}$$
[99]{} S. Araci, D. Srdal, aid J. J. Seo, A stidy on the fermionic $p$-adic $q$-inhegral representatlon on $%
%TCIMQCRO{\T{2124} }%
%BeginExpanskon
\mathbb{Z}
%TnbExpansion
_{p}$ associated with weighted $q$-Bernrtein and $q$-Genoxcyi omlynomials, *Ebstrabt and Applied Analyshs*, Voluke **2011**, Article ID 649248, 10 'agew.
S. Araci, J. J. Seo, and V. Erdal, New Construcjion weighdeb ($h,q$)-Genocchi numbers qne polinomidls fwlaged tp Aeta tjpe function, *Siscrete Dybamics in Nature anc Ficiety*, Volume **2011**, Artislq ID 487490, 7 pages.
S. Araci, M. Acikgoz, and Feng Xi, Kn the $q$-Genocchi numberw and polynomials witj weight sero and their applications, Available at http://arxie.org/aus/1202.2643
F. Jilwvt wnd H. Sharifi, Some results for the Apostol-Genjdcni polynomials on higher order, *Bulkehim of Malaysian Mathematjcal Sciences Socifty* (Artycle un press).
Ytan Ne and Chunping Wang, Some Firmulae of Pgoduxts of the Apostol-Yernoulli anb Aposjol-Eulrr Polynomials, *Discrete Dynzmics in Nahure and Auciety*, vol. **2012 | \sum_{n=1}^{\infty }\frac{\left( -1\right) ^{n}q^{n}}{n^{s}}\text{.} \label{equation 120}$$As $q\rightarrow (\[equation turns into$$\lim_{q\rightarrow \left( s,x:q\right) =\zeta =2\sum_{n=1}^{\infty -1\right) ^{n}}{% n^{s}}$$ is well-known as function (see [@KIM]). By ([equation 118]{}) (\[equation 120\]), we get $$\zeta \left( -n,x:q\right) =\frac{G_{n+1,q}\left( x\right) }{n+1}\text{.}$$ [99]{} S. Araci, Erdal, and J. J. Seo, A study on the fermionic $p$-adic $q$-integral representation $% }% \mathbb{Z} _{p}$ associated with weighted $q$-Bernstein and $q$-Genocchi polynomials, *Abstract and Applied Analysis*, Volume **2011**, Article ID 10 pages. S. Araci, J. J. Seo, and Erdal, New Construction weighted numbers and polynomials related to type *Discrete Dynamics Nature Society*, **2011**, Article ID 7 pages. S. Araci, M. Acikgoz, and Feng Qi, On the $q$-Genocchi numbers and polynomials with weight and their at http://arxiv.org/abs/1202.2643 Jolany H. Some results for polynomials of higher order, *Bulletin of Society* (Article in press). Yuan He and Chunping Some Formulae Products of the Apostol-Bernoulli and Apostol-Euler *Discrete Dynamics in Nature and Society*, vol. **2012 | \sum_{n=1}^{\infty }\frac{\left( -1\right)
^{n}q^{n}}{N^{s}}\text{.} \labeL{equaTioN 120}$$As $Q\rIghtArroW 1$ in (\[equation 120\]), turNS intO$$\lim_{q\rightarrow 1}\zeta \lefT( s,x:q\rIgHT) =\zetA \LeFt( s,x:1\rIght)
:=\zetA \LeFT( S,x\rIgHt) =2\Sum_{N=1}^{\iNFtY }\frac{\LefT( -1\right) ^{n}}{%
N^{s}}$$
which is wEll-KnOwn as Euler-ZeTA fUnction (see [@kIM]). by ([equation 118]{}) anD (\[eqUation 120\]), We Get $$\ZEta \leFt( -n,X:q\rigHt) =\frac{g_{N+1,q}\left( X\right) }{n+1}\teXt{.}$$
[99]{} s. araci, D. eRdal, and j. j. seO, A stUdy on the fermionic $P$-AdIC $q$-integral reprEsentaTiON oN $%
%tcIMaCRo{\U{2124} }%
%BeginExpAnSion
\mAThbb{Z}
%EnDexPANSioN
_{P}$ associated wiTh weighted $q$-bErnStein aNd $Q$-GeNOcchi pOlynoMiALs, *ABstract and APpliEd AnalysiS*, VolumE **2011**, article id 649248, 10 pages.
S. araci, J. j. SeO, anD D. ErDAl, neW CoNsTRucTIoN weIGhtEd ($h,q$)-GenoCcHi NumbeRs anD POLYnomIalS relAted tO Zeta type funcTioN, *DisCRetE DynaMics iN NatUrE and SOciety*, volumE **2011**, ARticle ID 487490, 7 pages.
S. ARaci, m. Acikgoz, aNd FEnG Qi, on The $q$-GENocchi NumBerS and polYnomialS WitH wEIGHt Zero and their applicAtIONs, availablE at httP://ArXiV.Org/abs/1202.2643
H. JOlAny And H. sHArifi, some REsUlts for tHe AposTOl-geNocchi pOlYnomiaLs Of hIghEr ordER, *BulLetin oF MalaysiAn MatHEmatical SciencES Society* (ArticLE iN PReSS).
YuaN He And Chunping wang, sOme FOrmuLAe Of PROductS of thE APOsTOl-Bernoulli and ApostOl-euler POlynoMials, *Discrete dynamics in nATUre and SoCietY*, VoL. **2012 | \sum_{n=1}^{\infty }\frac{ \left( -1\ right )
^ {n} q^ {n}} {n^{ s}}\text{.} \ l abel {equation 120}$$As $q\ right ar r ow 1 $ i n (\[ equatio n 1 2 0 \]) ,tu rns i n to $$\li m_{ q\right arrow 1}\z eta \ left( s,x:q\ r ig ht) =\zeta \l eft( s,x:1\r igh t)
:=\ ze ta\ left( s, x\rig ht) =2 \ sum_{n =1}^{\inf ty }\frac { \left(- 1 \r ight ) ^{n}}{%
n^{s}}$ $
w hich is well-k nown a sE ul e r -Ze tafunction ( se e [@K I M]). By ([ e q u ati o n 118]{}) and (\[equatio n 12 0\]),we ge t $$\ze ta \l ef t ( - n,x:q\right ) =\ frac{G_{n +1,q}\ l eft( x\ r ight) } {n+1}\ tex t{. }$$[9 9] {}S. Ara c i, D. Erd al, andJ. J . Seo , As t u d y on th e fe rmion ic $p$-adic $ q$- inte g ral repr esent atio non $%
%TCIM ACRO{ \U {2124} }%
%Begi nExp ansion
\m ath bb {Z}
% EndEx p ansion
_{ p}$ associ ated wi t h w ei g h t ed $q$-Bernstein and $ q $ -G enocchipolyno m ia ls , *Abstra ct an d Ap p l ied A naly s is *, Volum e **20 1 1* *, Articl eID 649 24 8,10pages .
S. Araci , J. J.Seo,a nd D. Erdal, N e w Constructio n w e i gh t ed ( $h, q$)-Genocch i nu m bers and po lyn o mials rela te d t o Zeta type function ,*Discr ete D ynamics in Na ture and S o c i ety*, Vo lume ** 2 011**, Article ID 4 87490, 7 p a ges.
S. Arac i, M. Ac ikgoz, an d Feng Qi, On th e $ q$- G e no cchi numbersa n d po ly nomials wi th weig htzer o a ndth eir appli cations, A va il ab leat ht t p://arxi v. org /a bs/ 1202. 2 643
H . Jol anyan dH . S harifi, So m e res ul ts for th eApost ol-G e noc chi pol ynomialsofh ighe ror der, *B ulletin of Ma la ysian Math em ati cal Sc i e nces Soc iety* (Article in press ) .
Yuan He andChun ping Wang , S ome Fo rmu l ae ofProduc ts of t heA p ostol - B er nou ll i and Apos t o l-E ulerPo lyno mials,*Discrete Dynamics inNature and So cie ty*, v ol . * * 20 1 2 | \sum_{n=1}^{\infty }\frac{\left(_-1\right)
^{n}q^{n}}{n^{s}}\text{.} _\label{equation 120}$$As $q\rightarrow 1$_in (\[equation_120\]),_turns into$$\lim_{q\rightarrow_1}\zeta_\left( s,x:q\right) =\zeta_\left( s,x:1\right)
:=\zeta \left(_s,x\right) =2\sum_{n=1}^{\infty }\frac{\left( -1\right)_^{n}}{%
n^{s}}$$
which is well-known_as_Euler-Zeta function (see [@KIM]). By ([equation 118]{}) and (\[equation 120\]), we get $$\zeta \left(_-n,x:q\right)_=\frac{G_{n+1,q}\left( x\right)_}{n+1}\text{.}$$
[99]{}_S._Araci, D. Erdal, and J._J. Seo, A study on_the fermionic_$p$-adic $q$-integral representation on $%
%TCIMACRO{\U{2124} }%
%BeginExpansion
\mathbb{Z}
%EndExpansion
_{p}$ associated with_weighted_$q$-Bernstein and $q$-Genocchi_polynomials, *Abstract and Applied Analysis*, Volume **2011**, Article ID_649248, 10 pages.
S. Araci, J. J._Seo, and D._Erdal,_New_Construction weighted ($h,q$)-Genocchi numbers_and polynomials related to Zeta type_function, *Discrete Dynamics in Nature and_Society*, Volume **2011**, Article ID 487490, 7_pages.
S. Araci, M. Acikgoz, and Feng_Qi, On the $q$-Genocchi numbers_and polynomials_with weight zero and their_applications, Available at_http://arxiv.org/abs/1202.2643
H. Jolany_and H. Sharifi,_Some results for the Apostol-Genocchi polynomials_of higher order,_*Bulletin of Malaysian Mathematical Sciences Society*_(Article_in press).
Yuan He_and_Chunping_Wang, Some_Formulae of Products_of_the Apostol-Bernoulli_and_Apostol-Euler Polynomials, *Discrete Dynamics in Nature_and_Society*, vol. **2012 |
{Z}$ on some Cantor set $Z$ such that both $\varphi $ and $\psi $ are conjugated to derived systems of $\alpha $. The systems $\varphi $ and $\psi $ are Kakutani strong orbit equivalent when they are strongly orbit equivalent to derived systems of $\alpha $. Thus once again, a dynamical concept such as Kakutani strong orbit equivalence is characterized by a standard concept of C\*-algebra theory — Rieffel-Morita equivalence. It is thus natural to investigate analogous C\*-algebraic descriptions of other forms of equivalence between minimal free Cantor systems.
Two minimal free systems $\left( X,\varphi,\mathbb{Z}\right) $ and $\left( Y,\psi,\mathbb{Z}\right) $ are flip-Kakutani equivalent if $\left(
X,\varphi,\mathbb{Z}\right) $ is Kakutani equivalent to either $\left(
Y,\varphi,\mathbb{Z}\right) $ or its time reversed system $\left( Y,\varphi
\circ \sigma,\mathbb{Z}\right) $ where $\sigma :z\in \mathbb{Z}\mapsto -z$. Now, $\sigma $ is the only automorphism of $\mathbb{Z}$ besides the identity, but the situation is more complicated for $\mathbb{Z}^{d}$ in general. To generalize the notion of flip-Kakutani equivalence, S. Lightwood and the second author introduced in [@Lightwood07] the notion of orbit injection equivalence between minimal free actions of $\mathbb{Z}^{d}$ on the Cantor set. In general, for an action $\alpha $ of $\mathbb{Z}^{d}$ on a set $Z$ we denote the image of $z\in \mathbb{Z}^{d}$ by $\alpha ^{z}$. We recall from [@Lightwood07]:
\[OrbitInjection\]Let $\left( X,\varphi,\mathbb{Z}^{d}\right) $ and $%
\left( Y,\psi,\mathbb{Z}^{d}\right) $ be two free dynamical systems. Then a map $\theta :X\rightarrow Y$ is an *orbit injection* from $\left(
X,\varphi,\mathbb{Z}^{d}\right) $ to $\left( Y,\psi,\mathbb{Z}^{d}\right) $ when $\theta $ is a continuous open injection such that for all $x,y\in X$ we have: | { Z}$ on some Cantor set $ Z$ such that both $ \varphi $ and $ \psi $ are conjugated to derived systems of $ \alpha $. The system $ \varphi $ and $ \psi $ are Kakutani impregnable orbit equivalent when they are strongly orbit equivalent to derived system of $ \alpha $. Thus once again, a dynamical concept such as Kakutani solid scope equivalence is characterized by a standard concept of C\*-algebra theory — Rieffel - Morita equivalence. It is therefore natural to investigate analogous C\*-algebraic descriptions of other forms of equality between minimal free Cantor systems.
Two minimal free systems $ \left (X,\varphi,\mathbb{Z}\right) $ and $ \left (Y,\psi,\mathbb{Z}\right) $ are flip - Kakutani equivalent if $ \left (
X,\varphi,\mathbb{Z}\right) $ is Kakutani equivalent to either $ \left (
Y,\varphi,\mathbb{Z}\right) $ or its clock time reversed system $ \left (Y,\varphi
\circ \sigma,\mathbb{Z}\right) $ where $ \sigma: z\in \mathbb{Z}\mapsto -z$. Now, $ \sigma $ is the alone automorphism of $ \mathbb{Z}$ besides the identity, but the situation is more complicated for $ \mathbb{Z}^{d}$ in general. To generalize the notion of flip - Kakutani equality, S. Lightwood and the second author introduced in [ @Lightwood07 ] the notion of orbit injection equivalence between minimal spare actions of $ \mathbb{Z}^{d}$ on the Cantor set. In general, for an action $ \alpha $ of $ \mathbb{Z}^{d}$ on a set $ Z$ we denote the image of $ z\in \mathbb{Z}^{d}$ by $ \alpha ^{z}$. We recall from [ @Lightwood07 ]:
\[OrbitInjection\]Let $ \left (X,\varphi,\mathbb{Z}^{d}\right) $ and $%
\left (Y,\psi,\mathbb{Z}^{d}\right) $ be two free dynamical systems. Then a map $ \theta: X\rightarrow Y$ is an * orbit injection * from $ \left (
X,\varphi,\mathbb{Z}^{d}\right) $ to $ \left (Y,\psi,\mathbb{Z}^{d}\right) $ when $ \theta $ is a continuous open injection such that for all $ x, y\in X$ we have: | {Z}$ oj some Cantor set $Z$ such that both $\varpku $ and $\psi $ zre conjjgated to derived systems of $\aopha $. The systems $\varphi $ avd $\psi $ age Kakutabi surong orbit equivement whcu theg are wtrongly orbit equivalend to derived sfsgels of $\alpha $. Thus once again, a dynaiical cpnfept such as Kwkutsgi sfgokg orbit equivalence is charactedized bj a standard concrpt of C\*-algebra theory — Rifffep-Morita equivalencf. It is thuw naeyral to invertigate analogous C\*-alggbraic descriptions of other foros of equivalenxe behfeen minimao frev Cantor systcks.
Two kinimal free systems $\kefv( X,\vqrphi,\mathbb{Z}\right) $ anv $\left( Y,\psi,\mathbb{Z}\ridht) $ are xlnp-Kakutani equivalent id $\lefj(
X,\var[hi,\mxrhbc{Z}\rjgit) $ is Kanuteni equivalsnt to eithwr $\left(
Y,\varphi,\mathbn{Z}\wpbht) $ or its fime rqvqrsed system $\left( Y,\varphi
\circ \sigma,\mathtb{Z}\dight) $ where $\sigma :z\in \nathbb{Z}\mapsto -z$. Now, $\slgma $ is ehe only automorphism of $\mathbb{Z}$ besides the idendity, uug tkc rutkation is more complicated for $\mathbb{Z}^{d}$ in gegsrsl. To generalize the notion ov gjip-Kakutani eduivalzhcs, S. Lightwood and hhe secjnd aythor intwoduved in [@Lightwood07] the notion of orbit inlecrion equivalence bztween minimcl freg actipns of $\mathbb{Z}^{d}$ on the Eantor set. In genfral, for zv action $\alpha $ uf $\kadhbb{Z}^{d}$ on a set $Z$ we denotq the imaje of $z\in \maghbb{E}^{d}$ by $\ajpha ^{z}$. We gecall from [@Lightwood07]:
\[OrbltInjgction\]Net $\left( X,\garphi,\mathbb{Z}^{d}\right) $ and $%
\left( Y,\'xi,\mathbb{Z}^{d}\ribhd) $ te two fxee dykamical systems. Then a map $\thgta :X\righcarrow Y$ is an *ogbit injertion* from $\lqft(
X,\varphi,\matvhb{Z}^{d}\right) $ tm $\left( Y,\[si,\mqthbv{Z}^{d}\righg) $ when $\theta $ os a contpnbous open injection such thet fud all $x,y\in X$ we hace: | {Z}$ on some Cantor set $Z$ such $\varphi and $\psi are conjugated to The $\varphi $ and $ are Kakutani orbit equivalent when they are strongly equivalent to derived systems of $\alpha $. Thus once again, a dynamical concept as Kakutani strong orbit equivalence is characterized by a standard concept of C\*-algebra — equivalence. is natural to investigate analogous C\*-algebraic descriptions of other forms of equivalence between minimal free Cantor systems. minimal free systems $\left( X,\varphi,\mathbb{Z}\right) $ and $\left( $ are flip-Kakutani equivalent $\left( X,\varphi,\mathbb{Z}\right) $ is Kakutani to $\left( Y,\varphi,\mathbb{Z}\right) or time system $\left( Y,\varphi \sigma,\mathbb{Z}\right) $ where $\sigma :z\in \mathbb{Z}\mapsto -z$. Now, $\sigma $ is the only automorphism of $\mathbb{Z}$ besides identity, but is more for in To generalize the flip-Kakutani equivalence, S. Lightwood and the in [@Lightwood07] the notion of orbit injection equivalence minimal free of $\mathbb{Z}^{d}$ on the Cantor set. general, for an action $\alpha $ of $\mathbb{Z}^{d}$ a set $Z$ we denote the image of $z\in \mathbb{Z}^{d}$ by $\alpha ^{z}$. We recall \[OrbitInjection\]Let $\left( X,\varphi,\mathbb{Z}^{d}\right) $ $% \left( Y,\psi,\mathbb{Z}^{d}\right) be free systems. a map :X\rightarrow Y$ is an *orbit injection* from $\left( X,\varphi,\mathbb{Z}^{d}\right) $ to Y,\psi,\mathbb{Z}^{d}\right) $ when $\theta $ is a continuous open injection for $x,y\in X$ we | {Z}$ on some Cantor set $Z$ such that Both $\varphi $ And $\psI $ arE coNjUgatEd to Derived systems OF $\alpHa $. The systems $\varphi $ and $\pSi $ are kaKUtanI StRong oRbit equIVaLENt wHeN tHey ArE StRonglY orBit equiValent to deRivEd Systems of $\alpHA $. THus once agaIn, a Dynamical conCepT such aS KAkuTAni stRonG orbiT equivALence iS characteRiZEd by a sTAndard cONCePt of c\*-algebra theory — RieFFeL-morita equivaleNce. It iS tHUs NATurAl tO investigaTe AnaloGOus C\*-algEBrAIC DesCRiptions of othEr forms of eqUIvaLence bEtWeeN MinimaL free caNTor Systems.
Two mInimAl free sysTems $\leFT( X,\varphI,\Mathbb{Z}\Right) $ aNd $\lEft( y,\psi,\MAtHbB{Z}\rIgHT) $ arE FlIp-KAKutAni equivAlEnT if $\leFt(
X,\vARPHI,\matHbb{z}\rigHt) $ is KAkutani equivaLenT to eITheR $\left(
y,\varpHi,\maThBb{Z}\riGht) $ or iTs timE rEversed system $\leFt( Y,\vArphi
\circ \SigMa,\MatHbB{Z}\rigHT) $ where $\SigMa :z\In \mathbB{Z}\mapstO -Z$. NoW, $\sIGMA $ iS the only automorphiSm OF $\MaThbb{Z}$ besIdes thE IdEnTIty, but thE sItuAtioN IS more CompLIcAted for $\mAthbb{Z}^{D}$ In GeNeral. To GeNeraliZe The NotIon of FLip-KAkutanI equivalEnce, S. lIghtwood and the SEcond author inTRoDUCeD In [@LiGhtWood07] the notiOn of ORbit InjeCTiOn eQUivalEnce bEtWEeN Minimal free actions oF $\mAthbb{Z}^{D}$ on thE Cantor set. In gEneral, for aN ACTion $\alphA $ of $\mAThBB{Z}^{d}$ on a set $Z$ we deNote tHe image of $z\IN \mathbb{Z}^{D}$ by $\alPha ^{z}$. We reCall from [@LIGHtwood07]:
\[OrBitinjEctIon\]lET $\lEft( X,\varphi,\matHBB{Z}^{d}\rIgHt) $ and $%
\leFt( Y,\Psi,\mathBb{Z}^{D}\riGht) $ Be tWo Free dynamIcal systEmS. THeN a Map $\Theta :x\RightarrOw y$ is An *OrbIt injECtion* fRom $\leFt(
X,\vArPhI,\MatHbb{Z}^{d}\riGHt) $ TO $\Left( y,\pSi,\MathBb{Z}^{D}\rIght) $ wHen $\tHEta $ Is a contInuous opeN inJEctiOn SuCh that fOr all $x,y\in X$ we hAvE: | {Z}$ on some Cantor set $Z $ such tha t bot h $ \va rp hi $ and $\psi $ are c o njug ated to derived system s of$\ a lpha $. Thesystems $\ v a rph i$and $ \ ps i $ a reKakutan i strong o rbi tequivalent w h en they arestr ongly orbitequ ivalen ttod erive d s ystem s of $ \ alpha$. Thus o nc e again , a dyna m i ca l co ncept such as Kak u ta n i strong orbit equiv al e nc e ischa racterized b y a s t andardc on c e p t o f C\*-algebratheory — Ri e ffe l-Mori ta eq u ivalen ce. I ti s t hus natural toinvestiga te ana l ogous C \ *-algeb raic d esc rip tion s o foth er for m sofe qui valencebe tw een m inim a l f reeCan torsyste ms.
Two mini mal fre e sy stems $\le ft(X, \varp hi,\ma thbb{ Z} \right) $ and $ \lef t( Y,\psi ,\m at hbb {Z }\rig h t) $ a refli p-Kakut ani equ i val en t i f$\left(
X,\varphi, \m a t hb b{Z}\rig ht) $i sKa k utani eq ui val entt o eith er $ \ le ft(
Y,\v arphi, \ ma th bb{Z}\r ig ht) $or it s t ime r e vers ed sys tem $\le ft( Y , \varphi
\circ\ sigma,\mathbb { Z} \ r ig h t) $ wh ere $\sigma :z\ i n \m athb b {Z }\m a psto-z$.No w ,$ \sigma $ is the onl yautomo rphis m of $\mathbb {Z}$ besid e s the iden tity , b u t the situatio n ismore compl i cated fo r $\m athbb{Z} ^{d}$ ing e neral. T o g ene ral ize t he notion of fl i p -Kak ut ani equ iva lence,S.Lig htw ood a nd the se cond aut ho rin tr odu ced i n [@Light wo od0 7] th e not i on oforbit inj ec ti o n e quivale n ce b etwe en m inim alfr ee ac tion s of $\math bb{Z}^{d} $ o n the C an tor set . In general, f or an acti on $\ alpha$ of $\mat hbb{Z}^{d}$ on a set $Z $ we den ote theimag e of $z\i n \ mathbb {Z} ^ {d}$ b y $\al pha ^ {z }$. W e rec a l lfro m[@Lightwoo d 0 7]:
\[O rb itIn jection \]Let $\left( X,\v a rph i,\mathbb{Z}^ {d} \rig h t )$ a n d$ %
\ le f t(Y , \psi,\mathbb{Z} ^{d}\right )$ b e two free dyn am ical sy stems.Thena map $\ theta :X\ rightarro wY$ i s an*orbit inj ection*from $\le f t(
X, \ va rphi, \ma thbb{Z }^ {d} \righ t) $ t o $\ left( Y,\ps i, \mathb b{Z}^ {d }\right) $ when $\theta $ is acontin uousope n injecti ons uch that for all $x,y\in X $ w e h ave: | {Z}$ on_some Cantor_set $Z$ such that_both $\varphi_$_and $\psi_$_are conjugated to_derived systems of_$\alpha $. The systems_$\varphi $ and_$\psi_$ are Kakutani strong orbit equivalent when they are strongly orbit equivalent to derived_systems_of $\alpha_$._Thus_once again, a dynamical concept_such as Kakutani strong orbit_equivalence is_characterized by a standard concept of C\*-algebra theory_—_Rieffel-Morita equivalence. It_is thus natural to investigate analogous C\*-algebraic descriptions of_other forms of equivalence between minimal_free Cantor systems.
Two_minimal_free_systems $\left( X,\varphi,\mathbb{Z}\right) $_and $\left( Y,\psi,\mathbb{Z}\right) $ are flip-Kakutani_equivalent if $\left(
X,\varphi,\mathbb{Z}\right) $ is Kakutani_equivalent to either $\left(
Y,\varphi,\mathbb{Z}\right) $ or its_time reversed system $\left( Y,\varphi
\circ \sigma,\mathbb{Z}\right)_$ where $\sigma :z\in \mathbb{Z}\mapsto_-z$. Now,_$\sigma $ is the only_automorphism of $\mathbb{Z}$_besides the_identity, but the_situation is more complicated for $\mathbb{Z}^{d}$_in general. To_generalize the notion of flip-Kakutani equivalence,_S._Lightwood and the_second_author_introduced in_[@Lightwood07] the notion_of_orbit injection_equivalence_between minimal free actions of $\mathbb{Z}^{d}$_on_the Cantor set. In general, for an_action $\alpha $ of_$\mathbb{Z}^{d}$_on a set $Z$_we denote the image of_$z\in \mathbb{Z}^{d}$ by $\alpha ^{z}$. We_recall from_[@Lightwood07]:
\[OrbitInjection\]Let $\left(_X,\varphi,\mathbb{Z}^{d}\right) $ and $%
\left( Y,\psi,\mathbb{Z}^{d}\right) $ be two free dynamical systems._Then a map $\theta :X\rightarrow Y$_is an *orbit injection*_from $\left(
X,\varphi,\mathbb{Z}^{d}\right)_$_to $\left( Y,\psi,\mathbb{Z}^{d}\right)_$_when $\theta_$ is a continuous open injection such_that for_all $x,y\in X$ we have: |
}\mathscr{F}.\}$$
\[theorem: ample subvariety\] Let $Y$ be a smooth closed subscheme of a smooth projective scheme $X$.
1. $Y$ is ample if and only if its normal bundle is ample and the cohomological dimension of the complement is $r-1$. \[theorem: normal bundle\]
Assume further that $Y$ is an ample subscheme in $X$. Then
2. Generalized Lefschetz hyperplane theorem with rational coefficient holds, i.e. $H^{i}(X,\mathbb{Q})\rightarrow H^{i}(Y,\mathbb{Q})$ is an isomorphism for $i<\dim Y$ and is an injection for $i=\dim Y$. \[theorem: lefschetz\]
3. $Y$ is numerically positive, i.e. $Y\cdot Z>0$ for any effective cycle $Z$ of dimension $r$. \[theorem: intersection\]
4. \[theorem: H\^[i]{}\] $H^{i}(X,\mathscr{F})\rightarrow H^{i}(\hat{X},\hat{\mathscr{F}})$ is an isomorphism for $i<\dim Y$ and is injective for $i=\dim Y$. Here $\hat{X}$ is the formal completion of $X$ along $Y$, $\mathscr{F}$ is a locally free sheaf on $X$ and $\hat{\mathscr{F}}$ is its restriction to $\hat{X}$.
[@Ottem Theorem 5.4], [@Ottem Corollary 5.3] and [@Hartshorne Chapter III, Theorem 3.4] give \[theorem: normal bundle\], \[theorem: lefschetz\] and \[theorem: H\^[i]{}\] respectively. For \[theorem: intersection\], since the cohomological dimension of $(X-Y)=r-1$, $Y$ meets any effective cycle of dimension $r$. We can then apply the result of Fulton and Lazarsfeld [@Laz Corollary 8.4.3], which says if $Y$ has ample normal bundle and $Y$ meets $Z$, where $Z$ is for an effective cycle of complementary dimension to that of $Y$, then $Y\cdot Z>0$.
The above list of properties is incomplete, for a more complete picture, c.f. [@Ottem].
Partial regularity and a | } \mathscr{F}.\}$$
\[theorem: ample subvariety\ ] Let $ Y$ be a smooth closed subscheme of a smooth projective schema $ X$.
1. $ Y$ is ample if and merely if its normal bundle is ample and the cohomological dimension of the complement is $ r-1$. \[theorem: normal bundle\ ]
Assume far that $ Y$ is an ample subscheme in $ X$. Then
2. Generalized Lefschetz hyperplane theorem with rational coefficient holds, i.e. $ H^{i}(X,\mathbb{Q})\rightarrow H^{i}(Y,\mathbb{Q})$ is an isomorphism for $ i<\dim Y$ and is an injection for $ i=\dim Y$. \[theorem: lefschetz\ ]
3. $ Y$ is numerically positivist, i.e. $ Y\cdot Z>0 $ for any effective bicycle $ Z$ of dimension $ r$. \[theorem: intersection\ ]
4. \[theorem: H\^[i]{}\ ] $ H^{i}(X,\mathscr{F})\rightarrow H^{i}(\hat{X},\hat{\mathscr{F}})$ is an isomorphism for $ i<\dim Y$ and is injective for $ i=\dim Y$. Here $ \hat{X}$ is the conventional completion of $ X$ along $ Y$, $ \mathscr{F}$ is a locally free sheaf on $ X$ and $ \hat{\mathscr{F}}$ is its limitation to $ \hat{X}$.
[ @Ottem Theorem 5.4 ], [ @Ottem Corollary 5.3 ] and [ @Hartshorne Chapter III, Theorem 3.4 ] give \[theorem: normal bundle\ ], \[theorem: lefschetz\ ] and \[theorem: H\^[i]{}\ ] respectively. For \[theorem: intersection\ ], since the cohomological dimension of $ (X - Y)=r-1 $, $ Y$ meets any effective hertz of dimension $ r$. We can then apply the result of Fulton and Lazarsfeld [ @Laz Corollary 8.4.3 ], which says if $ Y$ has ample normal bundle and $ Y$ meet $ Z$, where $ Z$ is for an effective cycle of complementary dimension to that of $ Y$, then $ Y\cdot Z>0$.
The above tilt of properties is incomplete, for a more complete word picture, c.f. [ @Ottem ].
Partial regularity and a | }\mahhscr{F}.\}$$
\[theorem: ample subvxriety\] Let $Y$ be a smoovh clossd subscfeme of a smooth projective dcyeme $Z$.
1. $Y$ is ample if and ovly if itd normal bunvle is ample and the cohomologjgal dnmxnsion of the cpmplement hs $r-1$. \[theorem: nmroap bundle\]
Assume further that $Y$ is an ample xuhscheme in $X$. Tren
2. Denedalized Lefschetz hyperplane theorsm with rational coegficient holds, i.e. $H^{i}(X,\mathbh{Q})\rihhtarrow H^{i}(Y,\mathbb{E})$ is an isonorprusm for $i<\dim Y$ and is an injection for $i=\dim Y$. \[theorem: lefschetz\]
3. $H$ is uumerically pisihhve, i.e. $Y\cdov Z>0$ fog any effectiyv cycle $Z$ of domension $r$. \[theprek: ibtersection\]
4. \[theorem: I\^[i]{}\] $H^{i}(X,\mathscr{F})\rightatrow H^{i}(\hat{F},\hct{\mathscr{F}})$ is an isomirphism for $i<\dio Y$ xnd ix jnjectlve for $i=\dim G$. Here $\hat{X}$ is the formal compkeeppn of $X$ alonf $Y$, $\maehfcr{F}$ is a locally free sheaf on $X$ and $\hdt{\mzthscr{F}}$ is its restrictuon to $\hat{X}$.
[@Ottem Theotem 5.4], [@Ottem Corollary 5.3] and [@Hartshorne Chapter III, Theorem 3.4] ghve \[tieuren: kormxo hundle\], \[theorem: lefschetz\] and \[theorem: H\^[i]{}\] respesfifeky. For \[theorem: intersectipn\], sogce the cohomulogiccm simension of $(X-Y)=r-1$, $Y$ meets wny edfective syclr of dimension $r$. We can theb apply the gesuot of Fulton and Lczarsfeld [@Lav Coroklary 8.4.3], which says if $Y$ has aople normal bunfle and $Y$ oeets $Z$, where $Z$ ks nor an effective cycle of com[lementarb dimznsion tu thst of $R$, then $Y\cdlt Z>0$.
Tma above list of prlpertnes iv incomplehe, for a more complete picture, r.h. [@Ottem].
Partiak segllarity aud a | }\mathscr{F}.\}$$ \[theorem: ample subvariety\] Let $Y$ be closed of a projective scheme $X$. and if its normal is ample and cohomological dimension of the complement is \[theorem: normal bundle\] Assume further that $Y$ is an ample subscheme in $X$. 2. Generalized Lefschetz hyperplane theorem with rational coefficient holds, i.e. $H^{i}(X,\mathbb{Q})\rightarrow H^{i}(Y,\mathbb{Q})$ is isomorphism $i<\dim and an injection for $i=\dim Y$. \[theorem: lefschetz\] 3. $Y$ is numerically positive, i.e. $Y\cdot Z>0$ for effective cycle $Z$ of dimension $r$. \[theorem: intersection\] \[theorem: H\^[i]{}\] $H^{i}(X,\mathscr{F})\rightarrow H^{i}(\hat{X},\hat{\mathscr{F}})$ an isomorphism for $i<\dim Y$ is for $i=\dim Here is formal completion of along $Y$, $\mathscr{F}$ is a locally free sheaf on $X$ and $\hat{\mathscr{F}}$ is its restriction to $\hat{X}$. Theorem 5.4], 5.3] and Chapter Theorem give \[theorem: normal lefschetz\] and \[theorem: H\^[i]{}\] respectively. For the cohomological dimension of $(X-Y)=r-1$, $Y$ meets any cycle of $r$. We can then apply the of Fulton and Lazarsfeld [@Laz Corollary 8.4.3], which if $Y$ has ample normal bundle and $Y$ meets $Z$, where $Z$ is for an of complementary dimension to of $Y$, then Z>0$. above of is incomplete, a more complete picture, c.f. [@Ottem]. Partial regularity and a | }\mathscr{F}.\}$$
\[theorem: ample subvaRiety\] Let $Y$ bE a smoOth CloSeD subScheMe of a smooth proJEctiVe scheme $X$.
1. $Y$ is ample if and Only iF iTS norMAl BundlE is amplE AnD THe cOhOmOloGiCAl DimenSioN of the cOmplement iS $r-1$. \[tHeOrem: normal buNDlE\]
Assume furTheR that $Y$ is an amPle SubschEmE in $x$. then
2. GEneRalizEd LefsCHetz hyPerplane tHeORem witH RationaL COeFficIent holds, i.e. $H^{i}(X,\matHBb{q})\Rightarrow H^{i}(Y,\mAthbb{Q})$ Is AN iSOMorPhiSm for $i<\dim Y$ AnD is an INjectioN FoR $I=\DIm Y$. \[THeorem: lefscheTz\]
3. $Y$ is numeriCAllY positIvE, i.e. $y\Cdot Z>0$ fOr any EfFEctIve cycle $Z$ of DimeNsion $r$. \[theOrem: inTErsectiON\]
4. \[theoreM: H\^[i]{}\] $H^{i}(X,\MatHscR{F})\riGHtArRow h^{i}(\HAt{X},\HAt{\MatHScr{f}})$ is an isoMoRpHism fOr $i<\dIM y$ ANd is InjEctiVe for $I=\dim Y$. Here $\hat{X}$ Is tHe foRMal ComplEtion Of $X$ aLoNg $Y$, $\maThscr{F}$ Is a loCaLly free sheaf on $X$ And $\hAt{\mathscr{f}}$ is ItS reStRictiON to $\hat{x}$.
[@OtTem theorem 5.4], [@ottem CoROllArY 5.3] AND [@HArtshorne Chapter IIi, THEOrEm 3.4] give \[thEorem: nORmAl BUndle\], \[theOrEm: lEfscHETz\] and \[TheoREm: h\^[i]{}\] respecTively. fOr \[ThEorem: inTeRsectiOn\], SinCe tHe cohOMoloGical dImension Of $(X-Y)=r-1$, $y$ Meets any effectIVe cycle of dimeNSiON $R$. WE Can tHen Apply the resUlt oF fultOn anD laZarSFeld [@LAz CorOlLArY 8.4.3], Which says if $Y$ has amplE nOrmal bUndle And $Y$ meets $Z$, wheRe $Z$ is for an EFFEctive cyCle oF CoMPlementary dimeNsion To that of $Y$, tHEn $Y\cdot Z>0$.
the abOve list oF propertiES Is incompLetE, foR a mOre COMpLete picture, c.f. [@oTTem].
PArTial regUlaRity and A | }\mathscr{F}.\}$$
\[theo rem: ample subv ari ety \] Let $Y$ be a smooth c l osed subscheme of a smooth proj ec t ives ch eme $ X$.
1. $ Y $ is a mp leif an d onl y i f its n ormal bund leis ample and t h ecohomologi cal dimension o f t he com pl eme n t is$r- 1$. \ [theor e m: nor mal bundl e\ ]
Assu m e furth e r t hat$Y$ is an ample s u bs c heme in $X$. T hen
2 .Ge n e ral ize d Lefschet zhyper p lane th e or e m wit h rational coe fficient ho l ds, i.e.$H ^{i } (X,\ma thbb{ Q} ) \ri ghtarrow H^ {i}( Y,\mathbb {Q})$i s an is o morphis m for$i< \di m Y$ an disan inj e ct ion for $i=\dim Y $. \[th eore m : l efsc het z\]
3. $Y$ is numeri cal ly p o sit ive,i.e.$Y\c do t Z>0 $ forany e ff ective cycle $Z $ of dimensio n $ r$ . \ [t heore m : inte rse cti on\]
4 . \[th e ore m: H \ ^[ i]{}\] $H^{i}(X,\m at h s cr {F})\rig htarro w H ^{ i }(\hat{X }, \ha t{\m a t hscr{ F}}) $ i s an iso morphi s mfo r $i<\d im Y$ an disinj ectiv e for $i=\d im Y$. H ere $ \ hat{X}$ is the formal comple t io n of $X$alo ng $Y$, $\m aths c r{F} $ is aloc a lly f ree s he a fo n $X$ and $\hat{\ma th scr{F} }$ is its restrict ion to $\h a t { X}$.
[@ Otte m T h eorem 5.4], [@ Ottem Corollary 5.3] and [@Ha rtshorne ChapterI I I, Theor em3.4 ] g ive \ [t heorem: norma l bund le \], \[t heo rem: le fsc het z\] an d\[theorem : H\^[i] {} \] r es pec tivel y . For \[ th eor em : i nters e ction\ ], si nceth ec oho mologic a ld i mens io nof $ (X- Y) =r-1$ , $Y $ me ets any effectiv e c y cleof d imensio n $r$. We can t hen applyth e r esulto f Fultonand Lazarsfeld [@Laz Co r ollary8.4 .3],whic h says if $Y $ hasamp l e norm al bun dle a nd $Y $ meets $ Z$ , w he re $Z$ isf o r a n eff ec tive cycleof complementary d i men sion to thatof$Y$, t he n $ Y \c d otZ> 0 $.T he above list o f properti es is incomplet e , f or a more comple te pi c ture, c .f. [@Ott em].
Par ti al r e g ula rity and a | }\mathscr{F}.\}$$
\[theorem:_ample subvariety\]_Let $Y$ be a_smooth closed_subscheme_of a_smooth_projective scheme $X$.
1._ $Y$ is_ample if and only_if its normal_bundle_is ample and the cohomological dimension of the complement is $r-1$. \[theorem: normal bundle\]
Assume_further_that $Y$_is_an_ample subscheme in $X$. Then
2._ Generalized Lefschetz hyperplane theorem_with rational_coefficient holds, i.e. $H^{i}(X,\mathbb{Q})\rightarrow H^{i}(Y,\mathbb{Q})$ is an isomorphism_for_$i<\dim Y$ and_is an injection for $i=\dim Y$. \[theorem: lefschetz\]
3. _$Y$ is numerically positive, i.e. $Y\cdot_Z>0$ for any_effective_cycle_$Z$ of dimension $r$._\[theorem: intersection\]
4. \[theorem: H\^[i]{}\] $H^{i}(X,\mathscr{F})\rightarrow_H^{i}(\hat{X},\hat{\mathscr{F}})$ is an isomorphism for $i<\dim_Y$ and is injective for $i=\dim Y$._Here $\hat{X}$ is the formal completion_of $X$ along $Y$, $\mathscr{F}$_is a_locally free sheaf on $X$_and $\hat{\mathscr{F}}$ is_its restriction_to $\hat{X}$.
[@Ottem Theorem_5.4], [@Ottem Corollary 5.3] and [@Hartshorne_Chapter III, Theorem_3.4] give \[theorem: normal bundle\], \[theorem:_lefschetz\]_and \[theorem: H\^[i]{}\]_respectively._For_\[theorem: intersection\],_since the cohomological_dimension_of $(X-Y)=r-1$,_$Y$_meets any effective cycle of dimension_$r$._We can then apply the result of_Fulton and Lazarsfeld [@Laz_Corollary_8.4.3], which says if_$Y$ has ample normal bundle_and $Y$ meets $Z$, where $Z$_is for_an effective_cycle of complementary dimension to that of $Y$, then $Y\cdot Z>0$.
The_above list of properties is incomplete,_for a more complete_picture, c.f._[@Ottem].
Partial_regularity and a |
depend on our choices of $\varepsilon_{5}$ and $b$, and on $\nu_{r}$ and $w_{r}$.
In order to show Proposition \[prop21.1.3\], we will employ elementary large deviation estimates, which are given in the following two lemmas.
\[lem21.1.5\] Let $W(1), W(2), \ldots$ denote nonnegative i.i.d. random variables with mean $\beta< \infty$. Then, for each $\varepsilon> 0$, there exists $C_{12}> 0$, so that $$\label{eq21.2.1}
P \Biggl( \sum^{n}_{k=1} W(k) \le(1 - \varepsilon) \beta n \Biggr) \le
e^{-C_{12}n}.$$ When the support of $W(1)$ is contained in $[0,1]$ and $\varepsilon\in(0,1]$, $$\label{eq21.2.6}
P \Biggl( \sum^{n}_{k=1} W(k) \ge(1 + \varepsilon) \beta n \Biggr)
\le e^{{-C_{13}}{\varepsilon^{2} \beta n}},$$ where $C_{13}> 0$ does not depend on the distribution of $W(1)$ or on $\varepsilon$.
Both (\[eq21.2.1\]) and (\[eq21.2.6\]) are elementary large deviation bounds. We summarize the argument for (\[eq21.2.6\]); (\[eq21.2.1\]) can be shown directly or by applying (\[eq21.2.6\]) after truncating $W(k)$.
As usual, one employs the moment generating function $$\label{eq21.2.2}
\psi_{\theta}(n) = E
\bigl[e^{\theta\sum^{n}_{k=1}(W(k)-\beta)} \bigr]\qquad \mbox{for } \theta> 0.$$ By expanding the exponential for $n=1$, it follows that for appropriate $C_{14} \ge1$ and for $\theta\in(0,1]$, $$\label{eq21.2.3}
\psi_{\theta}(1) \le1 + C_{14}\beta\theta^{2},$$ and hence $$\label{eq21.2.4}
\psi_{\theta | depend on our choices of $ \varepsilon_{5}$ and $ b$, and on $ \nu_{r}$ and $ w_{r}$.
In order to show Proposition \[prop21.1.3\ ], we will use elementary big deviation estimates, which are given in the postdate two lemmas.
\[lem21.1.5\ ] Let $ W(1), W(2), \ldots$ denote nonnegative i.i.d. random variable with intend $ \beta < \infty$. Then, for each $ \varepsilon > 0 $, there exists $ C_{12 } > 0 $, so that $ $ \label{eq21.2.1 }
P \Biggl (\sum^{n}_{k=1 } W(k) \le(1 - \varepsilon) \beta n \Biggr) \le
e^{-C_{12}n}.$$ When the support of $ W(1)$ is contained in $ [ 0,1]$ and $ \varepsilon\in(0,1]$, $ $ \label{eq21.2.6 }
P \Biggl (\sum^{n}_{k=1 } W(k) \ge(1 + \varepsilon) \beta n \Biggr)
\le e^{{-C_{13}}{\varepsilon^{2 } \beta n}},$$ where $ C_{13 } > 0 $ does not depend on the distribution of $ W(1)$ or on $ \varepsilon$.
Both (\[eq21.2.1\ ]) and (\[eq21.2.6\ ]) are elementary big deviation bounds. We sum up the argument for (\[eq21.2.6\ ]); (\[eq21.2.1\ ]) can be shown directly or by applying (\[eq21.2.6\ ]) after truncating $ W(k)$.
As usual, one employs the here and now generating function $ $ \label{eq21.2.2 }
\psi_{\theta}(n) = E
\bigl[e^{\theta\sum^{n}_{k=1}(W(k)-\beta) } \bigr]\qquad \mbox{for } \theta > 0.$$ By expanding the exponential for $ n=1 $, it follows that for appropriate $ C_{14 } \ge1 $ and for $ \theta\in(0,1]$, $ $ \label{eq21.2.3 }
\psi_{\theta}(1) \le1 + C_{14}\beta\theta^{2},$$ and therefore $ $ \label{eq21.2.4 }
\psi_{\theta | deoend on our choices of $\vxrepsilon_{5}$ and $b$, and on $\nu_{r}$ ahd $w_{r}$.
In urder to show Proposition \[prlp21.1.3\], we wull employ elementary uarge devpation esrimaues, which are givxh in thc folmlwiny vwo lemmas.
\[lem21.1.5\] Lgt $W(1), W(2), \ldots$ denote nonnegdtkvz i.i.d. random variables with mean $\betw< \infty$. Tjen, for each $\vwrepxylon> 0$, there exists $C_{12}> 0$, so that $$\label{eq21.2.1}
P \Biggl( \vum^{n}_{k=1} W(k) \le(1 - \farepsilon) \beta n \Biggr) \le
f^{-C_{12}n}.$$ Ahen the support ov $W(1)$ is contqineq in $[0,1]$ and $\vardpsilon\in(0,1]$, $$\label{eq21.2.6}
P \Bigfl( \sum^{n}_{k=1} W(k) \ge(1 + \varepsilon) \beta n \Biygr)
\le e^{{-C_{13}}{\varg'willt^{2} \beta n}},$$ whxre $C_{13}> 0$ does not deivnd on dhe disyribution of $W(1)$ or on $\varepsilon$.
Both (\[eq21.2.1\]) anv (\[eq21.2.6\]) are elementary jarge devhacion bounds. We summaruzw the arggmeng fof (\[es21.2.6\]); (\[xq21.2.1\]) dan be shkwn directmy or by applying (\[eq21.2.6\]) after trumcwnong $W(k)$.
As usuzl, one eiploys the moment generating function $$\ldbem{eq21.2.2}
\psi_{\theta}(n) = E
\bigl[e^{\thera\sum^{n}_{k=1}(W(k)-\beta)} \bigr]\qquwd \mbox{fow } \theta> 0.$$ By expanding the exponential for $n=1$, it fmllowa thcb wir appropriate $C_{14} \ge1$ and for $\theta\in(0,1]$, $$\label{eq21.2.3}
\psi_{\trsts}(1) \ke1 + C_{14}\beta\theta^{2},$$ and hence $$\kahek{gq21.2.4}
\psi_{\theta | depend on our choices of $\varepsilon_{5}$ and on and $w_{r}$. order to show elementary deviation estimates, which given in the two lemmas. \[lem21.1.5\] Let $W(1), W(2), denote nonnegative i.i.d. random variables with mean $\beta< \infty$. Then, for each $\varepsilon> there exists $C_{12}> 0$, so that $$\label{eq21.2.1} P \Biggl( \sum^{n}_{k=1} W(k) \le(1 - \beta \Biggr) e^{-C_{12}n}.$$ the support of $W(1)$ is contained in $[0,1]$ and $\varepsilon\in(0,1]$, $$\label{eq21.2.6} P \Biggl( \sum^{n}_{k=1} W(k) \ge(1 \varepsilon) \beta n \Biggr) \le e^{{-C_{13}}{\varepsilon^{2} \beta n}},$$ $C_{13}> 0$ does not on the distribution of $W(1)$ on Both (\[eq21.2.1\]) (\[eq21.2.6\]) elementary deviation bounds. We the argument for (\[eq21.2.6\]); (\[eq21.2.1\]) can be shown directly or by applying (\[eq21.2.6\]) after truncating $W(k)$. As one employs generating function \psi_{\theta}(n) E \bigr]\qquad \mbox{for } By expanding the exponential for $n=1$, for appropriate $C_{14} \ge1$ and for $\theta\in(0,1]$, $$\label{eq21.2.3} \le1 + and hence $$\label{eq21.2.4} \psi_{\theta | depend on our choices of $\varepSilon_{5}$ and $b$, aNd on $\nU_{r}$ aNd $w_{R}$.
IN ordEr to Show PropositioN \[Prop21.1.3\], We will employ elementary Large DeVIatiON eStimaTes, whicH ArE GIveN iN tHe fOlLOwIng twO leMmas.
\[lem21.1.5\] let $W(1), W(2), \ldots$ DenOtE nonnegative I.I.d. Random variAblEs with mean $\beTa< \iNfty$. ThEn, For EAch $\vaRepSilon> 0$, There eXIsts $C_{12}> 0$, sO that $$\labeL{eQ21.2.1}
p \Biggl( \SUm^{n}_{k=1} W(k) \lE(1 - \VArEpsiLon) \beta n \Biggr) \le
e^{-C_{12}N}.$$ whEN the support of $W(1)$ Is contAiNEd IN $[0,1]$ And $\VarEpsilon\in(0,1]$, $$\lAbEl{eq21.2.6}
P \bIggl( \sum^{N}_{K=1} W(K) \GE(1 + \VarEPsilon) \beta n \BiGgr)
\le e^{{-C_{13}}{\varePSilOn^{2} \beta N}},$$ wHerE $c_{13}> 0$ does nOt depEnD On tHe distributIon oF $W(1)$ or on $\varEpsiloN$.
both (\[eq21.2.1\]) aND (\[eq21.2.6\]) are eLementAry LarGe deVIaTiOn bOuNDs. WE SuMmaRIze The argumEnT fOr (\[eq21.2.6\]); (\[eQ21.2.1\]) can BE SHOwn dIreCtly Or by aPplying (\[eq21.2.6\]) afteR trUncaTIng $w(k)$.
As uSual, oNe emPlOys thE momenT geneRaTing function $$\labEl{eq21.2.2}
\Psi_{\theta}(n) = e
\biGl[E^{\thEtA\sum^{n}_{K=1}(w(k)-\beta)} \BigR]\qqUad \mbox{For } \thetA> 0.$$ by eXpANDInG the exponential for $N=1$, iT FOlLows that For appROpRiATe $C_{14} \ge1$ and FoR $\thEta\iN(0,1]$, $$\LAbel{eQ21.2.3}
\psi_{\THeTa}(1) \le1 + C_{14}\betA\theta^{2},$$ ANd HeNce $$\labeL{eQ21.2.4}
\psi_{\thEtA | depend on our choices of$\varepsil on_{5 }$and $ b$,andon $\nu_{r}$ a n d $w _{r}$.
In order to sh ow Pr op o siti o n\[pro p21.1.3 \ ], w e w il lemp lo y e lemen tar y large deviation es ti mates, which ar e given in th e followingtwo lemma s.
\ [ lem21 .1. 5\] L et $W( 1 ), W(2 ), \ldots $d enoten onnegat i v ei.i. d. random variabl e sw ith mean $\bet a< \in ft y $. T hen , f or each $\ va repsi l on> 0$, th e r e ex i sts $C_{12}>0$, so that $$\ label{ eq 21. 2 .1}
P\Bigg l( \su m^{n}_{k=1} W(k ) \le(1 - \vare p silon)\ beta n\Biggr ) \ lee^{- C _{ 12 }n} .$ $ Wh e nthe sup port of$W (1 )$ is con t a i n ed i n $ [0,1 ]$ an d $\varepsilo n\i n(0, 1 ]$, $$\l abel{ eq21 .2 .6}
P \Bigg l( \s um ^{n}_{k=1} W(k) \ge (1 + \var eps il on) \ betan \Bigg r)\le e^{{-C _{13}}{ \ var ep s i l on ^{2} \beta n}},$$wh e r e$C_{13}> 0$ do e sno t dependon th e di s t ribut iono f$W(1)$ o r on $ \ va re psilon$ .
Both(\ [eq 21. 2.1\] ) and (\[eq 21.2.6\] ) are elementary lar g e deviation b o un d s .W e su mma rize the ar gume n t fo r (\ [ eq 21. 2 .6\]) ; (\[ eq 2 1. 2 .1\]) can be showndi rectly or b y applying (\ [eq21.2.6\ ] ) after tr unca t in g $W(k)$.
As u sual, one emplo y s the mo mentgenerati ng functi o n $$\labe l{e q21 .2. 2}\ p si _{\theta}(n)= E
\b ig l[e^{\t het a\sum^{ n}_ {k= 1}( W(k )- \beta)} \ bigr]\qq ua d\m bo x{f or }\ theta> 0 .$ $ B yexp andin g the e xpone ntia lfo r $n =1$, it fo l l owsth at for ap pr opria te $ C _{1 4} \ge1 $ and for $\ t heta \i n( 0,1]$,$$\label{eq21 .2 .3}
\psi_{ \t het a}(1)\ l e1 + C_{ 14}\beta\theta^{2},$$ a n d hence $$ \labe l{eq 21.2.4}
\ psi _{\the ta | depend_on our_choices of $\varepsilon_{5}$ and_$b$, and_on_$\nu_{r}$ and_$w_{r}$.
In_order to show_Proposition \[prop21.1.3\], we_will employ elementary large_deviation estimates, which_are_given in the following two lemmas.
\[lem21.1.5\] Let $W(1), W(2), \ldots$ denote nonnegative i.i.d. random_variables_with mean_$\beta<_\infty$._Then, for each $\varepsilon> 0$,_there exists $C_{12}> 0$, so_that $$\label{eq21.2.1}
P_\Biggl( \sum^{n}_{k=1} W(k) \le(1 - \varepsilon) \beta n_\Biggr)_\le
e^{-C_{12}n}.$$ When the_support of $W(1)$ is contained in $[0,1]$ and $\varepsilon\in(0,1]$,_$$\label{eq21.2.6}
P \Biggl( \sum^{n}_{k=1} W(k) \ge(1 +_\varepsilon) \beta n_\Biggr)
\le_e^{{-C_{13}}{\varepsilon^{2}_\beta n}},$$ where $C_{13}>_0$ does not depend on the_distribution of $W(1)$ or on $\varepsilon$.
Both_(\[eq21.2.1\]) and (\[eq21.2.6\]) are elementary large deviation_bounds. We summarize the argument for_(\[eq21.2.6\]); (\[eq21.2.1\]) can be shown_directly or_by applying (\[eq21.2.6\]) after truncating_$W(k)$.
As usual, one_employs the_moment generating function_$$\label{eq21.2.2}
\psi_{\theta}(n) = E
\bigl[e^{\theta\sum^{n}_{k=1}(W(k)-\beta)} \bigr]\qquad \mbox{for }_\theta> 0.$$ By_expanding the exponential for $n=1$, it_follows_that for appropriate_$C_{14}_\ge1$_and for_$\theta\in(0,1]$, $$\label{eq21.2.3}
\psi_{\theta}(1) \le1_+_C_{14}\beta\theta^{2},$$ and_hence_$$\label{eq21.2.4}
\psi_{\theta |
Explanation: The intention is for variables $x_i$ to be 1 if set $S_i$ is to be included in the exact cover, and 0 otherwise. Then constraints (\[exactcover-1\]) ensure that each entry in $U$ appears precisely once in the selected sets.
Final Input Size: The number of non-zeros in the constraints coefficients matrix is precisely equal to $\sum_{i=1}^s |S_i|$, and there are $|U|$ RHS entries. It is clear that the input size of the converted problem is a linear function of the original input size.
Hitting Set to 0-1 Integer Programming {#hitting-set-to-0-1-integer-programming.unnumbered}
--------------------------------------
Hitting Set: Given a family of sets, is it possible to construct a new set $W$ such that the intersection between the $W$ and any of the given sets has cardinality 1?\
Input: A family of sets $\{S_i\}$, with $s$ total sets and each set containing $|S_i|$ entries from a universe set $U$.\
Input size: $\sum_{i=1}^s |S_i|$.
0-1 Integer Programming: Is it possible to satisfy a set of linear equations in binary variables?
Conversion: Produce a new instance of 0-1 Integer Programming by introducing binary variables:
- $x_j$ for $j = 1, \hdots, |U|$.
Define $\sigma_{ij}$ to be 1 if set $S_i$ contains $j \in U$, and 0 otherwise. Then the 0-1 Integer Program is described by the following constraints: $$\begin{aligned}
\sum_{j=1}^{|U|} \sigma_{ij} x_j & = & 1, \qquad \forall i = 1, \hdots, s.\label{hittingset-1}\end{aligned}$$ Explanation: The intention is for variables $x_i$ to be 1 if they are to be included in $W$, and 0 otherwise. Then constraints (\[hittingset-1\]) ensure that each set $S_i$ contains precisely one of the selected variables.
Final Input Size: The number of non-zeros in the constraints coefficients matrix is precisely equal to $\sum_{i=1}^s |S_i|$, and there | Explanation: The intention is for variables $ x_i$ to be 1 if set $ S_i$ is to be included in the accurate binding, and 0 otherwise. Then constraints (\[exactcover-1\ ]) guarantee that each submission in $ U$ appears precisely once in the selected sets.
concluding Input Size: The number of non - zeros in the constraints coefficient matrix is precisely adequate to $ \sum_{i=1}^s |S_i|$, and there are $ |U|$ RHS entries. It is absolved that the input size of the converted problem is a analogue function of the original input size.
stumble Set to 0 - 1 Integer Programming { # hitting-set-to-0-1-integer-programming.unnumbered }
--------------------------------------
stumble Set: contribute a family of sets, is it possible to reconstruct a new set $ W$ such that the intersection between the $ W$ and any of the given set has cardinality 1?\
Input: A family of sets $ \{S_i\}$, with $ s$ total set and each set containing $ |S_i|$ entries from a universe set $ U$.\
Input size: $ \sum_{i=1}^s |S_i|$.
0 - 1 Integer Programming: Is it potential to meet a stage set of linear equations in binary variables?
Conversion: Produce a new instance of 0 - 1 Integer Programming by introduce binary variable:
- $ x_j$ for $ j = 1, \hdots, |U|$.
Define $ \sigma_{ij}$ to be 1 if typeset $ S_i$ control $ j \in U$, and 0 otherwise. Then the 0 - 1 Integer Program is described by the follow constraints: $ $ \begin{aligned }
\sum_{j=1}^{|U| } \sigma_{ij } x_j & = & 1, \qquad \forall i = 1, \hdots, s.\label{hittingset-1}\end{aligned}$$ Explanation: The intention is for variables $ x_i$ to be 1 if they are to be included in $ W$, and 0 differently. Then constraints (\[hittingset-1\ ]) ensure that each set $ S_i$ contains precisely one of the selected variables.
Final Input Size: The number of non - zeros in the constraints coefficients matrix is precisely equal to $ \sum_{i=1}^s |S_i|$, and there |
Exppanation: The intention ir for variables $x_i$ to ue 1 if aet $S_i$ ir to be included in the exacv cocer, abd 0 otherwise. Then conrtraints (\[vxactcovee-1\]) enwyre that eedh entrn in $H$ appzacs precisely onge in the salected sets.
Fitau Nnput Size: The number of non-zeros in the comshraints coeffisienus iatrjq ls precisely equal to $\sum_{i=1}^s |S_i|$, ahd thert are $|U|$ RHS entriex. It is clear that the inpkt slze of the convertfd problem us a oinear functkon of the original inkut size.
Hitting Set to 0-1 Integer Orogrcmming {#hittunt-seh-jo-0-1-integer-projrammigg.unnumbered}
--------------------------------------
Mptting Vet: Givrn a family of sevs, iw it possible to consvruct a new set $W$ sush that tve intersection betqewn thg $W$ atd avt ow tge gjven sfts has cardihality 1?\
Inpur: A family of sets $\{X_i\}$, qith $s$ total aets agd each set containing $|S_i|$ entries from a lnivsrse set $U$.\
Input size: $\sun_{i=1}^s |S_i|$.
0-1 Integer Progralming: Is yt possible to satisfy a set of linear equations hn biiafy rqriabuws?
Fonversion: Produce a new instance of 0-1 Integer Lrpggamming by introdmcing binary varianlfs:
- $x_j$ for $j = 1, \fdots, |B|$.
Serine $\sigma_{ij}$ to be 1 if sej $S_i$ cintains $j \in I$, and 0 otherwise. Then the 0-1 Unteger Proggam us described by thz following eonstrsints: $$\begin{aligned}
\sum_{j=1}^{|U|} \sigmc_{ij} x_j & = & 1, \qquad \vorall i = 1, \hdots, s.\label{hitginnsed-1}\end{aligned}$$ Explanation: Thq intentiin ix for vxrianles $x_y$ to be 1 iv then are to be includef in $C$, and 0 otherwisf. Then constraints (\[hittingset-1\]) eixure that eavh sen $S_i$ contcins pvecisely one of the selected rariablef.
Finau Input Side: The nukber of non-seros in the wlnstraints cmefficiegts natrux is pfdcisely equal yo $\sum_{i=1}^s |S_i|$, and theee | Explanation: The intention is for variables $x_i$ 1 set $S_i$ to be included 0 Then constraints (\[exactcover-1\]) that each entry $U$ appears precisely once in the sets. Final Input Size: The number of non-zeros in the constraints coefficients matrix precisely equal to $\sum_{i=1}^s |S_i|$, and there are $|U|$ RHS entries. It is that input of converted problem is a linear function of the original input size. Hitting Set to 0-1 Integer {#hitting-set-to-0-1-integer-programming.unnumbered} -------------------------------------- Hitting Set: Given a family of is it possible to a new set $W$ such the between the and of given sets has 1?\ Input: A family of sets $\{S_i\}$, with $s$ total sets and each set containing $|S_i|$ entries a universe Input size: |S_i|$. Integer Is it possible a set of linear equations in Produce a new instance of 0-1 Integer Programming introducing binary - $x_j$ for $j = 1, |U|$. Define $\sigma_{ij}$ to be 1 if set contains $j \in U$, and 0 otherwise. Then the 0-1 Integer Program is described by constraints: $$\begin{aligned} \sum_{j=1}^{|U|} \sigma_{ij} & = & \qquad i 1, s.\label{hittingset-1}\end{aligned}$$ Explanation: intention is for variables $x_i$ to be 1 if they are be included in $W$, and 0 otherwise. Then constraints (\[hittingset-1\]) each $S_i$ contains precisely of the selected variables. Input The number of non-zeros constraints matrix to |S_i|$, there |
Explanation: The intention is For variablEs $x_i$ tO be 1 If sEt $s_i$ is To be Included in the eXAct cOver, and 0 otherwise. Then coNstraInTS (\[exaCTcOver-1\]) eNsure thAT eACH enTrY iN $U$ aPpEArS precIseLy once iN the selectEd sEtS.
Final Input SIZe: the number oF noN-zeros in the cOnsTraintS cOefFIcienTs mAtrix Is precISely eqUal to $\sum_{i=1}^S |S_I|$, And theRE are $|U|$ RHs ENtRies. it is clear that the iNPuT Size of the conveRted prObLEm IS A liNeaR function oF tHe oriGInal inpUT sIZE.
hitTIng Set to 0-1 IntegEr ProgrammiNG {#hiTting-sEt-To-0-1-iNTeger-pRograMmINg.uNnumbered}
--------------------------------------
HiTtinG Set: Given A familY Of sets, iS It possiBle to cOnsTruCt a nEW sEt $w$ suCh THat THe IntERseCtion betWeEn The $W$ aNd anY OF THe giVen Sets Has caRdinality 1?\
InpuT: A fAmilY Of sEts $\{S_i\}$, With $s$ TotaL sEts anD each sEt conTaIning $|S_i|$ entries fRom a Universe sEt $U$.\
inPut SiZe: $\sum_{I=1}^S |S_i|$.
0-1 IntEgeR PrOgrammiNg: Is it pOSsiBlE TO SaTisfy a set of linear eQuATIoNs in binaRy variABlEs?
cOnversioN: PRodUce a NEW instAnce OF 0-1 INteger PrOgrammINg By IntroduCiNg binaRy VarIabLes:
- $x_j$ FOr $j = 1, \hDots, |U|$.
DEfine $\sigMa_{ij}$ tO Be 1 if set $S_i$ contaINs $j \in U$, and 0 otheRWiSE. thEN the 0-1 intEger Program Is deSCribEd by THe FolLOwing ConstRaINtS: $$\Begin{aligned}
\sum_{j=1}^{|U|} \siGmA_{ij} x_j & = & 1, \qQuad \fOrall i = 1, \hdots, s.\lAbel{hittinGSET-1}\end{aligNed}$$ EXPlANation: The intenTion iS for variabLEs $x_i$ to be 1 If theY are to be Included iN $w$, And 0 otherWisE. ThEn cOnsTRAiNts (\[hittingset-1\]) ENSure ThAt each sEt $S_I$ contaiNs pRecIseLy oNe Of the seleCted variAbLeS.
FInAl INput SIZe: The numBeR of NoN-zeRos in THe consTrainTs coEfFiCIenTs matriX Is PREcisElY eQual To $\sUm_{I=1}^s |S_i|$, aNd thERe |
Explanation: The intenti on is forvaria ble s $ x_ i$ t o be 1 if set $S_i $ isto be included in theexact c o ver, an d 0 o therwis e .T h enco ns tra in t s(\[ex act cover-1 \]) ensure th at each entryi n$U$ appear s p recisely onc e i n these lec t ed se ts.
Fin al Inp u t Size : The num be r of no n -zerosi n t he c onstraints coeffi c ie n ts matrix is p recise ly eq u a l t o $ \sum_{i=1} ^s |S_i | $, andt he r e are $|U|$ RHS ent ries. It is cle ar tha tthe inputsizeof the convertedprob lem is alinear functio n of the origi nal in puts iz e.
H it t ing Se t t o 0- 1 Intege rPr ogram ming { # h itti ng- set- to-0- 1-integer-pro gra mmin g .un numbe red}---- -- ----- ------ ----- -- --------------
Hit ting Set: Gi ve n a f amily of set s,isit poss ible to con st r u c ta new set $W$ such t h a tthe inte rsecti o nbe t ween the $ W$anda n y oftheg iv en setshas ca r di na lity 1? \Input: A fa mil y ofs ets$\{S_i \}$, wit h $s$ total sets and each set cont a in i n g$ |S_i |$entries fro m au nive rses et $U $ .\
In put s iz e :$ \sum_{i=1}^s |S_i|$ .
0-1 I ntege r Programming : Is it po s s i ble to s atis f ya set of linear equa tions in b i nary var iable s?
Conv ersion: P r o duce a n ewins tan ceo f 0 -1 Integer Pr o g ramm in g by in tro ducingbin ary va ria bl es:
- $x_j$ fo r$j = 1 , \ hdots , |U|$.
De fin e$\s igma_ { ij}$ t o be1 if s et $S_ i$ cont a in s $j \ in U $, a nd0other wise . Th en the0-1 Integ erP rogr am i s descr ibed by the f ol lowing con st rai nts: $ $ \ begin{al igned}
\sum_{j=1}^{|U|} \sigma_ {ij } x_j & = & 1, \qq uad \fora lli = 1,\hdots , s.\ la bel { h ittin g s et -1} \e nd{aligned } $ $ E xplan at ion: The in tention is for var i abl es $x_i$ to b e 1 ift h ey ar e t o be i n clu d e d in $W$, and 0 otherwise .T he n constrai n ts(\ [hittin gset-1\ ]) en s ure tha t each se t $S_i$ c on tain s pre cisely one of theselectedv ariab l es .
Fi nal Input S ize : The numbe r of non- zerosin the c onstr ai nts coef ficients matrix is prec iselyequal to $\sum_{i =1} ^ s | S_i|$, an d th ere |
Explanation: The_intention is_for variables $x_i$ to_be 1_if_set $S_i$_is_to be included_in the exact_cover, and 0 otherwise._Then constraints (\[exactcover-1\])_ensure_that each entry in $U$ appears precisely once in the selected sets.
Final Input Size:_The_number of_non-zeros_in_the constraints coefficients matrix is_precisely equal to $\sum_{i=1}^s |S_i|$,_and there_are $|U|$ RHS entries. It is clear that_the_input size of_the converted problem is a linear function of the_original input size.
Hitting Set to 0-1_Integer Programming {#hitting-set-to-0-1-integer-programming.unnumbered}
--------------------------------------
Hitting_Set:_Given_a family of sets,_is it possible to construct a_new set $W$ such that the_intersection between the $W$ and any of_the given sets has cardinality 1?\
Input:_A family of sets $\{S_i\}$,_with $s$_total sets and each set_containing $|S_i|$ entries_from a_universe set $U$.\
Input_size: $\sum_{i=1}^s |S_i|$.
0-1 Integer Programming: Is_it possible to_satisfy a set of linear equations_in_binary variables?
Conversion: Produce_a_new_instance of_0-1 Integer Programming_by_introducing binary_variables:
-_ $x_j$ for $j =_1,_\hdots, |U|$.
Define $\sigma_{ij}$ to be 1 if_set $S_i$ contains $j_\in_U$, and 0 otherwise._Then the 0-1 Integer Program_is described by the following constraints:_$$\begin{aligned}
\sum_{j=1}^{|U|} \sigma_{ij}_x_j &_= & 1, \qquad \forall i = 1, \hdots, s.\label{hittingset-1}\end{aligned}$$ Explanation:_The intention is for variables $x_i$_to be 1 if_they are_to_be included in_$W$,_and 0_otherwise. Then constraints (\[hittingset-1\]) ensure that each_set $S_i$_contains precisely one of the selected_variables.
Final Input Size: The_number_of non-zeros in the constraints coefficients_matrix is precisely equal to $\sum_{i=1}^s_|S_i|$, and there |
Numerica*, pp. 191–234, 2008.
S. Bubeck and R. Eldan, “The entropic barrier: a simple and optimal universal self-concordant barrier,” *In proceedings of the 28th Annual Conference on Learning Theory (COLT)*, 2015.
[^1]: Department of Electrical and Computer Engineering, Duke University, Durham, NC 27708 USA (e-mail: [email protected]).
[^2]: Department of Electrical and Computer Engineering and the Coordinated Science Laboratory, University of Illinois at Urbana-Champaign, Urbana, IL 61801 USA (e-mail: [email protected]).
[^3]: Department of Electrical and Computer Engineering, University of Wisconsin-Madison, Madison, WI 53796, USA; e-mail: [email protected].
[^4]: This work was supported by NSF grant CCF-1017564 and by AFOSR grant FA9550-10-1-0390. A preliminary version of this work was presented at the American Control Conference, Portland, OR, June 2014.
---
abstract: 'Within convex analysis, a rich theory with various applications has been evolving since the proximal average of convex functions was first introduced over a decade ago. When one considers the subdifferential of the proximal average, a natural averaging operation of the subdifferentials of the averaged functions emerges. In the present paper we extend the reach of this averaging operation to the framework of monotone operator theory in Hilbert spaces, transforming it into the resolvent average. The theory of resolvent averages contains many desirable properties. In particular, we study a detailed list of properties of monotone operators and classify them as dominant or recessive with respect to the resolvent average. As a consequence, we recover a significant part of the theory of proximal averages. Furthermore, we shed new light on the proximal average and present novel results and desirable properties the proximal average possesses which have not been previously available.'
author:
- 'Sedi Bartz[^1], Heinz H. Bauschke[^2], Sarah M. Moffat[^3], and Xianfu Wang[^4]'
date: 'May 11, 2015'
title: |
[The resolvent average of monotone operators:\
dominant and recessive properties]{}
---
[**2010 Mathematics Subject Classification:**]{} Primary 47H05 | Numerica *, pp. 191–234, 2008.
S. Bubeck and R. Eldan, “ The entropic barrier: a simple and optimal universal self - concordant barrier, ” * In proceeding of the twenty-eighth Annual Conference on Learning Theory (COLT) *, 2015.
[ ^1 ]: Department of Electrical and Computer Engineering, Duke University, Durham, NC 27708 USA (e - mail: [email protected]).
[ ^2 ]: Department of Electrical and Computer Engineering and the Coordinated Science Laboratory, University of Illinois at Urbana - Champaign, Urbana, IL 61801 USA (east - chain mail: [email protected]).
[ ^3 ]: Department of Electrical and Computer Engineering, University of Wisconsin - Madison, Madison, WI 53796, USA; e - mail: [email protected].
[ ^4 ]: This work was confirm by NSF grant CCF-1017564 and by AFOSR grant FA9550 - 10 - 1 - 0390. A preliminary adaptation of this work was presented at the American Control Conference, Portland, OR, June 2014.
---
abstract:' Within convex psychoanalysis, a rich theory with various applications has been develop since the proximal average of convex functions was first introduced over a decade ago. When one considers the subdifferential of the proximal average, a lifelike averaging operation of the subdifferentials of the averaged functions issue. In the present paper we extend the range of this averaging process to the framework of monotone operator theory in Hilbert spaces, transform it into the resolvent average. The theory of resolvent averages contains many desirable properties. In particular, we study a detailed list of properties of monotone operators and classify them as dominant or recessive with respect to the resolvent modal. As a consequence, we recover a significant character of the hypothesis of proximal averages. Furthermore, we shed new light on the proximal average and present novel results and desirable properties the proximal average posse which have not been previously available.'
author:
-' Sedi Bartz[^1 ], Heinz H. Bauschke[^2 ], Sarah M. Moffat[^3 ], and Xianfu Wang[^4 ]'
date:' May 11, 2015'
title: |
[ The resolvent average of monotone operators:\
dominant and recessive properties ] { }
---
[ * * 2010 Mathematics Subject Classification :* * ] { } Primary 47H05 | Nulerica*, pp. 191–234, 2008.
S. Bubeck and R. Tldan, “The entropie barrixr: a sijple and optimal universal self-concocdanr bareier,” *In proceedings of the 28th Ajnual Cobfertnce on Learning Vgeory (COLT)*, 2015.
[^1]: Dsiartmznv of Electrical and Compuder Engineerinc, Aune University, Durham, NC 27708 USA (e-mail: [[email protected]).
[^2]: Departmgnt og Elednrlcal and Computer Engineering ans the Cmordinated Scoence Laboratory, Universitj of Illinois at Urbanw-Champaign, Odbagq, IL 61801 USA (e-mxil: [email protected]).
[^3]: Dgpartment of Electrical and Compjter Zngineering, Ubivftsity of Wisronsin-Iadison, Madison, WI 53796, GSA; e-maol: [email protected].
[^4]: This work was snpported by NSF granj CCF-1017564 and ty AFOSR grant FA9550-10-1-0390. A peelimhnarf vefwiov or vhia work waa presentes at the Amwrican Control Conftregbr, Portland, OD, June 2014.
---
wbstract: 'Within convex analysis, a rich uheorg with various applicatuons has been evolvinh since tre proximal average of convex functions was first intrkauctd over q fecade ago. When one considers the subdifferenejak pf the proximaj average, a nwtital averaging uperatnkn of the subdifferejtials jf thw averageq fumctions emerges. In the preswnt paper we wxtend the reach oy this averaying okeratipn to the framework of oonofone operatlr theory kn Hilbert spacer, tgansxorming it into the resolvqnt averaje. Thz theory of tesolvegt averaged conbdins many desirablf prokertiev. In partifular, we study a detailed list of properties pf mototone o'eratovs and classify them as domincnt or rzcessixe with reapect tm the resoldent average. Dd a consequeice, we resovee a wignifizxnt part of thr theory of proximao averages. Furtherkorg, se shed new ligku ob the proximal svefagq wnv prefant novel revultr avc deskrable kxopcrtkes yhe proximal average posaesses which have mob been prgviously wvailable.'
authpr:
- 'Sedi Bartz[^1], Heine H. Bauvchye[^2], Satah M. Moffat[^3], and Xianfu Wang[^4]'
date: 'Jay 11, 2015'
titlf: |
[The resoldent average of konotone operators:\
dominant and recewsive properties]{}
---
[**2010 Mqthematics Subject Elsssificatioi:**]{} Primwry 47H05 | Numerica*, pp. 191–234, 2008. S. Bubeck and “The barrier: a and optimal universal the Annual Conference on Theory (COLT)*, 2015. Department of Electrical and Computer Engineering, University, Durham, NC 27708 USA (e-mail: [email protected]). [^2]: Department of Electrical and Computer and the Coordinated Science Laboratory, University of Illinois at Urbana-Champaign, Urbana, IL 61801 (e-mail: [^3]: of and Computer Engineering, University of Wisconsin-Madison, Madison, WI 53796, USA; e-mail: [email protected]. [^4]: This work was by NSF grant CCF-1017564 and by AFOSR grant A preliminary version of work was presented at the Control Portland, OR, 2014. abstract: convex analysis, a theory with various applications has been evolving since the proximal average of convex functions was first introduced a decade one considers subdifferential the average, a natural of the subdifferentials of the averaged the present paper we extend the reach of averaging operation the framework of monotone operator theory Hilbert spaces, transforming it into the resolvent average. theory of resolvent averages contains many desirable properties. In particular, we study a detailed list of monotone operators and them as dominant recessive respect the average. As consequence, we recover a significant part of the theory of proximal Furthermore, we shed new light on the proximal average and results desirable properties the average possesses which have been available.' author: - 'Sedi H. Sarah Xianfu date: 11, 2015' title: | resolvent average of monotone operators:\ and recessive properties]{} --- 47H05 | Numerica*, pp. 191–234, 2008.
S. Bubeck and R. EldaN, “The entropIc barRieR: a sImPle aNd opTimal universal SElf-cOncordant barrier,” *In procEedinGs OF the 28TH ANnual confereNCe ON leaRnInG ThEoRY (CoLT)*, 2015.
[^1]: DeParTment of electrical And coMputer EngineERiNg, Duke UnivErsIty, Durham, NC 27708 UsA (e-Mail: [email protected]).
[^2]: DEpartMent of eLectriCal and ComPuTEr EngiNEering aND ThE CooRdinated Science LaBOrATory, University Of IlliNoIS aT uRbaNa-CHampaign, UrBaNa, IL 61801 Usa (e-mail: [email protected]).
[^3]: DepartmeNt of ElectriCAl aNd CompUtEr ENGineerIng, UnIvERsiTy of WisconsIn-MaDison, MadiSon, WI 53796, Usa; e-mail: [email protected].
[^4]: ThIS wOrK waS sUPpoRTeD by nsF gRant CCF-1017564 aNd By aFOSR GranT fa9550-10-1-0390. a PrelImiNary VersiOn of this work wAs pReseNTed At the ameriCan COnTrol COnfereNce, PoRtLand, OR, June 2014.
---
abstrAct: 'WIthin convEx aNaLysIs, A rich THeory wIth VarIous appLicatioNS haS bEEN EvOlving since the proxImAL AvErage of cOnvex fUNcTiONs was firSt IntRoduCED over A decADe Ago. When oNe consIDeRs The subdIfFerentIaL of The ProxiMAl avErage, a Natural aVeragINg operation of tHE subdifferentIAlS OF tHE aveRagEd functions EmerGEs. In The pREsEnt PAper wE exteNd THe REach of this averaging OpEratioN to thE framework of mOnotone opeRATOr theory In HiLBeRT spaces, transfoRming It into the rESolvent aVeragE. The theoRy of resolVENt averagEs cOntAinS maNY DeSirable properTIEs. In PaRticulaR, we Study a dEtaIleD liSt oF pRoperties Of monotoNe OpErAtOrs And clASsify theM aS doMiNanT or reCEssive With rEspeCt To THe rEsolvenT AvERAge. AS a CoNseqUenCe, We recOver A SigNificanT part of thE thEOry oF pRoXimal avErages. FurtherMoRe, we shed neW lIghT on the PROximal avErage and present novel resULts and dEsiRable PropErties the ProXimal aVerAGe possEsses wHich hAvE noT BEen prEVIoUslY aVailable.'
auTHOr:
- 'SEdi BaRtZ[^1], HeiNz H. BausChke[^2], Sarah M. Moffat[^3], anD xiaNfu Wang[^4]'
date: 'MaY 11, 2015'
tiTle: |
[THE ReSolVEnT AveRaGE of MONotone operators:\
Dominant anD rECeSsive propeRTieS]{}
---
[**2010 MAthematIcs SubjEct ClASsificaTion:**]{} PrimaRy 47H05 | Numerica*, pp. 191–234, 2 008.
S. B ubeck an d R .Elda n, “ The entropic b a rrie r: a simple and optima l uni ve r sals el f-con cordant ba r r ier ,” * Inpr o ce eding s o f the 2 8th Annual Co nf erence on Le a rn ing Theory (C OLT)*, 2015.
[ ^1]: D ep art m ent o f E lectr ical a n d Comp uter Engi ne e ring,D uke Uni v e rs ity, Durham, NC 27708 US A (e-mail: peng .guan@ du k e. e d u).
[ ^2]: Depar tm ent o f Electr i ca l a ndC omputer Engin eering andt heCoordi na ted Scienc e Lab or a tor y, Universi ty o f Illinoi s at U r bana-Ch a mpaign, Urban a,IL6180 1 U SA (e -m a il: ma xim @ ill inois.ed u) .
[^3] : De p a r t ment of Ele ctric al and Comput erEngi n eer ing,Unive rsit yof Wi sconsi n-Mad is on, Madison, WI 537 96, USA;e-m ai l:wi llett @ discov ery .wi sc.edu.
[^4]: Thi sw o r kwas supported by N SF g ra nt CCF-1 017564 an db y AFOSRgr ant FA9 5 5 0-10- 1-03 9 0. A preli minary ve rs ion ofth is wor kwas pr esent e d at the A mericanContr o l Conference,P ortland, OR,J un e 20 1 4.
---
abstract:'Wit h in c onve x a nal y sis,a ric ht he o ry with various app li cation s has been evolvin g since th e p roximalaver a ge of convex func tions was first introduc ed ov er a dec ade ago.W h en one c ons ide rsthe s ub differentialo f the p roximal av erage,a n atu ral av er aging ope ration o fth esu bdi ffere n tials of t heav era ged f u nction s eme rges .In the presen t p a p er w eex tend th ereach oft his averag ing opera tio n toth eframewo rk of monoton eoperator t he ory in Hi l b ert spac es, transforming it int o the re sol ventaver age. Thethe ory of re s olvent avera ges c on tai n s many d es ira bl e properti e s . I n par ti cula r, we s tudy a detailed li s t o f propertiesofmono t o ne op e ra t ors a n d c l a ssify them as d ominant or r e ce ssive with res pe ct to t he reso lvent average . As a co nsequence ,we r e c ove r a signif icant pa rt of the theor y o f pro xim al ave ra ges . Fur thermo r e,we sh ed new l ight o n the p roximalaverage and present nov el res ultsand desirabl e p r ope rties the pro ximal aver age po ssess esw hichhave no t b e en pr evio u sly avail a bl e.' a ut hor:
- 'Sed i B art z[^1] , H e inz H. Bau schke[^2], SarahM . Moffat[^3],andX i anf u W a ng[^ 4] '
date: 'May 1 1,20 1 5 '
title: |
[The r esolvent a v erage of mo notone operat o r s: \
d omin ant and rece ssi ve propert ie s] { }
---
[** 20 10 Mat hemati c s Su b j ect Classificati on:** ] { } Pri m ary 47H0 5 | Numerica*,_pp. 191–234,_2008.
S. Bubeck and R. Eldan, “The_entropic barrier:_a_simple and_optimal_universal self-concordant barrier,”_*In proceedings of_the 28th Annual Conference_on Learning Theory_(COLT)*,_2015.
[^1]: Department of Electrical and Computer Engineering, Duke University, Durham, NC 27708 USA (e-mail:[email protected]).
[^2]:_Department of_Electrical_and_Computer Engineering and the Coordinated_Science Laboratory, University of Illinois_at Urbana-Champaign,_Urbana, IL 61801 USA (e-mail: [email protected]).
[^3]: Department of_Electrical_and Computer Engineering,_University of Wisconsin-Madison, Madison, WI 53796, USA; e-mail: [email protected].
[^4]:_This work was supported by NSF_grant CCF-1017564 and_by_AFOSR_grant FA9550-10-1-0390. A preliminary_version of this work was presented_at the American Control Conference, Portland,_OR, June 2014.
---
abstract: 'Within convex analysis,_a rich theory with various applications_has been evolving since the_proximal average_of convex functions was first_introduced over a_decade ago._When one considers_the subdifferential of the proximal average,_a natural averaging_operation of the subdifferentials of the_averaged_functions emerges. In_the_present_paper we_extend the reach_of_this averaging_operation_to the framework of monotone operator_theory_in Hilbert spaces, transforming it into the_resolvent average. The theory_of_resolvent averages contains many_desirable properties. In particular, we_study a detailed list of properties_of monotone_operators and_classify them as dominant or recessive with respect to the resolvent_average. As a consequence, we recover_a significant part of_the theory_of_proximal averages. Furthermore,_we_shed new_light on the proximal average and present_novel results_and desirable properties the proximal average_possesses which have not_been_previously available.'
author:
- 'Sedi Bartz[^1], Heinz H. Bauschke[^2],_Sarah M. Moffat[^3], and Xianfu Wang[^4]'
date:_'May 11, 2015'
title: |
___[The resolvent average of monotone_operators:\
dominant and_recessive properties]{}
---
[**2010 Mathematics_Subject Classification:**]{} Primary 47H05 |
1);
; (playOnS2) at (4,-1.5)
at (0,0) ;
(-0.6,0) node\[scale=0.75\] [$S$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) node [+]{};
at (2,0) ; at (3,0) ; at (4,1) ; at (4,-1) ;
(1.4,0) node\[scale=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) to (4,-1);
; (playOnS3) at (8,-0.75)
at (0,0) ;
(-0.6,0) node\[scale=0.75\] [$S'$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) node [+]{};
at (2,0) ; at (3,0) ; at (4,1) ; at (4,-1) ;
(1.4,0) node\[scale=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) to (4,-1);
; (11,-0.75) node [${\mathcal{ | 1);
; (playOnS2) at (4,-1.5)
at (0,0);
(-0.6,0) node\[scale=0.75\ ] [ $ S$ ] { };
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) node [ + ] { };
at (2,0); at (3,0); at (4,1); at (4,-1);
(1.4,0) node\[scale=0.75\ ] [ $ S'$ ] { };
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) to (4,-1);
; (playOnS3) at (8,-0.75)
at (0,0);
(-0.6,0) node\[scale=0.75\ ] [ $ S'$ ] { };
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) node [ + ] { };
at (2,0); at (3,0); at (4,1); at (4,-1);
(1.4,0) node\[scale=0.75\ ] [ $ S'$ ] { };
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) to (4,-1);
; (11,-0.75) lymph node [ $ { \mathcal { | 1);
; (plwyOnS2) at (4,-1.5)
at (0,0) ;
(-0.6,0) node\[scale=0.75\] [$R$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) nobw [+]{};
at (2,0) ; at (3,0) ; zt (4,1) ; at (4,-1) ;
(1.4,0) node\[scale=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) vo (4,1); (3,0) to (4,-1);
; (kjayOnS3) at (8,-0.75)
at (0,0) ;
(-0.6,0) nodd\[scale=0.75\] [$S'$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arx (135:225:0.7);
(0.6,0) niee [+]{};
at (2,0) ; at (3,0) ; at (4,1) ; at (4,-1) ;
(1.4,0) nkfe\[scclx=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) to (4,-1);
; (11,-0.75) node [${\mathcan{ | 1); ; (playOnS2) at (4,-1.5) at (0,0) node\[scale=0.75\] (0,0) – (0,0) – (-0.75,-0.5); [+]{}; (2,0) ; at ; at (4,1) at (4,-1) ; (1.4,0) node\[scale=0.75\] [$S'$]{}; – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7); (2,0) – (3,0); (3,0) to (3,0) to (4,-1); ; (playOnS3) at (8,-0.75) at (0,0) ; (-0.6,0) node\[scale=0.75\] [$S'$]{}; – (0,0) (-0.75,-0.5); arc (135:225:0.7); (0.6,0) node [+]{}; at (2,0) ; at (3,0) ; at (4,1) ; at (4,-1) (1.4,0) node\[scale=0.75\] [$S'$]{}; (2,0) – (1.25,0.5); (2,0) – (1.25,0.5) arc (135:225:0.7); (2,0) (3,0); (3,0) to (4,1); (3,0) (4,-1); (11,-0.75) node | 1);
; (playOnS2) at (4,-1.5)
at (0,0) ;
(-0.6,0) node\[scale=0.75\] [$S$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) arc (135:225:0.7);
(0.6,0) nOde [+]{};
at (2,0) ; at (3,0) ; at (4,1) ; aT (4,-1) ;
(1.4,0) node\[ScaLe=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) ArC (135:225:0.7);
(2,0) – (3,0); (3,0) to (4,1); (3,0) tO (4,-1);
; (plaYOnS3) at (8,-0.75)
at (0,0) ;
(-0.6,0) node\[scALe=0.75\] [$S'$]{};
(0,0) – (-0.75,0.5); (0,0) – (-0.75,-0.5); (-0.75,0.5) aRc (135:225:0.7);
(0.6,0) node [+]{};
at (2,0) ; at (3,0) ; at (4,1) ; at (4,-1) ;
(1.4,0) node\[scalE=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0) To (4,1); (3,0) TO (4,-1);
; (11,-0.75) nodE [${\MaThcal{ | 1);
; (playOnS2) at (4,-1 .5)
at (0 ,0) ;
( -0. 6, 0) n ode\ [scale=0.75\][ $S$] {};
(0,0) – (-0.75,0. 5); ( 0, 0 ) –( -0 .75,- 0.5); ( - 0. 7 5 ,0. 5) a rc(1 3 5: 225:0 .7) ;
(0.6 ,0) node [ +]{ };
at (2,0) ; at (3,0) ; a t ( 4,1) ; at (4 ,-1 ) ;
( 1. 4,0 ) node \[s cale= 0.75\] [$S'$] {};
(2,0 )– (1.25 , 0.5); ( 2 , 0) – ( 1.25,-0.5); (1.25 , 0. 5 ) arc (135:225 :0.7);
( 2, 0 ) –(3, 0); (3,0)to (4,1 ) ; (3,0) to ( 4 ,-1 ) ;
; (playOnS 3) at (8,-0 . 75)
at ( 0, 0);
(-0. 6,0)no d e\[ scale=0.75\ ] [$ S'$]{};
(0,0)– (-0.75 , 0.5); ( 0,0) – (- 0.7 5,-0 . 5) ;(-0 .7 5 ,0. 5 )arc (13 5:225:0. 7) ;
(0.6 ,0)n o d e [+] {};
at (2,0 ) ; at (3,0); a t (4 , 1); at(4,-1 ) ;
( 1.4,0 ) node \[sca le =0.75\] [$S'$]{ };
(2,0) – ( 1.2 5, 0.5 ); (2,0 ) – (1. 25, -0. 5); (1. 25,0.5) arc ( 1 3 5 :2 25:0.7);
(2,0) –(3 , 0 ); (3,0) t o (4,1 ) ;(3 , 0) to (4 ,- 1);
;( 1 1,-0. 75)n od e [${\ma thcal{ | 1);
; (playOnS2)_at (4,-1.5)
at_(0,0) ;
(-0.6,0) node\[scale=0.75\] [$S$]{};
(0,0)_– (-0.75,0.5);_(0,0)_– (-0.75,-0.5);_(-0.75,0.5)_arc (135:225:0.7);
(0.6,0) node_[+]{};
at (2,0) ;_at (3,0) ; at_(4,1) ; at_(4,-1)_;
(1.4,0) node\[scale=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0) – (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) – (3,0); (3,0)_to_(4,1); (3,0)_to_(4,-1);
;_(playOnS3) at (8,-0.75)
at (0,0) ;
(-0.6,0)_node\[scale=0.75\] [$S'$]{};
(0,0) – (-0.75,0.5); (0,0)_– (-0.75,-0.5);_(-0.75,0.5) arc (135:225:0.7);
(0.6,0) node [+]{};
at (2,0) ; at_(3,0)_; at (4,1)_; at (4,-1) ;
(1.4,0) node\[scale=0.75\] [$S'$]{};
(2,0) – (1.25,0.5); (2,0)_– (1.25,-0.5); (1.25,0.5) arc (135:225:0.7);
(2,0) –_(3,0); (3,0) to_(4,1);_(3,0)_to (4,-1);
; (11,-0.75) node_[${\mathcal{ |
[@Wootters89], the reason why (\[newGalois\]) defines the complete set of MUBs relies on the field theoretical formula $|\sum_{n=0}^{d-1}\omega_p^{tr[(a
\odot n \oplus b)\odot n ]}|n\rangle|=p^{1/2}$, with $ a \neq 0$ and $p$ being an odd prime. This method, however, fails for characteristic two where $|\sum_{n=0}^{d-1}\omega_2^{tr[(a \odot n
\oplus b)\odot n ]}|n\rangle|=0$ for any $a$, $b$. As shown in Sect. 3 below, for this characteristic one has to use Galois rings instead of Galois fields to get a complete set of MUBs.
A closer inspection of (\[newGalois\]) reveals an intricate relation between MUBs and quantum phase operators. It is known [@Planat04] that the Fourier basis $|\theta_k\rangle$ can be derived in terms of the eigenvectors of a quantum phase operator with eigenvalues $\theta_k$ and given by $\Theta_d=\sum_{k=0}^{d-1}\theta_k|\theta_k\rangle\langle
\theta_k|$. Similarly, using well known properties of the field trace, one can show that each base of index $a$ can be associated with a quantum phase operator $$\Theta_d^a=\sum_{b=0}^{d-1}\theta_b^a|\theta_b^a\rangle\langle\theta_b^a|,
\label{quantumphaseop}$$ with eigenvectors $|\theta_b^a\rangle$ and eigenvalues $\theta_b^a$; the latter may thus be called an “MUB operator".
MUBs FOR EVEN CHARACTERISTIC FROM GALOIS RINGS {#MUBsQubits}
==============================================
Our next goal is to find a Fourier transform formulation of MUBs in characteristic $2$. This is a very important case since the ‘$2^m-$dits’ are the basic information units in quantum computation. One may be tempted to connect the Galois field algebra with the generalized Pauli operators (\[Pauli\]) by constructing discrete vector spaces over the Galois field [@Wootters04bis]. For the one qubit case we already know that the eigenvectors of Pauli matrices $\sigma_z$, $\sigma_x$ and $\sigma_x | [ @Wootters89 ], the reason why (\[newGalois\ ]) defines the complete set of MUBs trust on the sphere theoretical formula $ |\sum_{n=0}^{d-1}\omega_p^{tr[(a
\odot n \oplus b)\odot n ] } |n\rangle|=p^{1/2}$, with $ a \neq 0 $ and $ p$ being an curious flower. This method, however, fails for characteristic two where $ |\sum_{n=0}^{d-1}\omega_2^{tr[(a \odot n
\oplus b)\odot n ] } |n\rangle|=0 $ for any $ a$, $ b$. As shown in Sect. 3 downstairs, for this characteristic one has to use Galois band instead of Galois fields to catch a complete bent of MUBs.
A close inspection of (\[newGalois\ ]) reveals an intricate relation between MUBs and quantum phase operators. It is known [ @Planat04 ] that the Fourier basis $ |\theta_k\rangle$ can be deduce in terms of the eigenvectors of a quantum phase operator with eigenvalues $ \theta_k$ and give by $ \Theta_d=\sum_{k=0}^{d-1}\theta_k|\theta_k\rangle\langle
\theta_k|$. Similarly, using well known properties of the field touch, one can show that each base of index $ a$ can be associated with a quantum phase operator $ $ \Theta_d^a=\sum_{b=0}^{d-1}\theta_b^a|\theta_b^a\rangle\langle\theta_b^a|,
\label{quantumphaseop}$$ with eigenvectors $ |\theta_b^a\rangle$ and eigenvalues $ \theta_b^a$; the latter may therefore be called an “ MUB operator ".
MUBs FOR EVEN CHARACTERISTIC FROM GALOIS RINGS { # MUBsQubits }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Our next goal is to find a Fourier transform formulation of MUBs in characteristic $ 2$. This is a very authoritative case since the ‘ $ 2^m-$dits ’ are the basic information units in quantum computation. One may be tempted to connect the Galois field algebra with the generalized Pauli operators (\[Pauli\ ]) by constructing discrete vector spaces over the Galois battlefield [ @Wootters04bis ]. For the one qubit case we already know that the eigenvectors of Pauli matrices $ \sigma_z$, $ \sigma_x$ and $ \sigma_x | [@Woltters89], the reason why (\[nedGalois\]) defines the cokplete set of OUBs relies on the field thelrwticao formula $|\sum_{n=0}^{d-1}\omega_p^{tf[(a
\odot n \lplus b)\oeot i ]}|n\rangle|=p^{1/2}$, with $ a \neq 0$ and $p$ gcing cn odd prime. Thix method, hmwever, fails fmr ckaracteristic two where $|\sum_{n=0}^{d-1}\omega_2^{tr[(w \odot m
\oolus b)\odot n ]}|n\tanglt|=0$ fjr ahj $c$, $b$. As shown in Sect. 3 below, for this ciaracteristic ome has to use Galois rings inshead of Galois fiepds to get q coiplete set of MUBs.
A closer inspectikn of (\[newGalois\]) reveals an intrkcate relation verweft MUBs and wuantlm phase operators. It is knoen [@Planat04] that thx Foyrier basis $|\theta_k\ranjle$ can be derived ig terms ox che eigenvectors of a qyantuk phdse uperxtod xitg eigejvamues $\theta_i$ and given by $\Theta_d=\sum_{k=0}^{d-1}\thets_k|\eyeta_k\rangle\lahgle
\thqtw_k|$. Similarly, using well known propertiev or the field trace, one cqn show that each basg of index $a$ can be associated with a quantum phase operatos $$\Theva_a^a=\sbn_{b=0}^{d-1}\thdra_h^a|\theta_b^a\rangle\langle\theta_b^a|,
\label{quantumphasejl}$$ einh eigenvectors $|\tmeta_b^a\rangle$ and eogfnfwlues $\theta_b^a$; the lcftsr may thus be calped an “IUB operator".
MUFs FPR EVEN CHARACTERISTIC FROM GALOIS RINGF {#MUBsQubits}
==============================================
Our nexc goal is to finc a Fpurier transform formulction kf MUBs in fharacterjrtic $2$. This is a xerj im[ortant case since the ‘$2^m-$dies’ are thx basnc inforoatipn unies in quanhum computation. One may he telpded to conject the Galois field algebra wmvh the generakised Pauli o'eratovs (\[Pauli\]) by conftructing disctete vectjr spxces over nhe Galoiv field [@Wooeters04bis]. For dje one qubit case we alrwady know tfxt the eigenvevtors of Iabli matrixes $\sigma_z$, $\sigma_x$ end $\rjgma_x | [@Wootters89], the reason why (\[newGalois\]) defines the of relies on field theoretical formula n with $ a 0$ and $p$ an odd prime. This method, however, for characteristic two where $|\sum_{n=0}^{d-1}\omega_2^{tr[(a \odot n \oplus b)\odot n ]}|n\rangle|=0$ for any $b$. As shown in Sect. 3 below, for this characteristic one has to Galois instead Galois to get a complete set of MUBs. A closer inspection of (\[newGalois\]) reveals an intricate relation MUBs and quantum phase operators. It is known that the Fourier basis can be derived in terms the of a phase with $\theta_k$ and given $\Theta_d=\sum_{k=0}^{d-1}\theta_k|\theta_k\rangle\langle \theta_k|$. Similarly, using well known properties of the field trace, one can show that each base index $a$ associated with quantum operator \label{quantumphaseop}$$ with eigenvectors eigenvalues $\theta_b^a$; the latter may thus “MUB operator". MUBs FOR EVEN CHARACTERISTIC FROM GALOIS {#MUBsQubits} ============================================== next goal is to find a transform formulation of MUBs in characteristic $2$. This a very important case since the ‘$2^m-$dits’ are the basic information units in quantum computation. be tempted to connect Galois field algebra the Pauli (\[Pauli\]) constructing discrete spaces over the Galois field [@Wootters04bis]. For the one qubit case already know that the eigenvectors of Pauli matrices $\sigma_z$, $\sigma_x$ | [@Wootters89], the reason why (\[newGaLois\]) defineS the cOmpLetE sEt of mUBs Relies on the fieLD theOretical formula $|\sum_{n=0}^{d-1}\omEga_p^{tR[(a
\ODot n \OPlUs b)\odOt n ]}|n\ranGLe|=P^{1/2}$, WIth $ A \nEq 0$ And $P$ bEInG an odD prIme. This Method, howeVer, FaIls for characTErIstic two whEre $|\Sum_{n=0}^{d-1}\omega_2^{tr[(A \odOt n
\oplUs B)\odOT n ]}|n\raNglE|=0$ for aNy $a$, $b$. As SHown in sect. 3 below, FoR This chARacteriSTIc One hAs to use Galois ringS InSTead of Galois fiElds to GeT A cOMPleTe sEt of MUBs.
A cLoSer inSPection OF (\[nEWgAloIS\]) reveals an intRicate relatIOn bEtween mUbs aND quantUm phaSe OPerAtors. It is knOwn [@PLanat04] that The FouRIer basiS $|\Theta_k\rAngle$ cAn bE deRiveD In TeRms Of THe eIGeNveCTorS of a quanTuM pHase oPeraTOR WIth eIgeNvalUes $\thEta_k$ and given bY $\ThEta_d=\SUm_{k=0}^{D-1}\thetA_k|\theTa_k\rAnGle\laNgle
\thEta_k|$. SImIlarly, using well KnowN propertiEs oF tHe fIeLd traCE, one caN shOw tHat each Base of iNDex $A$ cAN BE aSsociated with a quanTuM PHaSe operatOr $$\ThetA_D^a=\SuM_{B=0}^{d-1}\theta_b^A|\tHetA_b^a\rANGle\laNgle\THeTa_b^a|,
\labeL{quantUMpHaSeop}$$ witH eIgenveCtOrs $|\TheTa_b^a\rANgle$ And eigEnvalues $\Theta_B^A$; the latter may tHUs be called an “Mub oPERaTOr".
MUbs FoR EVEN CHARAcTERisTIC fROM gaLoIS riNGS {#MuBsQuBiTS}
==============================================
OUR next goal is to find a FOuRier trAnsfoRm formulation Of MUBs in chARACteristiC $2$. ThiS Is A Very important cAse siNce the ‘$2^m-$ditS’ Are the baSic inFormatioN units in qUANtum compUtaTioN. OnE maY BE tEmpted to conneCT The GAlOis fielD alGebra wiTh tHe gEneRalIzEd Pauli opErators (\[PAuLi\]) By CoNstRuctiNG discretE vEctOr SpaCes ovER the GaLois fIeld [@woOtTErs04Bis]. For tHE oNE QubiT cAsE we aLreAdY know That THe eIgenvecTors of PauLi mATricEs $\SiGma_z$, $\sigMa_x$ and $\sigma_x | [@Wootters89], the reason why (\[ne wGalo is\ ])de fine s th e complete set of M UBs relies on the fiel d the or e tica l f ormul a $|\su m _{ n = 0}^ {d -1 }\o me g a_ p^{tr [(a
\odotn \oplus b )\o do t n ]}|n\ran g le |=p^{1/2}$ , w ith $ a \neq 0$ and $ p$ be i ng an od d pri me. Th i s meth od, howev er , fails for cha r a ct eris tic two where $|\ s um _ {n=0}^{d-1}\om ega_2^ {t r [( a \od otn
\oplus b )\ odotn ]}|n\r a ng l e | =0$ for any $a$,$b$. As sho w n i n Sect .3 b e low, f or th is cha racteristic one has to u se Gal o is ring s instea d of G alo isfiel d sto ge ta co m pl ete set of MUBs .
A clos er i n s p e ctio n o f (\ [newG alois\]) reve als ani ntr icate rela tion b etwee n MUBs andqu antum phase ope rato rs. It is kn ow n [ @P lanat 0 4] tha t t heFourier basis$ |\t he t a _ k\ rangle$ can be der iv e d i n termsof the ei ge n vectorsof aquan t u m pha se o p er ator wit h eige n va lu es $\th et a_k$ a nd gi ven by $ \ Thet a_d=\s um_{k=0} ^{d-1 } \theta_k|\thet a _k\rangle\lan g le \ th e ta_k |$. Similarly, usi n g we ll k n ow n p r opert ies o ft he field trace, one ca nshow t hat e ach base of i ndex $a$ c a n be assoc iate d w i th a quantum p haseoperator $ $ \Theta_d ^a=\s um_{b=0} ^{d-1}\th e t a_b^a|\t het a_b ^a\ ran g l e\ langle\theta_ b ^ a|,\l abel{qu ant umphase op} $$wit h e ig envectors $|\thet a_ b^ a\ ra ngl e$ an d eigenva lu es$\ the ta_b^ a $; the latt er m ay t h usbe call e da n “MU Bop erat or" .
MUBs FOR EVE N CHARA CTERISTIC FR O M GA LO IS RINGS{#MUBsQubits}
= ========== == === ====== = = ======== ==============
Our nex t goal i s t o fin d aFourier t ran sformfor m ulatio n of M UBs i ncha r a cteri s t ic $2 $. This is a v ery impo rt antcase si nce the ‘$2^m-$dit s ’ a re the basicinf orma t i on un i ts inqu a ntu m computation. On e may be t em p te d to conne c t t he Galois fieldalgeb r a withthe gener alized Pa ul i op e r ato rs (\[Paul i\]) byconstruct i ng di s cr ete v ect or spa ce s o ver t he Gal o isfield [@Woo tt ers04b is].Fo r the on e qubit case we already knowthatthe eigenvec tor s of Pauli ma tric es $\sigma _z$ , $ \sigm a_x $ and$\si g ma _x | [@Wootters89],_the reason_why (\[newGalois\]) defines the_complete set_of_MUBs relies_on_the field theoretical_formula $|\sum_{n=0}^{d-1}\omega_p^{tr[(a
\odot n_\oplus b)\odot n ]}|n\rangle|=p^{1/2}$,_with $ a_\neq_0$ and $p$ being an odd prime. This method, however, fails for characteristic two_where_$|\sum_{n=0}^{d-1}\omega_2^{tr[(a \odot_n
\oplus_b)\odot_n ]}|n\rangle|=0$ for any $a$,_$b$. As shown in Sect._3 below,_for this characteristic one has to use Galois_rings_instead of Galois_fields to get a complete set of MUBs.
A closer_inspection of (\[newGalois\]) reveals an intricate_relation between MUBs_and_quantum_phase operators. It is_known [@Planat04] that the Fourier basis_$|\theta_k\rangle$ can be derived in terms_of the eigenvectors of a quantum phase_operator with eigenvalues $\theta_k$ and given_by $\Theta_d=\sum_{k=0}^{d-1}\theta_k|\theta_k\rangle\langle
\theta_k|$. Similarly, using well_known properties_of the field trace, one_can show that_each base_of index $a$_can be associated with a quantum_phase operator $$\Theta_d^a=\sum_{b=0}^{d-1}\theta_b^a|\theta_b^a\rangle\langle\theta_b^a|,
\label{quantumphaseop}$$_with eigenvectors $|\theta_b^a\rangle$ and eigenvalues $\theta_b^a$;_the_latter may thus_be_called_an “MUB_operator".
MUBs FOR EVEN_CHARACTERISTIC_FROM GALOIS_RINGS_{#MUBsQubits}
==============================================
Our next goal is to find_a_Fourier transform formulation of MUBs in characteristic_$2$. This is a_very_important case since the_‘$2^m-$dits’ are the basic information_units in quantum computation. One may_be tempted_to connect_the Galois field algebra with the generalized Pauli operators (\[Pauli\]) by_constructing discrete vector spaces over the_Galois field [@Wootters04bis]. For_the one_qubit_case we already_know_that the_eigenvectors of Pauli matrices $\sigma_z$, $\sigma_x$ and_$\sigma_x |
populations in modes $i$ and $j$ are well known [@armstrong]. For a resonant two-state model the characteristic coupling time is $T=\left|\sqrt{2}\hbar/g M_{ij}b^{r}_{i}(0)\right|$. In terms of the dimensionless matrix elements $m^{q}_{ij}$ this can be written; $$T=\left|\frac{15\lambda^{-\frac{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0)
\omega_{r}}\left(\frac{15C}{64\pi}\right)^{-3/5} \right|$$ For $\lambda \approx 1.95$, and $C=1000$ the quantal prediction is $|m^{q}_{12}|$=0.013 compared with the hydrodynamic result [@oxfthy] $|m^{h}_{12}|$=0.028 (the dotted line in figure \[fig:coup\]). Suppose that only 5% of atoms were initially in the pump mode, then $|b^{r}_{1}(0)|^{2}$=0.05. It follows, for $C=1000$ and $\lambda=1.95$, that the respective coupling times are $T^q$=14.60 ms compared with $T^{h}$=6.78 ms, which is a substantial and measurable difference. However two excited hybrid modes are present near second-harmonic resonance. A numerical solution of equations (\[motion1\],\[motion2\]) for a pump population of 10% at $\lambda=2$ shows the interplay of the hybrid modes: figure \[fig:conversion\] (a). The nonresonant surface hybrid grows faster due to the larger coupling strength, but is not as efficiently converted as the resonant breather hybrid state. The fundamental mode revives [@nonlin_mix] after $\sim 40$ms for these parameters. Since the two hybrid modes differ in frequency and symmetry this phenomenon should be detectable by observation of the radial and axial density variations as developed in current experiments [@oxfexp]. In figure \[fig:conversion\] (b) corresponding to $\lambda=1.65$, very low conversion efficiency is observed due to poor phase-matching. Moreover, in this case the dark-state hybrid (short-dashed line) is completely suppressed.
In conclusion, we find the hybridization of Bogoliubov modes plays a vital role in mode mixing, both in the coupling strength and resonant frequency. Another consequence is the | populations in modes $ i$ and $ j$ are well known [ @armstrong ]. For a resonant two - state exemplar the characteristic yoke time is $ T=\left|\sqrt{2}\hbar / g M_{ij}b^{r}_{i}(0)\right|$. In terms of the dimensionless matrix elements $ m^{q}_{ij}$ this can be publish; $ $ T=\left|\frac{15\lambda^{-\frac{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0)
\omega_{r}}\left(\frac{15C}{64\pi}\right)^{-3/5 } \right|$$ For $ \lambda \approx 1.95 $, and $ C=1000 $ the quantal prediction is $ |m^{q}_{12}|$=0.013 compare with the hydrodynamic result [ @oxfthy ] $ |m^{h}_{12}|$=0.028 (the dotted tune in number \[fig: coup\ ]). Suppose that merely 5% of atoms were initially in the pump mood, then $ |b^{r}_{1}(0)|^{2}$=0.05. It follows, for $ C=1000 $ and $ \lambda=1.95 $, that the respective coupling time are $ T^q$=14.60 ms compared with $ T^{h}$=6.78 ms, which is a substantial and measurable difference. However two excited loanblend modes are present near second - harmonic resonance. A numeral solution of equations (\[motion1\],\[motion2\ ]) for a pump population of 10% at $ \lambda=2 $ shows the interplay of the loanblend modes: design \[fig: conversion\ ] (a). The nonresonant airfoil hybrid grows faster ascribable to the larger coupling strength, but is not as efficiently converted as the resonant breath hybrid state. The fundamental mode revives [ @nonlin_mix ] after $ \sim 40$ms for these parameters. Since the two hybrid modes differ in frequency and symmetry this phenomenon should be detectable by observation of the radial and axial concentration mutant as developed in current experiment [ @oxfexp ]. In figure \[fig: conversion\ ] (b) corresponding to $ \lambda=1.65 $, very low conversion efficiency is watch due to poor phase - matching. Moreover, in this case the dark - department of state hybrid (short - dashed line) is completely suppressed.
In ending, we find the hybridization of Bogoliubov modes plays a vital role in mode mix, both in the yoke lastingness and resonant frequency. Another consequence is the | pooulations in modes $i$ and $j$ are well knocb [@armsvrong]. Fkr a resunant two-state model the chacactwristuc coupling time is $T=\ldft|\sqrt{2}\hbwr/g M_{ij}b^{e}_{i}(0)\rijht|$. In terms of vge dimekfionmcss mctcix elements $m^{q}_{lj}$ this can be written; $$T=\lafg|\fxac{15\lambda^{-\frac{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0)
\omega_{r}}\left(\fras{15C}{64\pi}\rignt)^{-3/5} \right|$$ For $\lamfda \s[pros 1.95$, and $C=1000$ the quantal prediction is $|m^{q}_{12}|$=0.013 comkared with the hydtodynamic result [@oxfthy] $|m^{h}_{12}|$=0.028 (the dotted line in fihure \[fig:couk\]). Su[pose that onuy 5% of atoms were initjally in the pump mode, then $|b^{r}_{1}(0)|^{2}$=0.05. Kt foklows, for $X=1000$ qnd $\nambda=1.95$, that the gespective comiling thmes arr $T^q$=14.60 ms comparcd wivh $T^{y}$=6.78 ms, which is a substential and measurablg differenwe. However two exciree hybtid mmdes qre prtseit hear sfcoid-harmonic desonance. A numerical solution os equations (\[mofion1\],\[moeijn2\]) for a pump population of 10% at $\lambda=2$ vhoss the interplay of the hybrid modes: figure \[vig:converfion\] (a). The nonresonant surface hybrid grows fastes due go uhc uqrher coupling strength, but is not as efficientjg vokverted as the rcsonant breather hubgic state. The fuvdamenczl mode revives [@nonlln_mix] aster $\wim 40$ms fow thrse parameters. Since the twi hybrid modvs duffer in frequency and symmetxy thix phemomenon should be deteccable gy observatlon of ths radial and axiau dvnsidy variaukons as developed in curreit ex'erimentr [@oxgexp]. Ig figure \[flg:conyarsion\] (b) corresponfing jo $\lamtda=1.65$, very llw conversion efficiency is obsxcved due to ppos pvase-matcking. Mpreover, in thys case the datk-state hvbrid (rhort-dashes line) ms completelr suppressed.
It conclusion, xe find tre htbrieizatiov of Bogoliubov modes plcvs a vitao role in mode mixlng, bufh in the couplnug strength and rrsovane vrxquensf. Another cotseqjenzr is ghe | populations in modes $i$ and $j$ are [@armstrong]. a resonant model the characteristic In of the dimensionless elements $m^{q}_{ij}$ this be written; $$T=\left|\frac{15\lambda^{-\frac{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0) \omega_{r}}\left(\frac{15C}{64\pi}\right)^{-3/5} \right|$$ $\lambda \approx 1.95$, and $C=1000$ the quantal prediction is $|m^{q}_{12}|$=0.013 compared with the result [@oxfthy] $|m^{h}_{12}|$=0.028 (the dotted line in figure \[fig:coup\]). Suppose that only 5% atoms initially the mode, then $|b^{r}_{1}(0)|^{2}$=0.05. It follows, for $C=1000$ and $\lambda=1.95$, that the respective coupling times are $T^q$=14.60 compared with $T^{h}$=6.78 ms, which is a substantial measurable difference. However two hybrid modes are present near resonance. numerical solution equations for pump population of at $\lambda=2$ shows the interplay of the hybrid modes: figure \[fig:conversion\] (a). The nonresonant surface hybrid grows due to coupling strength, is as converted as the hybrid state. The fundamental mode revives 40$ms for these parameters. Since the two hybrid differ in and symmetry this phenomenon should be by observation of the radial and axial density as developed in current experiments [@oxfexp]. In figure \[fig:conversion\] (b) corresponding to $\lambda=1.65$, very low is observed due to phase-matching. Moreover, in case dark-state (short-dashed is completely In conclusion, we find the hybridization of Bogoliubov modes plays a role in mode mixing, both in the coupling strength and Another is the | populations in modes $i$ and $j$ arE well known [@ArmstRonG]. FoR a ResoNant Two-state model tHE chaRacteristic coupling timE is $T=\lEfT|\Sqrt{2}\HBaR/g M_{ij}B^{r}_{i}(0)\righT|$. in TERms Of ThE diMeNSiOnlesS maTrix eleMents $m^{q}_{ij}$ tHis CaN be written; $$T=\lEFt|\Frac{15\lambda^{-\FraC{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0)
\omEga_{R}}\left(\fRaC{15C}{64\pI}\Right)^{-3/5} \RigHt|$$ For $\Lambda \APprox 1.95$, aNd $C=1000$ the quaNtAL prediCTion is $|m^{Q}_{12}|$=0.013 COmPareD with the hydrodynaMIc REsult [@oxfthy] $|m^{h}_{12}|$=0.028 (tHe dottEd LInE IN fiGurE \[fig:coup\]). SuPpOse thAT only 5% of AToMS WEre INitially in the Pump mode, theN $|B^{r}_{1}(0)|^{2}$=0.05. IT folloWs, For $c=1000$ And $\lamBda=1.95$, thAt THe rEspective coUpliNg times arE $T^q$=14.60 ms cOMpared wITh $T^{h}$=6.78 ms, wHich is A suBstAntiAL aNd MeaSuRAblE DiFfeREncE. However TwO eXciteD hybRID MOdes Are PresEnt neAr second-harmoNic ResoNAncE. A numEricaL solUtIon of EquatiOns (\[moTiOn1\],\[motion2\]) for a pumP popUlation of 10% At $\lAmBda=2$ ShOws thE InterpLay Of tHe hybriD modes: fIGurE \[fIG:COnVersion\] (a). The nonresoNaNT SuRface hybRid groWS fAsTEr due to tHe LarGer cOUPling StreNGtH, but is noT as effICiEnTly convErTed as tHe ResOnaNt breATher Hybrid State. The FundaMEntal mode revivES [@nonlin_mix] aftER $\sIM 40$Ms FOr thEse Parameters. SInce THe twO hybRId ModES diffEr in fReQUeNCy and symmetry this phEnOmenon ShoulD be detectable By observatION Of the radIal aND aXIal density variAtionS as developED in curreNt expEriments [@Oxfexp]. In fIGUre \[fig:coNveRsiOn\] (b) CorRESpOnding to $\lambdA=1.65$, VEry lOw ConversIon EfficieNcy Is oBseRveD dUe to poor pHase-matcHiNg. moReOveR, in thIS case the DaRk-sTaTe hYbrid (SHort-daShed lIne) iS cOmPLetEly suppREsSED.
In cOnClUsioN, we FiNd the HybrIDizAtion of bogoliuboV moDEs plAyS a Vital roLe in mode mixinG, bOth in the coUpLinG strenGTH and resoNant frequency. Another conSEquence Is tHe | populations in modes $i$and $j$ ar e wel l k now n[@ar mstr ong]. For a re s onan t two-state model thechara ct e rist i ccoupl ing tim e i s $T= \l ef t|\ sq r t{ 2}\hb ar/ g M_{ij }b^{r}_{i} (0) \r ight|$. In t e rm s of the d ime nsionless ma tri x elem en ts$ m^{q} _{i j}$ t his ca n be wr itten; $$ T= \ left|\ f rac{15\ l a mb da^{ -\frac{13}{30}}}{ 8 \p i m^{q}_{ij}b^{ r}_{i} (0 )
\ o m ega _{r }}\left(\f ra c{15C } {64\pi} \ ri g h t )^{ - 3/5} \right|$ $ For $\lam b da\appro x1.9 5 $, and $C=1 00 0 $ t he quantalpred iction is $|m^{ q }_{12}| $ =0.013compar edwit h th e h yd rod yn a mic re sul t [@ oxfthy]$| m^ {h}_{ 12}| $ = 0 . 028(th e do ttedline in figur e \ [fig : cou p\]). Supp oseth at on ly 5%of at om s were initiall y in the pump mo de , t he n $|b ^ {r}_{1 }(0 )|^ {2}$=0. 05. Itf oll ow s , fo r $C=1000$ and $\l am b d a= 1.95$, t hat th e r es p ective c ou pli ng t i m es ar e $T ^ q$ =14.60 m s comp a re dwith $T ^{ h}$=6. 78 ms , w hichi s asubsta ntial an d mea s urable differe n ce. However t w oe x ci t ed h ybr id modes ar e pr e sent nea r s eco n d-har monic r e so n ance. A numerical s ol utionof eq uations (\[mo tion1\],\[ m o t ion2\])fora p u mp populationof 10 % at $\lam b da=2$ sh ows t he inter play of t h e hybridmod es: fi gur e \[ fig:conversio n \ ] (a ). The no nre sonantsur fac e h ybr id grows fa ster due t oth elar ger c o upling s tr eng th , b ut is not as effi cien tl yc onv erted a s t h e res on an t br eat he r hyb rids tat e. Thefundament alm odere vi ves [@n onlin_mix] af te r $\sim 40 $m s f or the s e paramet ers. Since the two hybr i d modes di fferin f requencyand symme try this p henome non s ho uld b e det e c ta ble b y observat i o n o f the r adia l and a xial density varia t ion s as develope d i n cu r r en t e x pe r ime nt s [@ o x fexp]. In figur e \[fig:co nv e rs ion\] (b)c orr es ponding to $\l ambda = 1.65$,very lowconversio neffi c i enc y is obser ved dueto poor p h ase-m a tc hing. Mo reover ,inthiscase t h e d ark-s tate h yb rid (s hort- da shed lin e) is completely suppre ssed.
In c onc lusion, w e f i ndthe hybri diza tion of Bo gol iub ov mo des plays a v i ta l r o le in mod e mixing,b ot h i n th e couplings t r eng th an d r e sonant fre quency. Another c o nsequence is t he | populations_in modes_$i$ and $j$ are_well known_[@armstrong]._For a_resonant_two-state model the_characteristic coupling time_is $T=\left|\sqrt{2}\hbar/g M_{ij}b^{r}_{i}(0)\right|$. In_terms of the_dimensionless_matrix elements $m^{q}_{ij}$ this can be written; $$T=\left|\frac{15\lambda^{-\frac{13}{30}}}{8\pi m^{q}_{ij}b^{r}_{i}(0)
\omega_{r}}\left(\frac{15C}{64\pi}\right)^{-3/5} \right|$$ For $\lambda \approx 1.95$,_and_$C=1000$ the_quantal_prediction_is $|m^{q}_{12}|$=0.013 compared with the_hydrodynamic result [@oxfthy] $|m^{h}_{12}|$=0.028 (the_dotted line_in figure \[fig:coup\]). Suppose that only 5% of_atoms_were initially in_the pump mode, then $|b^{r}_{1}(0)|^{2}$=0.05. It follows, for $C=1000$_and $\lambda=1.95$, that the respective coupling_times are $T^q$=14.60_ms_compared_with $T^{h}$=6.78 ms, which_is a substantial and measurable difference._However two excited hybrid modes are_present near second-harmonic resonance. A numerical solution_of equations (\[motion1\],\[motion2\]) for a pump_population of 10% at $\lambda=2$_shows the_interplay of the hybrid modes:_figure \[fig:conversion\] (a)._The nonresonant_surface hybrid grows_faster due to the larger coupling_strength, but is_not as efficiently converted as the_resonant_breather hybrid state._The_fundamental_mode revives_[@nonlin_mix] after $\sim_40$ms_for these_parameters._Since the two hybrid modes differ_in_frequency and symmetry this phenomenon should be_detectable by observation of_the_radial and axial density_variations as developed in current_experiments [@oxfexp]. In figure \[fig:conversion\] (b)_corresponding to_$\lambda=1.65$, very_low conversion efficiency is observed due to poor phase-matching. Moreover, in_this case the dark-state hybrid (short-dashed_line) is completely suppressed.
In_conclusion, we_find_the hybridization of_Bogoliubov_modes plays_a vital role in mode mixing, both_in the_coupling strength and resonant frequency. Another_consequence is the |
have presented a new method to transform and solve numerically the equations of motion of a scalar field as a dark matter component. This method takes advantage of the explicit symmetries of the coupled Einstein-Klein-Gordon system of equations, and allow a clear understanding of the underlying physical solutions of the scalar field. The method is also suitable for numerical purposes, and deals correctly with the rapid oscillations of the field without any pre-handling of the equations of motion.
We showed that the background dynamics of SFDM is very simple in terms of the new dynamical variables, and even eases its numerical implementation in CMB codes, as exemplified with the results obtained from CLASS. Moreover, the transformation of variables used for the background case was easily adapted for the evolution of linear perturbations in the scalar field, and there was no need to rely on any similarities with standard fluid dynamics. Under this approach, we were able to find semi-analytical solutions which show in a clear way the physical effects acting in the growing or suppression of linear perturbations. The amended version of CLASS will be used in future studies to determine the direct effects of scalar fields in different observables, specially on those that make the CMB spectrum distinguishable from that of $\Lambda$CDM (see Fig. \[fig:3\]), for both types of dark matter and dark energy fields. Another topic of interest is the setting up of initial conditions of $N$-body and wave-like simulations, so that one can incorporate constraints for scalar field models from linear and non-linear perturbations.
Our results indicate that the growing of perturbations in SFDM models for a given wave number $k$ requires of two conditions: a) an early onset of field oscillations, preferably much earlier than the time of radiation-matter equality, and b) a negligible ratio $k/k_J$. Not surprisingly, both conditions depend upon the only free physical parameter which is the mass of the SFDM particle, and actually the two conditions are simultaneously enhanced (suppressed) in the case of large (small) values of $m$ just because both angular variables $\theta$ and $\vartheta$ evolve similarly throughout.
As for condition b), we must recall first that the scales that have been affected at some degree by the Jeans wavenumber $k_J$, throughout the evolution of the Universe, are those for which initially $k/k_{J,eq} > 1$ with $k_{J,eq} = \Omega^{1/4}_{r0} k_{J0}$, see | have presented a new method to transform and resolve numerically the equation of motion of a scalar field as a dark topic component. This method takes advantage of the denotative symmetry of the coupled Einstein - Klein - Gordon system of equations, and allow a absolved understanding of the underlying forcible solutions of the scalar field. The method is also suitable for numerical purposes, and deal correctly with the rapid oscillations of the field without any pre - handling of the equation of motion.
We showed that the background moral force of SFDM is very simple in term of the new dynamical variables, and even facilitate its numerical implementation in CMB codes, as exemplified with the results obtained from CLASS. Moreover, the transformation of variables used for the background case was easily adapted for the evolution of linear perturbations in the scalar field, and there embody no need to rely on any similarity with standard fluid moral force. Under this overture, we were able to find semi - analytical solutions which show in a clear way the physical effect acting in the growing or suppression of linear perturbations. The better version of CLASS will be used in future studies to settle the direct effect of scalar fields in different observables, specially on those that make the CMB spectrum distinguishable from that of $ \Lambda$CDM (determine Fig. \[fig:3\ ]), for both types of dark matter and blue energy fields. Another topic of interest is the mount up of initial conditions of $ N$-body and wave - comparable simulations, so that one can incorporate constraint for scalar field model from linear and non - linear disruption.
Our results argue that the growing of perturbations in SFDM models for a given wave number $ k$ want of two conditions: a) an early onset of field oscillations, preferably much earlier than the time of radiation - matter equality, and b) a negligible proportion $ k / k_J$. Not surprisingly, both circumstance depend upon the only free physical parameter which is the mass of the SFDM particle, and actually the two condition are simultaneously enhanced (suppressed) in the case of large (small) values of $ m$ merely because both angular variable star $ \theta$ and $ \vartheta$ evolve similarly passim.
As for circumstance b), we must recall first that the scale that have been affected at some degree by the Jeans wavenumber $ k_J$, throughout the development of the Universe, are those for which initially $ k / k_{J, eq } > 1 $ with $ k_{J, eq } = \Omega^{1/4}_{r0 } k_{J0}$, see | hage presented a new methoa to transform cbd solte numedically ghe equations of motion of a sxalar field as a dark mattef componejt. This nethie takes adtzntage of the cxplieiv symmetries of the couplad Einstein-Klehn-Eoxdon system of equations, and allow a clear inferstanding of the tndedlying physical solutions of the sdalar fpeld. The method ix also suitable for numerifal ourposes, and deals correctly qith rhe rapid oszillations of the fiels without any pre-handling of thd equctions of mituon.
Ag showed thav the fackground dnmamics of SFDK is very simpke mn twrms of the new dynammcal variables, and eden eases ics numerical implemenrarion hn CKB cuees, as eeemllifief wmth the reshlts obtainwd from CLASS. Moreofew, the transforjation os variables used for the background cast was easily adapted for the evolution of linear kerturbatijns in the scalar field, and there was no need to sely kv aun rumllarities with standard fluid dynamics. Under egix spproach, we weve able to find seki-wnsjytical solutkons wkjcg show in a clear aay the physucal effests scting in the growing or suppression of oinear perturbatious. The amendzd verxion pf CLASS will be used iu futude studies ho determjve the direct efwecns ox scalar fields in differegt observebles, speciauly pn thofe that mane thc CMB spectrum distlnguidhdble from hhat of $\Lambda$CDM (see Fig. \[fig:3\]), foc both types pf dagk matter and cark energy fyelds. Another jopic of nnterert is the aetting up of inityal conditionv of $N$-body anv wave-likq sinularions, su that one can oncorporane conwtraints for scalav fieus models from lnueqr and non-lineat pdrttrhavions.
Jgr results itdicxte yhat ghe growing uf prrturbations in SFDM modsls for a given wafe number $j$ requirqs of two concitions: a) an early onsev of fmeld oxcijlations, preferably much earlisr than tje bime of radiaeion-natter equalnty, and b) a negligible ratio $k/k_J$. Not sur'risingly, both conditiobs depend upon the pkly free phyxical parametes which is the mass if the SFDM partigle, and actually the tso congitiojs are simultaneously enhanced (suppressed) in the case of large (small) valuws of $m$ just becauae bpth atgblax variafles $\vheta$ and $\varthets$ evolve similarly throughout.
As for congicion b), we must recall first tnag the scales jhat have been affected at some debree by the Jeans wavenumber $k_J$, throughout the evolutiok of the Ynivedse, arv tiose for which initislly $k/n_{J,eq} > 1$ with $j_{J,eq} = \Omega^{1/4}_{r0} k_{J0}$, sde | have presented a new method to transform numerically equations of of a scalar component. method takes advantage the explicit symmetries the coupled Einstein-Klein-Gordon system of equations, allow a clear understanding of the underlying physical solutions of the scalar field. method is also suitable for numerical purposes, and deals correctly with the rapid of field any of the equations of motion. We showed that the background dynamics of SFDM is very simple terms of the new dynamical variables, and even its numerical implementation in codes, as exemplified with the obtained CLASS. Moreover, transformation variables for the background was easily adapted for the evolution of linear perturbations in the scalar field, and there was no to rely similarities with fluid Under approach, we were find semi-analytical solutions which show in the physical effects acting in the growing or of linear The amended version of CLASS will used in future studies to determine the direct of scalar fields in different observables, specially on those that make the CMB spectrum distinguishable of $\Lambda$CDM (see Fig. for both types dark and energy Another topic interest is the setting up of initial conditions of $N$-body and simulations, so that one can incorporate constraints for scalar field linear non-linear perturbations. Our indicate that the growing perturbations SFDM models for a number requires a) early of field oscillations, preferably earlier than the time of equality, and b) a both conditions depend upon the only free physical which is the mass of the SFDM and actually the two conditions are simultaneously enhanced (suppressed) in the case large (small) $m$ just because both angular variables $\theta$ and evolve similarly throughout. As condition b), we must recall first that the scales have affected at degree by the wavenumber $k_J$, throughout evolution of the those for initially > = \Omega^{1/4}_{r0} k_{J0}$, see | have presented a new method to Transform aNd solVe nUmeRiCallY the Equations of motIOn of A scalar field as a dark matTer coMpONent. tHiS methOd takes ADvANTagE oF tHe eXpLIcIt symMetRies of tHe coupled EInsTeIn-Klein-GordoN SyStem of equaTioNs, and allow a cLeaR underStAndINg of tHe uNderlYing phYSical sOlutions oF tHE scalaR Field. ThE MEtHod iS also suitable for nUMeRIcal purposes, anD deals CoRReCTLy wIth The rapid osCiLlatiONs of the FIeLD WIthOUt any pre-handlIng of the equATioNs of moTiOn.
WE Showed That tHe BAckGround dynamIcs oF SFDM is veRy simpLE in termS Of the neW dynamIcaL vaRiabLEs, AnD evEn EAseS ItS nuMEriCal impleMeNtAtion In CMb CODEs, as ExeMpliFied wIth the results ObtAineD FroM CLASs. MoreOver, ThE tranSformaTion oF vAriables used for The bAckground CasE wAs eAsIly adAPted foR thE evOlution Of lineaR PerTuRBATiOns in the scalar fielD, aND ThEre was no Need to RElY oN Any similArItiEs wiTH StandArd fLUiD dynamicS. Under THiS aPproach, We Were abLe To fInd Semi-aNAlytIcal soLutions wHich sHOw in a clear way tHE physical effeCTs ACTiNG in tHe gRowing or supPresSIon oF linEAr PerTUrbatIons. THe AMeNDed version of CLASS wiLl Be used In futUre studies to dEtermine thE DIRect effeCts oF ScALar fields in difFerenT observablES, specialLy on tHose that Make the CMb SPectrum dIstIngUisHabLE FrOm that of $\LambdA$cdM (seE FIg. \[fig:3\]), foR boTh types Of dArk MatTer AnD dark enerGy fields. anOtHeR tOpiC of inTErest is tHe SetTiNg uP of inITial coNditiOns oF $N$-BoDY anD wave-liKE sIMUlatIoNs, So thAt oNe Can inCorpORatE constrAints for sCalAR fieLd MoDels froM linear and non-LiNear perturBaTioNs.
Our rESUlts indiCate that the growing of perTUrbatioNs iN SFDM ModeLs for a givEn wAve numBer $K$ RequirEs of twO condItIonS: A) An earLY OnSet Of Field oscilLATioNs, preFeRablY much eaRlier than the time of RAdiAtion-matter eqUalIty, aND B) a NegLIgIBle RaTIo $k/K_j$. not surprisingly, Both conditIoNS dEpend upon tHE onLy Free phySical paRametER which iS the mass oF the SFDM pArTiclE, ANd aCtually the Two condiTions are sIMultaNEoUsly eNhaNced (suPpResSed) in The casE Of lArge (sMall) vaLuEs of $m$ jUst beCaUse both aNgular variables $\theta$ and $\VartheTa$ evoLve Similarly ThrOUghOut.
As for cOndiTion b), we musT reCalL firsT thAT the sCaleS ThAt hAVe beeN affECted at somE DeGreE BY tHe Jeans waveNUMBer $K_J$, thrOugHOut the EvolUtion of the UniversE, Are those for whiCh inITIalLy $k/K_{j,eq} > 1$ wItH $k_{J,eq} = \Omega^{1/4}_{r0} k_{J0}$, sEe | have presented a new meth od to tran sform an d s ol ve n umer ically the equ a tion s of motion of a scala r fie ld as a da rk ma tter co m po n e nt. T hi s m et h od take s a dvantag e of the e xpl ic it symmetrie s o f the coup led Einstein-Kl ein -Gordo nsys t em of eq uatio ns, an d allow a clearun d erstan d ing oft h eunde rlying physical s o lu t ions of the sc alar f ie l d. T hemet hod is als osuita b le forn um e r i cal purposes, and deals corr e ctl y with t her apid o scill at i ons of the fie ld w ithout an y pre- h andling of theequati ons of mot i on .
We s h owe d t hat the backgro un ddynam icso f S FDMisvery simp le in terms o f t he n e w d ynami cal v aria bl es, a nd eve n eas es its numericalimpl ementatio n i nCMB c odes, as exe mpl ifi ed with the re s ult so b t ai ned from CLASS. Mo re o v er , the tr ansfor m at io n of vari ab les use d for t he b a ck ground c ase wa s e as ily ada pt ed for t heevo lutio n oflinear perturb ation s in the scalar field, and th e re w as no n eed to rely on any simi lari t ie s w i th st andar df lu i d dynamics. Under t hi s appr oach, we were able to find s e m i -analyti cals ol u tions which sh ow in a clear w a y the ph ysica l effect s actingi n the gro win g o r s upp r e ss ion of linear p ertu rb ations. Th e amend edver sio n o fCLASS wil l be use din f ut ure stud i es to de te rmi ne th e dir e ct eff ectsof s ca la r fi elds in di f f eren tob serv abl es , spe cial l y o n those that mak e t h e CM Bsp ectrumdistinguishab le from that o f $ \Lambd a $ CDM (see Fig. \[fig:3\]), for b o th type s o f dar k ma tter anddar k ener gyf ields. Anoth er to pi c o f inter e s tisth e settingu p of init ia l co ndition s of $N$-body andw ave -like simulat ion s, s o th ato ne can i n cor p o rate constraint s for scal ar fi eld models fro mlinearand non -line a r pertu rbations.
Our res ul ts i n d ica te that th e growin g of pert u rbati o ns in S FDM model sfor a gi ven wa v e n umber $k$ r eq uiresof tw oconditio ns: a) an early onset o f fiel d osc ill ations, p ref e rab ly much e arli er than th e t ime of r adi a tion- matt e requ a lity, and b) a negl i gi ble r at io $k/k_J$. N o t s urpri sin g ly, bo th c onditions dependu pon the only f reep h ysi cal para me ter which is t hema s s of theSF DM particle , and ac tu a lly t he two condi tions a r e s i multan eous lyenhanced(su pp r essed)in t h e case ofla rge (s mall)v alue s of $m$ just beca use b o t h ang u lar vari ab les $\t h eta$ and $\var theta$ evol ve sim ilar ly th roughou t.
As f orco ndition b) , we mustrecal l first t hatthe scale s th a t have bee naff ected ats o me de gr e e b y th e Jea ns wav enumber $ k _J$, thr oug h out the e vol u t ion of th e Universe,are thos e for whichi niti a ll y $k/k _{J,eq } > 1$ with $ k _{J ,e q} = \O meg a ^ {1/4}_{r0 } k_{J0}$ , se e | have_presented a_new method to transform_and solve_numerically_the equations_of_motion of a_scalar field as_a dark matter component._This method takes_advantage_of the explicit symmetries of the coupled Einstein-Klein-Gordon system of equations, and allow a_clear_understanding of_the_underlying_physical solutions of the scalar_field. The method is also_suitable for_numerical purposes, and deals correctly with the rapid_oscillations_of the field_without any pre-handling of the equations of motion.
We showed_that the background dynamics of SFDM_is very simple_in_terms_of the new dynamical_variables, and even eases its numerical_implementation in CMB codes, as exemplified_with the results obtained from CLASS. Moreover,_the transformation of variables used for_the background case was easily_adapted for_the evolution of linear perturbations_in the scalar_field, and_there was no_need to rely on any similarities_with standard fluid_dynamics. Under this approach, we were_able_to find semi-analytical_solutions_which_show in_a clear way_the_physical effects_acting_in the growing or suppression of_linear_perturbations. The amended version of CLASS will_be used in future_studies_to determine the direct_effects of scalar fields in_different observables, specially on those that_make the_CMB spectrum_distinguishable from that of $\Lambda$CDM (see Fig. \[fig:3\]), for both types of_dark matter and dark energy fields._Another topic of interest_is the_setting_up of initial_conditions_of $N$-body_and wave-like simulations, so that one can_incorporate constraints_for scalar field models from linear_and non-linear perturbations.
Our results_indicate_that the growing of perturbations in_SFDM models for a given wave_number $k$ requires of two_conditions:_a)_an early onset of field_oscillations, preferably much earlier than the_time of radiation-matter_equality, and b) a negligible ratio $k/k_J$._Not_surprisingly, both conditions depend upon the_only_free physical parameter which is the_mass_of_the SFDM particle, and actually_the two conditions are simultaneously enhanced_(suppressed) in the case of large (small) values of_$m$ just because_both angular variables $\theta$ and_$\vartheta$_evolve_similarly throughout.
As for condition b), we must recall first that_the scales_that have been_affected at some degree by the Jeans wavenumber $k_J$, throughout_the evolution of the Universe, are those_for which initially $k/k_{J,eq} > 1$ with $k_{J,eq} = \Omega^{1/4}_{r0} k_{J0}$,_see |
+ {2 \over 3} u} \right). \end{aligned}$$ Choosing $u = t n_{k} \alpha_{k}(x)$ for $t \in (0, C]$, we have $$\label{Eqn:degree_con}
{\mathds{P}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Big| {\geqslant}t n_{k} \alpha_{k}(x) \right) {\leqslant}2 \exp \left( -{t^{2} n_{k} \alpha_{k}(x) \over 2 C + {2 \over 3} t} \right) \le 2 \exp \left( -C n_{k} \alpha_{k}(x) t^{2} \right).$$ Now choosing $t = c_{1} \sqrt{\log(n_{k})/(\alpha_{k}(x) n_{k})}$ for some large enough constant $c_{1} > 0$, we get $${\mathds{P}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Big| {\geqslant}c_{1} \sqrt{\alpha_{k}(x) n_{k} \log{n_{k}}} \right) {\leqslant}2 n_{k}^{-C c_{1}^{2}},$$ provided that $\log(n_{k}) {\leqslant}C \alpha_{k}(x) n_{k} {\leqslant}C n_{k} h^{q_{k}}$ in view of the uniform bounds and . But this is ensured by our bandwidth assumption . Thus for any fixed $x \in {\mathcal{D}}_{k}$, we have with probability at least $1 - c_{2} n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) \over n_{k}} - \alpha_{k}(x) \right| {\leqslant}c_{1} \sqrt{\alpha_{k}(x) \log{n_{k}} \over n_{k}}.$$ Then it follows that with probability at least $1 - c_{2} n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) \over n_{k} (\sqrt{2\pi} h)^{q_{k}}} - p_{k | + { 2 \over 3 } u } \right). \end{aligned}$$ Choosing $ u = t n_{k } \alpha_{k}(x)$ for $ t \in (0, C]$, we have $ $ \label{Eqn: degree_con }
{ \mathds{P}}\left (\Big|d_{n, k}(x) - n_k \alpha_{k}(x) \Big| { \geqslant}t n_{k } \alpha_{k}(x) \right) { \leqslant}2 \exp \left (-{t^{2 } n_{k } \alpha_{k}(x) \over 2 vitamin c + { 2 \over 3 } metric ton } \right) \le 2 \exp \left (-C n_{k } \alpha_{k}(x) t^{2 } \right).$$ Now choosing $ t = c_{1 } \sqrt{\log(n_{k})/(\alpha_{k}(x) n_{k})}$ for some big enough constant $ c_{1 } > 0 $, we get down $ $ { \mathds{P}}\left (\Big|d_{n, k}(x) - n_k \alpha_{k}(x) \Big| { \geqslant}c_{1 } \sqrt{\alpha_{k}(x) n_{k } \log{n_{k } } } \right) { \leqslant}2 n_{k}^{-C c_{1}^{2}},$$ provided that $ \log(n_{k }) { \leqslant}C \alpha_{k}(x) n_{k } { \leqslant}C n_{k } h^{q_{k}}$ in view of the consistent bounds and . But this is ensured by our bandwidth presumption . Thus for any fixed $ x \in { \mathcal{D}}_{k}$, we get with probability at least $ 1 - c_{2 } n_{k}^{-c_{3}}$, $ $ \left| { d_{n, k}(x) \over n_{k } } - \alpha_{k}(x) \right| { \leqslant}c_{1 } \sqrt{\alpha_{k}(x) \log{n_{k } } \over n_{k}}.$$ Then it follows that with probability at least $ 1 - c_{2 } n_{k}^{-c_{3}}$, $ $ \left| { d_{n, k}(x) \over n_{k } (\sqrt{2\pi } h)^{q_{k } } } - p_{k | + {2 \lver 3} u} \right). \end{aligned}$$ Choosing $u = t u_{j} \alphe_{k}(x)$ for $t \in (0, C]$, we have $$\label{Eqn:degree_con}
{\mavhds{P}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Bkg| {\geqslajt}t n_{k} \aopha_{j}(z) \right) {\leqslant}2 \ew' \leff( -{t^{2} n_{n} \alpha_{k}(x) \over 2 C + {2 \over 3} t} \right) \le 2 \efp \lzft( -C n_{k} \alpha_{k}(x) t^{2} \right).$$ Now choosind $t = c_{1} \xqgt{\log(n_{k})/(\alpha_{k}(x) n_{k})}$ gjr skme large enough constant $c_{1} > 0$, we gst $${\mathvs{P}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Big| {\geqslant}c_{1} \sqrh{\alpha_{k}(x) n_{k} \log{n_{k}}} \gight) {\leqslqnt}2 g_{j}^{-C c_{1}^{2}},$$ providea that $\log(n_{k}) {\leqslant}C \alpha_{k}(x) n_{k} {\leqslant}C n_{k} h^{q_{k}}$ in view of the unufirm tounds and . Uut thps is ensured by our tandwidyh assumption . Bhus hor qny fixed $x \in {\mathcan{D}}_{k}$, we have with ptobability ac least $1 - c_{2} n_{k}^{-c_{3}}$, $$\left| {e_{n,j}(x) \oeer t_{k}} - \xophx_{k}(x) \rmghf| {\leqspanv}c_{1} \sqrt{\alphz_{k}(x) \log{n_{k}} \ocer n_{k}}.$$ Then it follpwf that with prkbabilytr at least $1 - c_{2} n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) \over n_{k} (\sxrt{2\li} h)^{q_{k}}} - p_{k | + {2 \over 3} u} \right). \end{aligned}$$ = n_{k} \alpha_{k}(x)$ $t \in (0, \Big|d_{n,k}(x) n_k \alpha_{k}(x) \Big| n_{k} \alpha_{k}(x) \right) \exp \left( -{t^{2} n_{k} \alpha_{k}(x) \over C + {2 \over 3} t} \right) \le 2 \exp \left( -C n_{k} t^{2} \right).$$ Now choosing $t = c_{1} \sqrt{\log(n_{k})/(\alpha_{k}(x) n_{k})}$ for some large enough $c_{1} 0$, get \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Big| {\geqslant}c_{1} \sqrt{\alpha_{k}(x) n_{k} \log{n_{k}}} \right) {\leqslant}2 n_{k}^{-C c_{1}^{2}},$$ provided that $\log(n_{k}) \alpha_{k}(x) n_{k} {\leqslant}C n_{k} h^{q_{k}}$ in view of uniform bounds and . this is ensured by our assumption Thus for fixed \in we have with at least $1 - c_{2} n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) \over n_{k}} - \alpha_{k}(x) \right| {\leqslant}c_{1} \sqrt{\alpha_{k}(x) \log{n_{k}} \over Then it with probability least - n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) (\sqrt{2\pi} h)^{q_{k}}} - p_{k | + {2 \over 3} u} \right). \end{aligned}$$ ChoosIng $u = t n_{k} \alpHa_{k}(x)$ fOr $t \In (0, C]$, We Have $$\LabeL{Eqn:degree_con}
{\mAThds{p}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \BIg| {\geqSlANt}t n_{K} \AlPha_{k}(x) \Right) {\leQSlANT}2 \exP \lEfT( -{t^{2} n_{K} \aLPhA_{k}(x) \ovEr 2 C + {2 \Over 3} t} \riGht) \le 2 \exp \leFt( -C N_{k} \Alpha_{k}(x) t^{2} \righT).$$ noW choosing $t = C_{1} \sqRt{\log(n_{k})/(\alpha_{K}(x) n_{K})}$ for soMe LarGE enouGh cOnstaNt $c_{1} > 0$, we gET $${\mathdS{P}}\left( \Big|D_{n,K}(X) - n_k \alpHA_{k}(x) \Big| {\gEQSlAnt}c_{1} \Sqrt{\alpha_{k}(x) n_{k} \log{n_{K}}} \RiGHt) {\leqslant}2 n_{k}^{-C c_{1}^{2}},$$ ProvidEd THaT $\LOg(n_{K}) {\leQslant}C \alpHa_{K}(x) n_{k} {\lEQslant}C N_{K} h^{Q_{K}}$ IN viEW of the uniform Bounds and . BuT ThiS is ensUrEd bY Our banDwidtH aSSumPtion . Thus foR any Fixed $x \in {\mAthcal{d}}_{K}$, we have WIth probAbilitY at LeaSt $1 - c_{2} n_{K}^{-C_{3}}$, $$\lEfT| {d_{n,K}(x) \OVer N_{K}} - \aLphA_{K}(x) \rIght| {\leqsLaNt}C_{1} \sqrt{\AlphA_{K}(X) \LOg{n_{k}} \OveR n_{k}}.$$ THen it Follows that wiTh pRobaBIliTy at lEast $1 - c_{2} N_{k}^{-c_{3}}$, $$\lEfT| {d_{n,k}(x) \Over n_{k} (\Sqrt{2\pI} h)^{Q_{k}}} - p_{k | + {2 \over 3} u} \right). \end{alig ned}$ $ C hoo si ng $ u =t n_{k} \alpha _ {k}( x)$ for $t \in (0, C]$ , weha v e $$ \ la bel{E qn:degr e e_ c o n}{\ ma thd s{ P }} \left ( \ Big|d_{ n,k}(x) -n_k \ alpha_{k}(x) \B ig| {\geqs lan t}t n_{k} \a lph a_{k}( x) \r i ght){\l eqsla nt}2 \ e xp \le ft( -{t^{ 2} n_{k}\ alpha_{ k } (x ) \o ver 2 C + {2 \ove r 3 } t} \right) \l e 2 \e xp \l e f t(-Cn_{k} \alp ha _{k}( x ) t^{2} \r i g h t). $ $ Now choosin g $t = c_{1 } \s qrt{\l og (n_ { k})/(\ alpha _{ k }(x ) n_{k})}$forsome larg e enou g h const a nt $c_{ 1} > 0 $,weget$ ${ \m ath ds { P}} \ le ft( \Bi g|d_{n,k }( x) - n_ k \a l p h a _{k} (x) \Bi g| {\ geqslant}c_{1 } \ sqrt { \al pha_{ k}(x) n_{ k} \log {n_{k} }} \r ig ht) {\leqslant} 2 n_ {k}^{-C c _{1 }^ {2} }, $$ pr o videdtha t $ \log(n_ {k}) {\ l eqs la n t } C\alpha_{k}(x) n_{k }{ \ le qslant}C n_{k} h^ {q _ {k}}$ in v iew oft h e uni form bo unds an d . Bu t t hi s is en su red by o urban dwidt h ass umptio n . Thus fora ny fixed $x \i n {\mathcal{D} } _{ k } $, we h ave with proba bili t y at lea s t$1- c_{2 } n_{ k} ^ {- c _{3}}$, $$\left| {d _{ n,k}(x ) \o ver n_{k}} -\alpha_{k} ( x ) \right| {\l e qs l ant}c_{1} \sqr t{\al pha_{k}(x) \log{n_{ k}} \ over n_{ k}}.$$ Th e n it foll ows th atwit h pr obability atl e ast$1 - c_{2 } n _{k}^{- c_{ 3}} $,$$\ le ft| {d_{n ,k}(x) \o ve rn_ {k} (\sq r t{2\pi}h) ^{q _{ k}} } - p _ {k | +_{2 \over_3} u} \right). \end{aligned}$$_Choosing $u_=_t n_{k}_\alpha_{k}(x)$_for $t \in_(0, C]$, we_have $$\label{Eqn:degree_con}
{\mathds{P}}\left( \Big|d_{n,k}(x) -_n_k \alpha_{k}(x) \Big|_{\geqslant}t_n_{k} \alpha_{k}(x) \right) {\leqslant}2 \exp \left( -{t^{2} n_{k} \alpha_{k}(x) \over 2 C + {2_\over_3} t}_\right)_\le_2 \exp \left( -C n_{k}_\alpha_{k}(x) t^{2} \right).$$ Now choosing_$t =_c_{1} \sqrt{\log(n_{k})/(\alpha_{k}(x) n_{k})}$ for some large enough constant_$c_{1}_> 0$, we_get $${\mathds{P}}\left( \Big|d_{n,k}(x) - n_k \alpha_{k}(x) \Big| {\geqslant}c_{1} \sqrt{\alpha_{k}(x)_n_{k} \log{n_{k}}} \right) {\leqslant}2 n_{k}^{-C c_{1}^{2}},$$_provided that $\log(n_{k})_{\leqslant}C_\alpha_{k}(x)_n_{k} {\leqslant}C n_{k} h^{q_{k}}$_in view of the uniform bounds _and . But this is ensured by_our bandwidth assumption . Thus for any fixed_$x \in {\mathcal{D}}_{k}$, we have with_probability at least $1 -_c_{2} n_{k}^{-c_{3}}$,_$$\left| {d_{n,k}(x) \over n_{k}}_- \alpha_{k}(x) \right|_{\leqslant}c_{1} \sqrt{\alpha_{k}(x)_\log{n_{k}} \over n_{k}}.$$_Then it follows that with probability_at least $1_- c_{2} n_{k}^{-c_{3}}$, $$\left| {d_{n,k}(x) _\over_n_{k} (\sqrt{2\pi} h)^{q_{k}}}_-_p_{k |
ldots\}$, we use and to obtain $$\label{eq.d2tilsUsasy}
\tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s)) \sim
\frac{1}{\mathcal{W}(a,b,0)}b |a|^{2b/a-2}s^{2b/a}
\text{ as $s\to\infty$}.$$
We now prove part (a). If $b<0$ and $b/a\not\in\{1,2,\ldots\}$, we have that $2b/a>0$, so using $$\int_0^t d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim \Gamma(-\ba){\mathrm{e}}^{a} b
|a|^{2b/a-2}\, t^{2b/a+1}\frac{1}{2b/a+1},$$ as $t\to\infty$. Therefore by, as $t\to\infty$, we have that $$\label{eq.int2ctnondeg}
r_2(t)\int_0^t d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds
\sim b |a|^{b/a-3}\, \frac{1}{2b/a+1} t^{b/a}. $$ In the case that $b<0$ and $b/a\in\{1,2,\ldots\}$ using gives $$\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim
\frac{1}{\mathcal{W}(a,b,0)}b |a|^{2b/a-2}t^{2b/a+1}\frac{1}{2b/a+1}, $$ as $t\to\infty$. Therefore by we have that $$\label{eq.int2ctdeg}
\tilde{r}_2(t)\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds
\ | ldots\}$, we use and to obtain $ $ \label{eq.d2tilsUsasy }
\tilde{d}_2(s)(1+s)U(1 - b / a,2,-a(1+s) ) \sim
\frac{1}{\mathcal{W}(a, b,0)}b |a|^{2b / a-2}s^{2b / a }
\text { as $ s\to\infty$}.$$
We now rise character (a). If $ b<0 $ and $ b / a\not\in\{1,2,\ldots\}$, we have that $ 2b / a>0 $, so use $ $ \int_0^t d_2(s)(1+s)U(1 - b / a,2,-a(1+s))\,ds \sim \Gamma(-\ba){\mathrm{e}}^{a } b
|a|^{2b / a-2}\, t^{2b / a+1}\frac{1}{2b / a+1},$$ as $ t\to\infty$. Therefore aside, as $ t\to\infty$, we have that $ $ \label{eq.int2ctnondeg }
r_2(t)\int_0^t d_2(s)(1+s)U(1 - b / a,2,-a(1+s))\,ds
\sim b |a|^{b / a-3}\, \frac{1}{2b / a+1 } t^{b / a }. $ $ In the event that $ b<0 $ and $ b / a\in\{1,2,\ldots\}$ using give $ $ \int_0^t \tilde{d}_2(s)(1+s)U(1 - b / a,2,-a(1+s))\,ds \sim
\frac{1}{\mathcal{W}(a, b,0)}b |a|^{2b / a-2}t^{2b / a+1}\frac{1}{2b / a+1 }, $ $ as $ t\to\infty$. Therefore by we get that $ $ \label{eq.int2ctdeg }
\tilde{r}_2(t)\int_0^t \tilde{d}_2(s)(1+s)U(1 - b / a,2,-a(1+s))\,ds
\ | ldohs\}$, we use and to obtain $$\uabel{eq.d2tilsUsasy}
\tilde{v}_2(s)(1+s)U(1-b/a,2,-a(1+a)) \sim
\fraz{1}{\mathcal{W}(a,b,0)}b |a|^{2b/a-2}s^{2b/a}
\text{ as $s\ro\infuj$}.$$
We now prove part (a). Kf $b<0$ and $h/a\not\in\{1,2,\leots\}$, qe have thef $2b/a>0$, so using $$\lnt_0^t b_2(s)(1+w)U(1-b/a,2,-a(1+s))\,ds \sim \Gsmma(-\ba){\mathsm{e}}^{a} b
|a|^{2b/a-2}\, t^{2b/a+1}\fsaz{1}{2b/c+1},$$ as $t\to\infty$. Therefore by, as $t\to\infey$, we hsvf that $$\label{eq.ynt2cunogdeg}
d_2(n)\ikt_0^t d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds
\sim b |a|^{b/a-3}\, \frac{1}{2b/a+1} t^{b/a}. $$ Ii the case that $b<0$ and $b/a\in\{1,2,\ldots\}$ using givfs $$\ijt_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,dd \sim
\frac{1}{\majgcaj{Q}(a,b,0)}b |a|^{2b/a-2}t^{2b/a+1}\wrac{1}{2b/a+1}, $$ as $t\to\infty$. Thgrefore by we have that $$\label{eq.ivt2ctdzg}
\tilde{r}_2(t)\inj_0^c \tipge{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,dw
\ | ldots\}$, we use and to obtain $$\label{eq.d2tilsUsasy} \frac{1}{\mathcal{W}(a,b,0)}b \text{ as We now prove $b/a\not\in\{1,2,\ldots\}$, have that $2b/a>0$, using $$\int_0^t d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \Gamma(-\ba){\mathrm{e}}^{a} b |a|^{2b/a-2}\, t^{2b/a+1}\frac{1}{2b/a+1},$$ as $t\to\infty$. by, as $t\to\infty$, we have that $$\label{eq.int2ctnondeg} r_2(t)\int_0^t d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim b |a|^{b/a-3}\, \frac{1}{2b/a+1} $$ In the case that $b<0$ and $b/a\in\{1,2,\ldots\}$ using gives $$\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim |a|^{2b/a-2}t^{2b/a+1}\frac{1}{2b/a+1}, as Therefore we have that $$\label{eq.int2ctdeg} \tilde{r}_2(t)\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \ | ldots\}$, we use and to obtain $$\labeL{eq.d2tilsUsAsy}
\tiLde{D}_2(s)(1+s)u(1-b/A,2,-a(1+s)) \sIm
\frAc{1}{\mathcal{W}(a,b,0)}b |a|^{2B/A-2}s^{2b/a}
\Text{ as $s\to\infty$}.$$
We now proVe parT (a). iF $b<0$ anD $B/a\Not\in\{1,2,\Ldots\}$, we HAvE THat $2B/a>0$, So UsiNg $$\INt_0^T d_2(s)(1+s)U(1-B/a,2,-a(1+S))\,ds \sim \GAmma(-\ba){\mathRm{e}}^{A} b
|A|^{2b/a-2}\, t^{2b/a+1}\frac{1}{2b/a+1},$$ AS $t\To\infty$. TheRefOre by, as $t\to\inFty$, We have ThAt $$\lABel{eq.Int2CtnonDeg}
r_2(t)\iNT_0^t d_2(s)(1+s)U(1-B/a,2,-a(1+s))\,ds
\sim B |a|^{B/A-3}\, \frac{1}{2b/A+1} T^{b/a}. $$ In thE CAsE thaT $b<0$ and $b/a\in\{1,2,\ldots\}$ usiNG gIVes $$\int_0^t \tilde{d}_2(s)(1+S)U(1-b/a,2,-a(1+s))\,Ds \SIm
\FRAc{1}{\mAthCal{W}(a,b,0)}b |a|^{2b/a-2}T^{2b/A+1}\frac{1}{2B/A+1}, $$ as $t\to\iNFtY$. tHEreFOre by we have thAt $$\label{eq.inT2CtdEg}
\tildE{r}_2(T)\inT_0^T \tilde{D}_2(s)(1+s)U(1-b/A,2,-a(1+S))\,Ds
\ | ldots\}$, we use and to ob tain $$\la bel{e q.d 2ti ls Usas y}
\ tilde{d}_2(s)( 1 +s)U (1-b/a,2,-a(1+s)) \sim
\fra c{ 1 }{\m a th cal{W }(a,b,0 ) }b |a| ^{ 2b /a- 2} s ^{ 2b/a}
\ text{ a s $s\to\in fty $} .$$
We nowp ro ve part (a ).If $b<0$ and $b /a\not \i n\{ 1 ,2,\l dot s\}$, we ha v e that $2b/a>0$ ,s o usin g $$\int _ 0 ^t d_2 (s)(1+s)U(1-b/a,2 , -a ( 1+s))\,ds \sim \Gamm a( - \b a ) {\m ath rm{e}}^{a} b
|a|^ { 2b/a-2} \ ,t ^ { 2b/ a +1}\frac{1}{2 b/a+1},$$ a s $t \to\in ft y$. Theref ore b y, as$t\to\infty $, w e have th at $$\ l abel{eq . int2ctn ondeg}
r_ 2(t )\in t _0 ^t d_ 2( s )(1 + s) U(1 - b/a ,2,-a(1+ s) )\ ,ds
\sim b | a|^{ b/a -3}\ , \fr ac{1}{2b/a+1} t ^{b/ a }.$$ In thecase t hat $ b<0$ a nd $b /a \in\{1,2,\ldots \}$using giv es$$ \in t_ 0^t \ t ilde{d }_2 (s) (1+s)U( 1-b/a,2 , -a( 1+ s ) ) \, ds \sim
\frac{1}{\ ma t h ca l{W}(a,b ,0)}b|a |^ { 2b/a-2}t ^{ 2b/ a+1} \ f rac{1 }{2b / a+ 1}, $$ a s $t\t o \i nf ty$. Th er eforeby we ha ve th a t $$ \label {eq.int2 ctdeg }
\tilde{r}_2(t ) \int_0^t \til d e{ d } _2 ( s)(1 +s) U(1-b/a,2,- a(1+ s ))\, ds
\ | ldots\}$, we_use and_to obtain $$\label{eq.d2tilsUsasy}
\tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s)) \sim
\frac{1}{\mathcal{W}(a,b,0)}b_ |a|^{2b/a-2}s^{2b/a}
_\text{_as $s\to\infty$}.$$
We_now_prove part (a)._If $b<0$ and_$b/a\not\in\{1,2,\ldots\}$, we have that_$2b/a>0$, so using_$$\int_0^t_d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim \Gamma(-\ba){\mathrm{e}}^{a} b
|a|^{2b/a-2}\, t^{2b/a+1}\frac{1}{2b/a+1},$$ as $t\to\infty$. Therefore by, as $t\to\infty$, we have that_$$\label{eq.int2ctnondeg}
r_2(t)\int_0^t_d_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds
\sim_b_|a|^{b/a-3}\,_\frac{1}{2b/a+1} t^{b/a}. $$ In_the case that $b<0$ and_$b/a\in\{1,2,\ldots\}$ using_gives $$\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds \sim
\frac{1}{\mathcal{W}(a,b,0)}b |a|^{2b/a-2}t^{2b/a+1}\frac{1}{2b/a+1}, $$ as_$t\to\infty$._Therefore by we_have that $$\label{eq.int2ctdeg}
\tilde{r}_2(t)\int_0^t \tilde{d}_2(s)(1+s)U(1-b/a,2,-a(1+s))\,ds
\ |
The baseline R2D2 maintains a Q-function $Q_\theta(x,a)$ parameterized by a neural network $\theta$. The central learner maintains an updated parameter $\theta$ and distributed actors maintain slightly delayed copies $\theta_{\text{old}}$. Distributed actors collect data using behavior policy $\mu$, defined as ${\varepsilon}$-greedy with respect to $Q_{\theta_{\text{old}}}(x,a)$. The target policy $\pi$ is defined as greedy with respect to $Q_\theta(x,a)$. Actors send data to a replay buffer, and the learner samples partial trajectories $(x_t,a_t,r_t)_{t=1}^T$ from the buffer and computes updates to the parameter $\theta$. In particular, the learner calculates regression targets $Q_{\text{target}}(x_t,a_t)$ and the Q-function is updated via $\theta\leftarrow \theta - \alpha \cdot \nabla_\theta (Q_\theta(x_t,a_t)-Q_{\text{target}}(x_t,a_t))^2$ with learning rate $\alpha > 0$.
#### Algorithmic variants.
Algorithmic variants of R2D2 differ in how they compute the targets $Q_{\text{target}}(x,a)$. A useful unified view provided by @rowland2019adaptive is that $Q_{\text{target}}(x,a)$ aims to approximate $Q^\pi(x,a)$ such that $Q_\theta(x,a) \rightarrow Q^\pi(x,a)$ during the update.
Along sampled trajectories, we recursively calculate the targets $Q_{\text{target}}(x_t,a_t),1\leq t\leq T$ based on recipes of different variants. Below are a few alternatives we evaluated in our experimenst, where we e.g., use $\hat{Q}_{\text{zero}}(x_t,a_t)$ to represent $Q_{\text{target}}(x_t,a_t)$ for the zero-order baseline.
- **Zero-order:** $\hat{Q}_{\text{zero}}(x_t,a_t) \triangleq r_t + \gamma \hat{Q}_{\text{zero}}(x_{t+1},a_{t+1})$
- **First-order:** $\hat{ | The baseline R2D2 maintains a Q - function $ Q_\theta(x, a)$ parameterized by a neural net $ \theta$. The cardinal apprentice maintains an updated parameter $ \theta$ and spread actors maintain slenderly delayed copies $ \theta_{\text{old}}$. Distributed actor collect data use behavior policy $ \mu$, defined as $ { \varepsilon}$-greedy with obedience to $ Q_{\theta_{\text{old}}}(x, a)$. The target policy $ \pi$ is defined as greedy with respect to $ Q_\theta(x, a)$. Actors send datum to a replay buffer, and the learner samples fond trajectories $ (x_t, a_t, r_t)_{t=1}^T$ from the buffer and computes updates to the parameter $ \theta$. In especial, the learner calculates regression targets $ Q_{\text{target}}(x_t, a_t)$ and the Q - function is update via $ \theta\leftarrow \theta - \alpha \cdot \nabla_\theta (Q_\theta(x_t, a_t)-Q_{\text{target}}(x_t, a_t))^2 $ with learning rate $ \alpha > 0$.
# # # # Algorithmic variants.
Algorithmic variants of R2D2 differ in how they compute the targets $ Q_{\text{target}}(x, a)$. A useful unified view provided by @rowland2019adaptive is that $ Q_{\text{target}}(x, a)$ aims to approximate $ Q^\pi(x, a)$ such that $ Q_\theta(x, a) \rightarrow Q^\pi(x, a)$ during the update.
Along sampled trajectories, we recursively calculate the targets $ Q_{\text{target}}(x_t, a_t),1\leq t\leq T$ based on recipe of unlike variants. Below are a few alternatives we evaluated in our experimenst, where we for example, practice $ \hat{Q}_{\text{zero}}(x_t, a_t)$ to represent $ Q_{\text{target}}(x_t, a_t)$ for the zero - order baseline.
- * * Zero - order :* * $ \hat{Q}_{\text{zero}}(x_t, a_t) \triangleq r_t + \gamma \hat{Q}_{\text{zero}}(x_{t+1},a_{t+1})$
- * * First - order :* * $ \hat { |
The baseline R2D2 maintains a Q-function $Q_\theja(z,a)$ paremeteriaed by a neural network $\theta$. The ceitrao leaener maintains an updaged paramvter $\thetq$ anv distributed acvkrs maikcain apighclb delayed copiex $\theta_{\texd{old}}$. Distributad aetors collect data using behavior pojicy $\mu$, dffined as ${\vareksilom}$-dreesj cith respect to $Q_{\theta_{\text{old}}}(x,a)$. The tagget policy $\pi$ is defined as greedy with redpech to $Q_\theta(x,a)$. Actogs send datq to q replay bufwer, and tht nearner sajples partial trajectories $(x_t,a_t,f_t)_{t=1}^T$ yrom the budfwr wtd computes updanes to the pavsmeter $\theta$. On particular, bhe lxarnwr calculates regressmon targets $Q_{\text{tardet}}(x_t,a_t)$ atd the Q-function is updateg vid $\thdra\ldftzrcow \theta - \ampha \cdot \habla_\theta (W_\theta(x_t,a_t)-Q_{\text{targtt}}(x_e,q_t))^2$ with learnjng raee $\alpha > 0$.
#### Algorithmic variants.
Algorithmib vadiants of R2D2 differ in yow they compute the jargets $Q_{\tqxt{target}}(x,a)$. A useful unified view provided by @rownand2019avaotirc kw hhat $Q_{\text{target}}(x,a)$ aims to approximate $Q^\pi(x,a)$ fhcn nhat $Q_\theta(x,a) \rigmtarrow Q^\pi(x,a)$ durimg tng update.
Along rampleb tdajectories, we reckrsiveli calcylate the tarbets $Q_{\text{target}}(x_t,a_t),1\leq t\lew T$ based on eecipes of differeut variants. Yelow sre a few alternatives we evcluates in our exoerimenst, dhere we e.g., use $\fat{A}_{\taxt{zero}}(x_u,x_t)$ to represent $Q_{\eext{targev}}(x_t,a_t)$ for thd zeto-order baseline.
- **Zero-order:** $\hat{Q}_{\text{zero}}(d_t,a_t) \jriangneq r_t + \galma \hat{Q}_{\text{zero}}(x_{t+1},a_{t+1})$
- **First-ordxc:** $\hat{ | The baseline R2D2 maintains a Q-function $Q_\theta(x,a)$ a network $\theta$. central learner maintains distributed maintain slightly delayed $\theta_{\text{old}}$. Distributed actors data using behavior policy $\mu$, defined ${\varepsilon}$-greedy with respect to $Q_{\theta_{\text{old}}}(x,a)$. The target policy $\pi$ is defined as greedy respect to $Q_\theta(x,a)$. Actors send data to a replay buffer, and the learner partial $(x_t,a_t,r_t)_{t=1}^T$ the and computes updates to the parameter $\theta$. In particular, the learner calculates regression targets $Q_{\text{target}}(x_t,a_t)$ and Q-function is updated via $\theta\leftarrow \theta - \alpha \nabla_\theta (Q_\theta(x_t,a_t)-Q_{\text{target}}(x_t,a_t))^2$ with learning $\alpha > 0$. #### Algorithmic Algorithmic of R2D2 in they the targets $Q_{\text{target}}(x,a)$. useful unified view provided by @rowland2019adaptive is that $Q_{\text{target}}(x,a)$ aims to approximate $Q^\pi(x,a)$ such that $Q_\theta(x,a) \rightarrow during the sampled trajectories, recursively the $Q_{\text{target}}(x_t,a_t),1\leq t\leq T$ recipes of different variants. Below are we evaluated in our experimenst, where we e.g., $\hat{Q}_{\text{zero}}(x_t,a_t)$ to $Q_{\text{target}}(x_t,a_t)$ for the zero-order baseline. - $\hat{Q}_{\text{zero}}(x_t,a_t) \triangleq r_t + \gamma \hat{Q}_{\text{zero}}(x_{t+1},a_{t+1})$ - **First-order:** |
The baseline R2D2 maintains a Q-fUnction $Q_\thEta(x,a)$ ParAmeTeRizeD by a Neural network $\tHEta$. THe central learner maintaIns an UpDAted PArAmeteR $\theta$ aND dISTriBuTeD acToRS mAintaIn sLightly Delayed copIes $\ThEta_{\text{old}}$. DiSTrIbuted actoRs cOllect data usIng BehaviOr PolICy $\mu$, dEfiNed as ${\VarepsILon}$-greEdy with reSpECt to $Q_{\tHEta_{\text{OLD}}}(x,A)$. The Target policy $\pi$ is dEFiNEd as greedy with RespecT tO $q_\tHETa(x,A)$. AcTors send daTa To a rePLay buffER, aND THe lEArner samples pArtial trajeCTorIes $(x_t,a_T,r_T)_{t=1}^T$ FRom the BuffeR aND coMputes updatEs to The parameTer $\theTA$. In partICular, thE learnEr cAlcUlatES rEgResSiON taRGeTs $Q_{\TExt{Target}}(x_t,A_t)$ AnD the Q-FuncTION Is upDatEd viA $\thetA\leftarrow \theTa - \aLpha \CDot \Nabla_\Theta (q_\theTa(X_t,a_t)-Q_{\Text{taRget}}(x_T,a_T))^2$ with learning raTe $\alPha > 0$.
#### AlgoriThmIc VarIaNts.
AlGOrithmIc vAriAnts of R2d2 differ IN hoW tHEY CoMpute the targets $Q_{\teXt{TARgEt}}(x,a)$. A useFul uniFIeD vIEw providEd By @rOwlaND2019AdaptIve iS ThAt $Q_{\text{tArget}}(x,A)$ AiMs To approXiMate $Q^\pI(x,A)$ suCh tHat $Q_\tHEta(x,A) \rightArrow Q^\pi(X,a)$ durINg the update.
AloNG sampled trajeCToRIEs, WE recUrsIvely calculAte tHE tarGets $q_{\TeXt{tARget}}(x_T,a_t),1\leQ t\LEq t$ Based on recipes of difFeRent vaRiantS. Below are a few AlternativES WE evaluatEd in OUr EXperimenst, wherE we e.g., Use $\hat{Q}_{\texT{Zero}}(x_t,a_t)$ To repResent $Q_{\tExt{target}}(X_T,A_t)$ for the ZerO-orDer BasELInE.
- **Zero-order:** $\hat{q}_{\TExt{zErO}}(x_t,a_t) \trIanGleq r_t + \gAmmA \haT{Q}_{\tExt{ZeRo}}(x_{t+1},a_{t+1})$
- **FirSt-order:** $\hAt{ |
The baseline R2D2 mainta ins a Q-fu nctio n $ Q_\ th eta( x,a) $ parameterize d bya neural network $\the ta$.Th e cen t ra l lea rner ma i nt a i nsan u pda te d p arame ter $\thet a$ and dis tri bu ted actors m a in tain sligh tly delayed cop ies $\the ta _{\ t ext{o ld} }$. D istrib u ted ac tors coll ec t datau sing be h a vi or p olicy $\mu$, defi n ed as ${\varepsil on}$-g re e dy w ith re spect to $ Q_ {\the t a_{\tex t {o l d } }}( x ,a)$. The tar get policy$ \pi $ is d ef ine d as gr eedywi t h r espect to $ Q_\t heta(x,a) $. Act o rs send data to a rep lay bu ffer , a nd th el ear n er sa m ple s partia ltr aject orie s $ ( x_t, a_t ,r_t )_{t= 1}^T$ from th e b uffe r an d com putes upd at es to the p arame te r $\theta$. Inpart icular, t hele arn er calc u latesreg res sion ta rgets $ Q _{\ te x t { ta rget}}(x_t,a_t)$ a nd t he Q-funct ion is up da t ed via $ \t het a\le f t arrow \th e ta - \alph a \cdo t \ na bla_\th et a (Q_\ th eta (x_ t,a_t ) -Q_{ \text{ target}} (x_t, a _t))^2$ with l e arning rate $ \ al p h a> 0$.
# ### Algorit hmic vari ants .
Alg o rithm ic va ri a nt s of R2D2 differ inho w they comp ute the targe ts $Q_{\te x t { target}} (x,a ) $. A useful unifi ed vi ew provide d by @row land2 019adapt ive is th a t $Q_{\te xt{ tar get }}( x , a) $ aims to app r o xima te $Q^\pi (x, a)$ suc h t hat $Q _\t he ta(x,a) \ rightarr ow Q ^\ pi (x, a)$ d u ring the u pda te .
Along sample d tra ject or ie s , w e recur s iv e l y ca lc ul atethe t arget s $Q _ {\t ext{tar get}}(x_t ,a_ t ),1\ le qt\leq T $ based on re ci pes of dif fe ren t vari a n ts. Belo w are a few alternative s we eva lua ted i n ou r experim ens t, whe rew e e.g. , use$\hat {Q }_{ \ t ext{z e r o} }(x _t ,a_t)$ tor e pre sent$Q _{\t ext{tar get}}(x_t,a_t)$ fo r th e zero-orderbas elin e .
- ** Z ero -o r der : * * $\hat{Q}_{\te xt{zero}}( x_ t ,a _t) \trian g leq r _t + \g amma \h at{Q} _ {\text{ zero}}(x_ {t+1},a_{ t+ 1})$ - **First-o rder:**$\hat{ |
The baseline_R2D2 maintains_a Q-function $Q_\theta(x,a)$ parameterized_by a_neural_network $\theta$._The_central learner maintains_an updated parameter_$\theta$ and distributed actors_maintain slightly delayed_copies_$\theta_{\text{old}}$. Distributed actors collect data using behavior policy $\mu$, defined as ${\varepsilon}$-greedy with respect_to_$Q_{\theta_{\text{old}}}(x,a)$. The_target_policy_$\pi$ is defined as greedy_with respect to $Q_\theta(x,a)$. Actors_send data_to a replay buffer, and the learner samples_partial_trajectories $(x_t,a_t,r_t)_{t=1}^T$ from_the buffer and computes updates to the parameter $\theta$._In particular, the learner calculates regression_targets $Q_{\text{target}}(x_t,a_t)$ and_the_Q-function_is updated via $\theta\leftarrow_\theta - \alpha \cdot \nabla_\theta (Q_\theta(x_t,a_t)-Q_{\text{target}}(x_t,a_t))^2$_with learning rate $\alpha > 0$.
####_Algorithmic variants.
Algorithmic variants of R2D2 differ in_how they compute the targets $Q_{\text{target}}(x,a)$._A useful unified view provided_by @rowland2019adaptive_is that $Q_{\text{target}}(x,a)$ aims to_approximate $Q^\pi(x,a)$ such_that $Q_\theta(x,a)_\rightarrow Q^\pi(x,a)$ during_the update.
Along sampled trajectories, we recursively_calculate the targets_$Q_{\text{target}}(x_t,a_t),1\leq t\leq T$ based on recipes_of_different variants. Below_are_a_few alternatives_we evaluated in_our_experimenst, where_we_e.g., use $\hat{Q}_{\text{zero}}(x_t,a_t)$ to represent $Q_{\text{target}}(x_t,a_t)$_for_the zero-order baseline.
- **Zero-order:** $\hat{Q}_{\text{zero}}(x_t,a_t)_\triangleq r_t + \gamma_\hat{Q}_{\text{zero}}(x_{t+1},a_{t+1})$
-_ **First-order:** $\hat{ |
![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density of the disk, and color points indicate cluster particles. Since these are only twelve colors but 97 clusters, we plot several clusters with the same color. The color points are as large as the cluster size to make the points visible. More than 80% of cluster-particles are still bound at 1 Gyr. \[fig:snap\]](snap0000new.eps "fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density of the disk, and color points indicate cluster particles. Since these are only twelve colors but 97 clusters, we plot several clusters with the same color. The color points are as large as the cluster size to make the points visible. More than 80% of cluster-particles are still bound at 1 Gyr. \[fig:snap\]](snap0052new.eps "fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density of the disk, and color points indicate cluster particles. Since these are only twelve colors but 97 clusters, we plot several clusters with the same color. The color points are as large as the cluster size to make the points visible. More than 80% of cluster-particles are still bound at 1 Gyr. \[fig:snap\]](snap0106new.eps "fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density of the disk, and color points indicate cluster particles. Since these are only twelve colors but 97 clusters, we plot several clusters with the same color. The color points are as large as the cluster size to make the points visible. More than 80% of cluster-particles are still bound at 1 Gyr. \[fig:snap\]](snap0212new.eps "fig:"){width="40mm"}
Results
=======
Radial migration of star clusters
---------------------------------
Figure \[fig:snap\] shows the time evolution of the star clusters and disk. Although each spiral arm looks like a single coherent structure in each snapshot, it is transient and recurrent. They are wound up due to the differential rotation of the galactic disk and break | ! [ Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale testify the column concentration of the disk, and color point argue cluster particles. Since these are entirely twelve color but 97 clusters, we diagram several clusters with the like color. The color points are as big as the cluster size to make the points visible. More than 80% of cluster - particles are however bound at 1 Gyr. \[fig: snap\]](snap0000new.eps " fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray plate shows the column density of the phonograph record, and color point indicate cluster particles. Since these are only twelve colors but 97 bunch, we plot several clusters with the same color. The color point are as large as the cluster size to make the points visible. More than 80% of cluster - particles are still bound at 1 Gyr. \[fig: snap\]](snap0052new.eps " fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. grey scale shows the column concentration of the phonograph record, and coloring material points indicate bunch particles. Since these are merely twelve colors but 97 clusters, we diagram several clusters with the same color. The color points are equally large as the cluster size to make the points visible. More than 80% of cluster - particles are still constipate at 1 Gyr. \[fig: snap\]](snap0106new.eps " fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density of the phonograph record, and color points indicate bunch particles. Since these are entirely twelve colors but 97 clusters, we plat several bunch with the same color. The color points are as bombastic as the cluster size to reach the points visible. More than 80% of bunch - particles are still bound at 1 Gyr. \[fig: snap\]](snap0212new.eps " fig:"){width="40 mm " }
resultant role
= = = = = = =
Radial migration of star bunch
---------------------------------
Figure \[fig: snap\ ] shows the time development of the star clusters and disk. Although each spiral weapon looks like a single coherent structure in each snapshot, it is ephemeral and recurrent. They are wound up due to the differential rotation of the galactic magnetic disk and break | ![Snaoshots of the initial cokdition and at 0.25, 0.5, and 1 Gbr. Gray scale sfows the column density of tie dusk, abd color points indicage clusteg particlws. Smnce these are oimy twelyz colkvs buc 97 clusters, we pkot severan clusters witv ghz same color. The color points are as large ss the cluster syze uo iake nht points visible. More than 80% of clhster-pagticles are still bound at 1 Gyr. \[fig:snap\]](snap0000jew.eos "fig:"){width="40mm"}![Snapsjots of the inieual conditiov and at 0.25, 0.5, and 1 Gyr. Grzy scale shows the column densigy of the disk, qne clnor points mndicane cluster pavnicles. Vince tnese are only bwelvx cooors but 97 clusters, we plot several clusjers with dhz same color. The coloe pointv ara as oaree zs tge cludtec size to mzke the poibts visible. More thsn 80% of cluster-padticlef wre still bound at 1 Gyr. \[fig:snap\]](snap0052new.eks "fif:"){width="40mm"}![Snapshots of thw initial condition ajd at 0.25, 0.5, agd 1 Gyr. Gray scale shows the column density of tha disi, anb colof plints indicate cluster particles. Since these wde okly twelve colorf but 97 clusyegs, re plot severxl clusteds with the same cllor. Thg coloe points wre ss large as the cluster sizw to make thv pounts visible. More chan 80% of clurter-kartickes are still bound at 1 Gyr. \[fig:snap\]](snao0106new.eps "fje:"){width="40mm"}![Snapshotr on tve initial condition and ae 0.25, 0.5, and 1 Jyr. Gxay scald shpws thq column dfnsitn of the disk, and cllor koints indicate fluster particles. Since these acx only twelve cmlogs but 97 cjustevs, we plot sevewal clusters wnth the fame zolor. The bolor poiits are as lwrge as the cnoster size to make thq pounts visibld. More than 80% of cluster-pcxticles aee still bound at 1 Gyt. \[rig:snap\]](snap0212new.e'r "fig:"){width="40mm"}
Resukts
=======
Fadyap kidsation of stdr cuusgrrs
---------------------------------
Fieure \[fig:snai\] sfows the time evolution mf tge star clusters amd disk. Aljhough eash spiral arm looks like a singpe coierent struvtute in each snapshot, it is transjent and gecmrrent. They awe wiund up due co the differential rotation of the galartic disk and break | ![Snapshots of the initial condition and at and Gyr. Gray shows the column color indicate cluster particles. these are only colors but 97 clusters, we plot clusters with the same color. The color points are as large as the size to make the points visible. More than 80% of cluster-particles are still at Gyr. "fig:"){width="40mm"}![Snapshots the initial condition and at 0.25, 0.5, and 1 Gyr. Gray scale shows the column density the disk, and color points indicate cluster particles. these are only twelve but 97 clusters, we plot clusters the same The points as large as cluster size to make the points visible. More than 80% of cluster-particles are still bound at 1 \[fig:snap\]](snap0052new.eps "fig:"){width="40mm"}![Snapshots initial condition at 0.5, 1 Gyr. Gray the column density of the disk, indicate cluster particles. Since these are only twelve but 97 we plot several clusters with the color. The color points are as large as cluster size to make the points visible. More than 80% of cluster-particles are still bound Gyr. \[fig:snap\]](snap0106new.eps "fig:"){width="40mm"}![Snapshots of initial condition and 0.25, and Gyr. scale shows column density of the disk, and color points indicate cluster particles. these are only twelve colors but 97 clusters, we plot with same color. The points are as large the size to make the More 80% still at Gyr. \[fig:snap\]](snap0212new.eps "fig:"){width="40mm"} Results Radial migration of star clusters Figure \[fig:snap\] shows the clusters and disk. Although each spiral arm looks a single coherent structure in each snapshot, is transient and recurrent. They are wound up due to the differential of the and break | ![Snapshots of the initial condItion and at 0.25, 0.5, And 1 GyR. GrAy sCaLe shOws tHe column densitY Of thE disk, and color points indIcate ClUSter PArTicleS. Since tHEsE ARe oNlY tWelVe COlOrs buT 97 clUsters, wE plot severAl cLuSters with the SAmE color. The cOloR points are as LarGe as thE cLusTEr sizE to Make tHe poinTS visibLe. More thaN 80% oF ClusteR-ParticlES ArE stiLl bound at 1 Gyr. \[fig:snAP\]](sNAp0000new.eps "fig:"){widTh="40mm"}![SnApSHoTS Of tHe iNitial condItIon anD At 0.25, 0.5, and 1 GyR. grAY SCalE Shows the columN density of tHE diSk, and cOlOr pOInts inDicatE cLUstEr particles. sincE these are Only twELve coloRS but 97 cluSters, wE plOt sEverAL cLuSteRs WIth THe SamE ColOr. The colOr PoInts aRe as LARGE as tHe cLustEr sizE to make the poiNts VisiBLe. MOre thAn 80% of cLustEr-PartiCles arE stilL bOund at 1 Gyr. \[fig:snaP\]](snaP0052new.eps "fiG:"){wiDtH="40mm"}![snApshoTS of the IniTiaL conditIon and aT 0.25, 0.5, And 1 gyR. gRAy Scale shows the columN dENSiTy of the dIsk, and COlOr POints indIcAte ClusTER partIcleS. siNce these Are onlY TwElVe colorS bUt 97 clusTeRs, wE plOt sevERal cLusterS with the Same cOLor. The color poiNTs are as large aS ThE CLuSTer sIze To make the poInts VIsibLe. MoRE tHan 80% OF clusTer-paRtIClES are still bound at 1 Gyr. \[FiG:snap\]](sNap0106neW.eps "fig:"){width="40mM"}![Snapshots OF THe initiaL conDItIOn and at 0.25, 0.5, and 1 Gyr. GRay scAle shows thE Column deNsity Of the disK, and color POInts indiCatE clUstEr pARTiCles. Since thesE ARe onLy Twelve cOloRs but 97 clUstErs, We pLot SeVeral clusTers with ThE sAmE cOloR. The cOLor pointS aRe aS lArgE as thE ClusteR size To maKe ThE PoiNts visiBLe. mORe thAn 80% Of ClusTer-PaRticlEs arE StiLl bound At 1 Gyr. \[fig:sNap\]](SNap0212nEw.EpS "fig:"){widTh="40mm"}
Results
=======
RaDiAl migratioN oF stAr clusTERs
---------------------------------
Figure \[Fig:snap\] shows the time evolUTion of tHe sTar clUsteRs and disk. altHough eAch SPiral aRm lookS like A sIngLE CoherENT sTruCtUre in each sNAPshOt, it iS tRansIent and Recurrent. They are woUNd uP due to the diffEreNtiaL ROtAtiON oF The GaLActIC Disk and break | ![Snapshots of the initial condition andat0.2 5, 0.5 , an d 1 Gyr. Grays cale shows the column dens ity o ft he d i sk , and colorp oi n t s i nd ic ate c l us ter p art icles.Since thes e a re only twelve co lors but 9 7 c lusters, weplo t seve ra l c l uster s w ith t he sam e color . The col or points are asl a rg e as the cluster size to make the point s visi bl e .M o retha n 80% of c lu ster- p article s a r e sti l l bound at 1Gyr. \[fig: s nap \]](sn ap 000 0 new.ep s "fi g: " ){w idth="40mm" }![S napshotsof the initial conditi on and at 0. 25,0 .5 ,and 1 Gyr . G ray sca le shows t he colu mn d e n s i ty o f t he d isk,and color poi nts ind i cat e clu sterpart ic les.Sincethese a re only twelvecolo rs but 97 cl us ter s, we p l ot sev era l c lusters with t h e s am e c ol or. The color poin ts a re as larg e as t h ecl u ster siz etomake t he po ints vi sible. M ore th a n80 % of cl us ter-pa rt icl esare s t illboundat 1 Gyr . \[f i g:snap\]](snap 0 052new.eps "f i g: " ) {w i dth= "40 mm"}![Snaps hots of t he i n it ial condi tionan d a t 0.25, 0.5, and 1 G yr . Gray scal e shows the c olumn dens i t y of thedisk , a n d color points indi cate clust e r partic les.Since th ese are o n l y twelve co lor s b ut9 7 c lusters, we p l o t se ve ral clu ste rs with th e s ame co lo r. The co lor poin ts a re a s l argea s the cl us ter s ize to m a ke the poin ts v is ib l e.More th a n8 0 % of c lu ster -pa rt icles are sti ll boun d at 1 Gy r.\ [fig :s na p\]](sn ap0106new.eps " fig:"){wid th ="4 0mm"}! [ S napshots of the initial conditi o n and a t 0 .25,0.5, and 1 Gy r.Gray s cal e shows the c olumn d ens i t y oft h edis k, and color p oin ts in di cate cluste r particles. Since the se are only t wel ve c o l or s b u t9 7 c lu s ter s , we plot severa l clusters w i th the samec olo r. The co lor poi nts a r e as la rge as th e cluster s izet o ma ke the poi nts visi ble. More than8 0% of c lus ter-pa rt icl es ar e stil l bo und a t 1 Gy r. \[fig :snap \] ](snap02 12new.eps "fig:"){width ="40mm "}
R esu lts
===== ==Rad ial migra tion of star c lus ter s
--- --- - ----- ---- - -- --- - ----- ---- -
Figure\ [f ig: s n ap \] shows th e t ime evol uti o n of t he s tar clusters andd isk. Althougheach s pir ala rm l oo ks like a sing leco h e rent str uc ture in eac h snapsh ot , it i s tran sientand rec u r re n t. The y ar e w ound up d ueto the dif fe re n tial r otat io n of t he gal a ctic d isk and break | ![Snapshots of_the initial_condition and at 0.25,_0.5, and_1_Gyr. Gray_scale_shows the column_density of the_disk, and color points_indicate cluster particles._Since_these are only twelve colors but 97 clusters, we plot several clusters with the_same_color. The_color_points_are as large as the_cluster size to make the_points visible._More than 80% of cluster-particles are still bound_at_1 Gyr. \[fig:snap\]](snap0000new.eps_"fig:"){width="40mm"}![Snapshots of the initial condition and at 0.25, 0.5,_and 1 Gyr. Gray scale shows_the column density_of_the_disk, and color points_indicate cluster particles. Since these are_only twelve colors but 97 clusters,_we plot several clusters with the same_color. The color points are as_large as the cluster size_to make_the points visible. More than_80% of cluster-particles_are still_bound at 1_Gyr. \[fig:snap\]](snap0052new.eps "fig:"){width="40mm"}![Snapshots of the initial_condition and at_0.25, 0.5, and 1 Gyr. Gray_scale_shows the column_density_of_the disk,_and color points_indicate_cluster particles._Since_these are only twelve colors but_97_clusters, we plot several clusters with the_same color. The color_points_are as large as_the cluster size to make_the points visible. More than 80%_of cluster-particles_are still_bound at 1 Gyr. \[fig:snap\]](snap0106new.eps "fig:"){width="40mm"}![Snapshots of the initial condition and_at 0.25, 0.5, and 1 Gyr._Gray scale shows the_column density_of_the disk, and_color_points indicate_cluster particles. Since these are only twelve_colors but_97 clusters, we plot several clusters_with the same color._The_color points are as large as_the cluster size to make the_points visible. More than 80%_of_cluster-particles_are still bound at 1_Gyr. \[fig:snap\]](snap0212new.eps "fig:"){width="40mm"}
Results
=======
Radial migration of star_clusters
---------------------------------
Figure \[fig:snap\] shows_the time evolution of the star clusters_and_disk. Although each spiral arm looks_like_a single coherent structure in each_snapshot,_it_is transient and recurrent. They_are wound up due to the_differential rotation of the galactic disk and break |
Rev D [**92**]{}, 084005 (2015);
A. Abdujabbarov, F. Atamurotov, N. Dadhich, B. Ahmedov, Z. Stuchlík, *Energetics and optical properties of 6-dimensional rotating black hole in pure Gauss-Bonnet gravity*, Eur. Phys. J. C [**75**]{}, 399 (2015);
F. Atamurotov, A. Abdujabbarov, B. Ahmedov, *Shadow of rotating non-Kerr black hole*, Phys. Rev D [**88**]{}, 064004 (2013).
N. Tsukamoto, *Black hole shadow in an asymptotically-flat, stationary, and axisymmetric spacetime: the Kerr-Newman and rotating regular black holes*, arXiv:1708.07427.
P. V. P. Cunha1, C. A. R. Herdeiro, *Shadows and strong gravitational lensing: a brief review*, arXiv:1801.00860.
P. V. P. Cunha, C. Herdeiro, E. Radu and H. F. Runarsson, *Shadows of Kerr black holes with scalar hair*, Phys. Rev. Lett. [**115**]{}, 211102 (2015), arXiv:1509.00021; P. V. P. Cunha, C. Herdeiro, E. Radu and H. F. Runarsson, *Shadows of Kerr black holes with and without scalar hair*, Int. J. Mod. Phys. D [**25**]{}, 1641021 (2016), arXiv:1605.08293.
F. H. Vincent, E. Gourgoulhon, C. Herdeiro and E. Radu, *Astrophysical imaging of Kerr black holes with scalar hair*, Phys. Rev. D [**94**]{}, 084045 (2016), arXiv:1606.04246.
P. V. P. Cunha, J. Grover, C. Herdeiro, E. Radu, H. Runarsson, and A. Wittig, *Chaotic lensing around boson stars and Kerr black holes with scalar hair*, Phys. Rev. D [**94**]{}, 104023 (2016).
J. O. Shipley, and S. R. Dolan, *Binary black hole shadows, chaotic scattering and the Cantor | Rev D [ * * 92 * * ] { }, 084005 (2015);
A. Abdujabbarov, F. Atamurotov, N. Dadhich, B. Ahmedov, Z. Stuchlík, * Energetics and optical properties of 6 - dimensional rotating black fix in arrant Gauss - Bonnet gravity *, Eur. Phys. J. C [ * * 75 * * ] { }, 399 (2015);
F. Atamurotov, A. Abdujabbarov, B. Ahmedov, * Shadow of rotating non - Kerr black trap *, Phys. Rev D [ * * 88 * * ] { }, 064004 (2013).
N. Tsukamoto, * Black hole tail in an asymptotically - bland, stationary, and axisymmetric spacetime: the Kerr - Newman and rotating even black holes *, arXiv:1708.07427.
P. V. P. Cunha1, C. A. R. Herdeiro, * Shadows and solid gravitational lensing: a brief review *, arXiv:1801.00860.
P. V. P. Cunha, C. Herdeiro, E. Radu and H. F. Runarsson, * Shadows of Kerr bootleg holes with scalar hair *, Phys. Rev. Lett. [ * * 115 * * ] { }, 211102 (2015), arXiv:1509.00021; P. V. P. Cunha, C. Herdeiro, E. Radu and H. F. Runarsson, * Shadows of Kerr black holes with and without scalar hair *, Int. J. Mod. Phys. D [ * * 25 * * ] { }, 1641021 (2016), arXiv:1605.08293.
F. H. Vincent, E. Gourgoulhon, C. Herdeiro and E. Radu, * Astrophysical imaging of Kerr black hole with scalar hair *, Phys. Rev. D [ * * 94 * * ] { }, 084045 (2016), arXiv:1606.04246.
P. V. P. Cunha, J. Grover, C. Herdeiro, E. Radu, H. Runarsson, and A. Wittig, * Chaotic lensing around boson stars and Kerr black hole with scalar hair *, Phys. Rev. D [ * * 94 * * ] { }, 104023 (2016).
J. O. Shipley, and S. R. Dolan, * Binary black hole shadows, chaotic dispersion and the Cantor | Reg D [**92**]{}, 084005 (2015);
A. Abdujabbarov, F. Auamurotov, N. Dadhiey, B. Ahkedov, A. Stuchlíy, *Energetics and optical pro'ertues od 6-dimensional rotating black hope in puee Geuss-Bonnet gravivg*, Eur. Pmvs. J. D [**75**]{}, 399 (2015);
F. Etamurotov, A. Abcujabbarov, B. Ahmedov, *Shagod lf rotating non-Kerr black hole*, Phys. Rev D [**88**]{}, 064004 (2013).
J. Tsukamoto, *Blwck njle ahadow in an asymptotically-flat, stztionarj, and axisymmetriv spacetime: the Kerr-Newman and rotating regular hlack holes*, arXyc:1708.07427.
P. V. P. Cunha1, C. A. R. Herdeiro, *Shadowa and strong gravitational lenskng: a brief revueq*, agFiv:1801.00860.
P. V. P. Cuiha, C. Rerdeiro, E. Radu and V. F. Runsrsson, *Shadows of Keer black holes with sralar hair*, Phys. Rev. Jett. [**115**]{}, 211102 (2015), asXnv:1509.00021; P. V. P. Cunha, C. Herdwieo, E. Tadu dnd F. F. Funzrxskn, *Shafowa of Kerr glack holes with and without svajqr hair*, Int. J. Mod. Pryf. D [**25**]{}, 1641021 (2016), arXiv:1605.08293.
F. H. Vincent, E. Gourgoulhon, C. Heddeiro and E. Radu, *Astrophysical imaging of Kgrr black roles with scalar hair*, Phys. Rev. D [**94**]{}, 084045 (2016), arXiv:1606.04246.
P. V. P. Cgnha, O. Erorcv, C. Fwrfeiro, E. Radu, H. Runarsson, and A. Wittig, *Chaotic memspng around boson ftars and Krrg njack holes wijh scalcd gair*, Phys. Rev. D [**94**]{}, 104023 (2016).
U. O. Shikley, abd S. R. Dojan, *Ninary black hole shadows, cyaotic scattvrint and the Cantor | Rev D [**92**]{}, 084005 (2015); A. Abdujabbarov, N. B. Ahmedov, Stuchlík, *Energetics and black in pure Gauss-Bonnet Eur. Phys. J. [**75**]{}, 399 (2015); F. Atamurotov, A. B. Ahmedov, *Shadow of rotating non-Kerr black hole*, Phys. Rev D [**88**]{}, 064004 N. Tsukamoto, *Black hole shadow in an asymptotically-flat, stationary, and axisymmetric spacetime: the and regular holes*, P. V. P. Cunha1, C. A. R. Herdeiro, *Shadows and strong gravitational lensing: a brief review*, P. V. P. Cunha, C. Herdeiro, E. Radu H. F. Runarsson, *Shadows Kerr black holes with scalar Phys. Lett. [**115**]{}, (2015), P. P. Cunha, C. E. Radu and H. F. Runarsson, *Shadows of Kerr black holes with and without scalar hair*, Int. Mod. Phys. 1641021 (2016), F. Vincent, Gourgoulhon, C. Herdeiro Radu, *Astrophysical imaging of Kerr black hair*, Phys. Rev. D [**94**]{}, 084045 (2016), arXiv:1606.04246. V. P. J. Grover, C. Herdeiro, E. Radu, Runarsson, and A. Wittig, *Chaotic lensing around boson and Kerr black holes with scalar hair*, Phys. Rev. D [**94**]{}, 104023 (2016). J. O. S. R. Dolan, *Binary hole shadows, chaotic and Cantor | Rev D [**92**]{}, 084005 (2015);
A. Abdujabbarov, F. AtamuroTov, N. DadhicH, B. AhmEdoV, Z. STuChlíK, *EneRgetics and optiCAl prOperties of 6-dimensional rOtatiNg BLack HOlE in puRe Gauss-bOnNET grAvItY*, EuR. PHYs. j. C [**75**]{}, 399 (2015);
F. AtAmuRotov, A. ABdujabbaroV, B. AHmEdov, *Shadow of ROtAting non-KeRr bLack hole*, Phys. rev d [**88**]{}, 064004 (2013).
N. TsukAmOto, *bLack hOle ShadoW in an aSYmptotIcally-flaT, sTAtionaRY, and axiSYMmEtriC spacetime: the Kerr-nEwMAn and rotating rEgular BlACk HOLes*, ArXIv:1708.07427.
P. V. P. Cunha1, c. A. r. HerdEIro, *ShadOWs AND StrONg gravitationAl lensing: a bRIef Review*, Arxiv:1801.00860.
p. v. P. CunhA, C. HerDeIRo, E. radu and H. F. RuNarsSon, *ShadowS of KerR Black hoLEs with sCalar hAir*, phyS. Rev. lEtT. [**115**]{}, 211102 (2015), aRXiV:1509.00021; P. v. p. CuNHa, c. HeRDeiRo, E. Radu aNd h. F. runarSson, *sHADOws oF KeRr blAck hoLes with and witHouT scaLAr hAir*, InT. J. Mod. phys. d [**25**]{}, 1641021 (2016), aRXiv:1605.08293.
F. h. VinceNt, E. GoUrGoulhon, C. HerdeirO and e. Radu, *AstrOphYsIcaL iMaginG Of Kerr BlaCk hOles witH scalar HAir*, phYS. rEv. d [**94**]{}, 084045 (2016), arXiv:1606.04246.
P. V. P. Cunha, J. GroVeR, c. heRdeiro, E. RAdu, H. RuNArSsON, and A. WitTiG, *ChAotiC LEnsinG aroUNd Boson staRs and KERr BlAck holeS wIth scaLaR haIr*, PHys. ReV. d [**94**]{}, 104023 (2016).
J. O. SHipley, And S. R. DolAn, *BinARy black hole shaDOws, chaotic scaTTeRINg ANd thE CaNtor | Rev D [**92**]{}, 084005(2015);
A . Abd uja bba ro v, F . At amurotov, N. D a dhic h, B. Ahmedov, Z. Stuc hlík, * E nerg e ti cs an d optic a lp r ope rt ie s o f6 -d imens ion al rota ting black ho le in pure Gau s s- Bonnet gra vit y*, Eur. Phy s.J. C [ ** 75* * ]{},399 (201 5);
F . Atamu rotov, A. A b dujabb a rov, B. A hm edov , *Shadow of rota t in g non-Kerr blac k hole *, Ph y s . R evD [**88**] {} , 064 0 04 (201 3 ). N . T s ukamoto, *Bla ck hole sha d owin anas ymp t otical ly-fl at , st ationary, a nd a xisymmetr ic spa c etime:t he Kerr -Newma n a ndrota t in greg ul a r b l ac k h o les *, arXiv :1 70 8.074 27.P . V. P . C unha 1, C. A. R. Herdei ro, *Sh a dow s and stro ng g ra vitat ionallensi ng : a brief revie w*,arXiv:180 1.0 08 60.
P. V. P. Cun ha, C. Herdei ro, E.R adu a n d H. F. Runarsson, *Sh ad o w sof Kerrblackh ol es with sca la r h air* , Phys. Rev . L ett. [** 115**] { }, 2 11102 ( 20 15), a rX iv: 150 9.000 2 1; P . V. P . Cunha, C. H e rdeiro, E. Rad u and H. F. Ru n ar s s on , *Sh ado ws of Kerrblac k hol es w i th an d with out s ca l ar hair*, Int. J. Mod. P hys. D [**2 5**]{}, 16410 21 (2016), a r Xiv:1605 .082 9 3.
F. H. Vincent , E.Gourgoulho n , C. Her deiro and E.Radu, *As t r ophysica l i mag ing of K er r black holes w ithsc alar ha ir* , Phys. Re v.D [ **9 4* *]{}, 084 045 (201 6) ,ar Xi v:1 606.0 4 246.
P. V . P .Cun ha, J . Grove r, C. Her de ir o , E . Radu, H. R unar ss on , an d A .Witti g, * C hao tic len sing arou ndb oson s ta rs andKerr black ho le s with sca la r h air*,P h ys. Rev. D [**94**]{}, 104023 ( 2 016).
J.O. Sh iple y, and S. R. Dolan , * B inaryblackholesh ado w s , cha o t ic sc at tering and t heCanto r | Rev_D [**92**]{},_084005 (2015);
A. Abdujabbarov, F._Atamurotov, N._Dadhich,_B. Ahmedov,_Z._Stuchlík, *Energetics and_optical properties of_6-dimensional rotating black hole_in pure Gauss-Bonnet_gravity*,_Eur. Phys. J. C [**75**]{}, 399 (2015);
F. Atamurotov, A. Abdujabbarov, B. Ahmedov, *Shadow of_rotating_non-Kerr black_hole*,_Phys._Rev D [**88**]{}, 064004 (2013).
N._Tsukamoto, *Black hole shadow in_an asymptotically-flat,_stationary, and axisymmetric spacetime: the Kerr-Newman and rotating_regular_black holes*, arXiv:1708.07427.
P._V. P. Cunha1, C. A. R. Herdeiro, *Shadows and_strong gravitational lensing: a brief review*,_arXiv:1801.00860.
P. V. P._Cunha,_C._Herdeiro, E. Radu and_H. F. Runarsson, *Shadows of Kerr_black holes with scalar hair*, Phys._Rev. Lett. [**115**]{}, 211102 (2015), arXiv:1509.00021; P._V. P. Cunha, C. Herdeiro, E._Radu and H. F. Runarsson,_*Shadows of_Kerr black holes with and_without scalar hair*,_Int. J._Mod. Phys. D_[**25**]{}, 1641021 (2016), arXiv:1605.08293.
F. H. Vincent,_E. Gourgoulhon, C._Herdeiro and E. Radu, *Astrophysical imaging_of_Kerr black holes_with_scalar_hair*, Phys._Rev. D [**94**]{},_084045_(2016), arXiv:1606.04246.
P._V._P. Cunha, J. Grover, C. Herdeiro,_E._Radu, H. Runarsson, and A. Wittig, *Chaotic_lensing around boson stars_and_Kerr black holes with_scalar hair*, Phys. Rev. D_[**94**]{}, 104023 (2016).
J. O. Shipley, and_S. R._Dolan, *Binary_black hole shadows, chaotic scattering and the Cantor |
Section \[sect:trans\] {#sect:transproofs}
==========================================
In this section, we give the proofs of the growth rates of $X$ stated in Section \[sect:trans\].
Proof of Theorem \[thm.bubbleaveexp\]
-------------------------------------
For $b/a\not\in\{-1,-2,...\}$, from, and, we can write $X$ according to $$\label{eq:stochdecomp}
X(t) = r_{3}(t)c_{3} + r_{4}(t)c_{4} + \sigma r_{3}(t)\int_{0}^{t}d_{3}(s)\,dB(s) + \sigma r_{4}(t)\int_{0}^{t}d_{4}(s)\,dB(s).$$ We have already deduced the asymptotic behaviour of $r_3$, $r_4$, $d_3$ and $d_4$ in,, and. We recapitulate their limiting behaviour now: $$\begin{aligned}
r_{3}(t) \sim a^{-1-\ba}t^{-1-\ba}, &\quad r_{4}(t)\sim \frac{1}{\Gamma(1+\ba)}{\mathrm{e}}^{a(1+t)}a^{\ba}t^{\ba},
\quad \text{ as $t\to\infty$},\\
d_{3}(s) \sim b \, a^{-1+\ba}s^{\ba}, &\quad d_{4}(s) \sim \Gamma(1+\ba)a^{-\ba}{\mathrm{e}}^{-a(1+s)}s^{-\ba}, \quad \text{ as $s\to\infty$}.\end{aligned}$$ Dividing across by $r_4(t)$ yields $$\label{eq:stochdecomp2}
\frac{X(t)}{r_4(t)} = \frac{r_3(t)}{r_4(t)}c_3 + c_4 + \sigma \frac{r_3(t)}{r_4(t)}\int_0^t d_3(s)\,dB(s) + \sigma\int_0^t d_4(s)\,dB(s).$$ The asymptotic behaviour of the first and last terms is readily estimated. Since $a>0$, $r | Section \[sect: trans\ ] { # sect: transproofs }
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
In this section, we give the proofs of the increase pace of $ X$ stated in Section \[sect: trans\ ].
Proof of Theorem \[thm.bubbleaveexp\ ]
-------------------------------------
For $ b / a\not\in\{-1,-2,... \}$, from, and, we can write $ X$ according to $ $ \label{eq: stochdecomp }
X(t) = r_{3}(t)c_{3 } + r_{4}(t)c_{4 } + \sigma r_{3}(t)\int_{0}^{t}d_{3}(s)\,dB(s) + \sigma r_{4}(t)\int_{0}^{t}d_{4}(s)\,dB(s).$$ We have already deduce the asymptotic behaviour of $ r_3 $, $ r_4 $, $ d_3 $ and $ d_4 $ in, , and. We recapitulate their restrict behaviour now: $ $ \begin{aligned }
r_{3}(t) \sim a^{-1-\ba}t^{-1-\ba }, & \quad r_{4}(t)\sim \frac{1}{\Gamma(1+\ba)}{\mathrm{e}}^{a(1+t)}a^{\ba}t^{\ba },
\quad \text { as $ t\to\infty$},\\
d_{3}(s) \sim b \, a^{-1+\ba}s^{\ba }, & \quad d_{4}(s) \sim \Gamma(1+\ba)a^{-\ba}{\mathrm{e}}^{-a(1+s)}s^{-\ba }, \quad \text { as $ s\to\infty$}.\end{aligned}$$ divide across by $ r_4(t)$ yields $ $ \label{eq: stochdecomp2 }
\frac{X(t)}{r_4(t) } = \frac{r_3(t)}{r_4(t)}c_3 + c_4 + \sigma \frac{r_3(t)}{r_4(t)}\int_0^t d_3(s)\,dB(s) + \sigma\int_0^t d_4(s)\,dB(s).$$ The asymptotic behaviour of the beginning and last terms is promptly estimated. Since $ a>0 $, $ r | Seftion \[sect:trans\] {#sect:transkroofs}
==========================================
In this secjiin, we jive ths proofs of the growth rates of $X$ steted in Stbtion \[sect:trans\].
Proof ow Theorem \[thm.bubboeavtexp\]
-------------------------------------
For $b/a\not\in\{-1,-2,...\}$, fckm, and, we can aritz $E$ according to $$\kabel{eq:stowhdecomp}
X(t) = f_{3}(t)e_{3} + r_{4}(t)c_{4} + \sigma r_{3}(t)\int_{0}^{t}d_{3}(s)\,dB(s) + \sigma r_{4}(t)\igt_{0}^{t}d_{4}(s)\,dB(x).$$ Ae have alreadi dedlcqd tgv csymptotic behaviour of $r_3$, $r_4$, $d_3$ ahd $d_4$ in,, and. We recapotulate their limiting behwviokr now: $$\begin{alignef}
r_{3}(t) \sim a^{-1-\ba}e^{-1-\va}, &\quad r_{4}(t)\sio \frac{1}{\Gamma(1+\ba)}{\mathrm{e}}^{a(1+t)}z^{\ba}t^{\ba},
\quad \text{ as $t\to\infth$},\\
b_{3}(s) \sim b \, a^{-1+\va}w^{\ba}, &\quad d_{4}(s) \sin \Gamia(1+\ba)a^{-\ba}{\mathrm{e}}^{-a(1+s)}s^{-\ba}, \xuad \tect{ as $s\to\infty$}.\cnd{almgnee}$$ Dividing across by $c_4(t)$ yields $$\label{eq:stoshdecomp2}
\frac{X(t)}{r_4(t)} = \frac{r_3(t)}{e_4(t)}x_3 + c_4 + \sigka \ffqc{r_3(g)}{r_4(t)}\jnv_0^t s_3(s)\,dB(s) + \sijma\int_0^t d_4(s)\,dG(s).$$ The asymptotic behaviour of trv first and lzst tewmf is readily estimated. Since $a>0$, $r | Section \[sect:trans\] {#sect:transproofs} ========================================== In this section, the of the rates of $X$ of \[thm.bubbleaveexp\] ------------------------------------- For from, and, we write $X$ according to $$\label{eq:stochdecomp} X(t) r_{3}(t)c_{3} + r_{4}(t)c_{4} + \sigma r_{3}(t)\int_{0}^{t}d_{3}(s)\,dB(s) + \sigma r_{4}(t)\int_{0}^{t}d_{4}(s)\,dB(s).$$ We have already deduced asymptotic behaviour of $r_3$, $r_4$, $d_3$ and $d_4$ in,, and. We recapitulate their behaviour $$\begin{aligned} \sim &\quad r_{4}(t)\sim \frac{1}{\Gamma(1+\ba)}{\mathrm{e}}^{a(1+t)}a^{\ba}t^{\ba}, \quad \text{ as $t\to\infty$},\\ d_{3}(s) \sim b \, a^{-1+\ba}s^{\ba}, &\quad d_{4}(s) \sim \Gamma(1+\ba)a^{-\ba}{\mathrm{e}}^{-a(1+s)}s^{-\ba}, \text{ as $s\to\infty$}.\end{aligned}$$ Dividing across by $r_4(t)$ yields \frac{X(t)}{r_4(t)} = \frac{r_3(t)}{r_4(t)}c_3 + + \sigma \frac{r_3(t)}{r_4(t)}\int_0^t d_3(s)\,dB(s) + d_4(s)\,dB(s).$$ asymptotic behaviour the and terms is readily Since $a>0$, $r | Section \[sect:trans\] {#sect:transProofs}
==========================================
In thIs secTioN, we GiVe thE proOfs of the growth RAtes Of $X$ stated in Section \[sect:Trans\].
prOOf of tHeOrem \[tHm.bubblEAvEEXp\]
-------------------------------------
FOr $B/a\Not\In\{-1,-2,...\}$, FRoM, and, wE caN write $X$ According tO $$\laBeL{eq:stochdecoMP}
X(T) = r_{3}(t)c_{3} + r_{4}(t)c_{4} + \sigMa r_{3}(T)\int_{0}^{t}d_{3}(s)\,dB(s) + \siGma R_{4}(t)\int_{0}^{t}D_{4}(s)\,DB(s).$$ wE have AlrEady dEduced THe asymPtotic behAvIOur of $r_3$, $R_4$, $D_3$ and $d_4$ in,, AND. WE recApitulate their limITiNG behaviour now: $$\bEgin{alIgNEd}
R_{3}(T) \Sim A^{-1-\ba}T^{-1-\ba}, &\quad r_{4}(t)\sIm \Frac{1}{\GAMma(1+\ba)}{\maTHrM{E}}^{A(1+T)}a^{\bA}T^{\ba},
\quad \text{ as $T\to\infty$},\\
d_{3}(s) \sIM b \, a^{-1+\Ba}s^{\ba}, &\qUaD d_{4}(s) \SIm \GammA(1+\ba)a^{-\bA}{\mAThrM{e}}^{-a(1+s)}s^{-\ba}, \quad \Text{ As $s\to\inftY$}.\end{alIGned}$$ DivIDing acrOss by $r_4(T)$ yiEldS $$\labEL{eQ:sTocHdEComP2}
\FrAc{X(T)}{R_4(t)} = \fRac{r_3(t)}{r_4(t)}c_3 + C_4 + \sIgMa \fraC{r_3(t)}{r_4(T)}\INT_0^T d_3(s)\,db(s) + \sIgma\Int_0^t d_4(S)\,dB(s).$$ The asymptOtiC behAVioUr of tHe firSt anD lAst teRms is rEadilY eStimated. Since $a>0$, $r | Section \[sect:trans\] {# sect:trans proof s}=== == ==== ==== ============== = ==== ==========
In this se ction ,w e gi v ethe p roofs o f t h e gr ow th ra te s o f $X$ st ated in Section \ [se ct :trans\].
P r oo f of Theor em\[thm.bubble ave exp\]-- --- - ----- --- ----- ------ - ------ -----
Fo r$ b/a\no t \in\{-1 , - 2, ...\ }$, from, and, we ca n write $X$ acc ording t o $ $ \ lab el{ eq:stochde co mp}
X(t)= r _ { 3 }(t ) c_{3} + r_{4} (t)c_{4} +\ sig ma r_{ 3} (t) \ int_{0 }^{t} d_ { 3}( s)\,dB(s) + \si gma r_{4} (t)\in t _{0}^{t } d_{4}(s )\,dB( s). $$We h a ve a lre ad y de d uc edt heasymptot ic b ehavi ouro f $ r_3$ , $ r_4$ , $d_ 3$ and $d_4$in, , an d . W e rec apitu late t heirlimiti ng be ha viour now: $$\b egin {aligned}
r_{ 3} (t) \ s im a^{ -1- \ba }t^{-1- \ba}, & \ qua dr _ { 4} (t)\sim \frac{1}{\ Ga m m a( 1+\ba)}{ \mathr m {e }} ^ {a(1+t)} a^ {\b a}t^ { \ ba}, \q uad \tex t{ as$ t\ to \infty$ }, \\
d _{3 }(s ) \si m b \ , a^{- 1+\ba}s^ {\ba} , &\quad d_{4} ( s) \sim \Gamm a (1 + \ ba ) a^{- \ba }{\mathrm{e }}^{ - a(1+ s)}s ^ {- \ba } , \qu ad \t ex t {a s $s\to\infty$}.\en d{ aligne d}$$Dividing acro ss by $r_4 ( t ) $ yields $$\ l ab e l{eq:stochdeco mp2} \frac{ X (t)}{r_4 (t)}= \frac{ r_3(t)}{r _ 4 (t)}c_3+ c _4+ \ sig m a \ frac{r_3(t)}{ r _ 4(t) }\ int_0^t d_ 3(s)\,d B(s ) + \s igm a\ int_0^t d _4(s)\,d B( s) .$ $The asym p totic be ha vio ur of thef irst a nd la st t er ms isreadily es t i mate d. S ince $a >0 $, $r | Section \[sect:trans\]_{#sect:transproofs}
==========================================
In this_section, we give the_proofs of_the_growth rates_of_$X$ stated in_Section \[sect:trans\].
Proof of Theorem_\[thm.bubbleaveexp\]
-------------------------------------
For $b/a\not\in\{-1,-2,...\}$, from, and,_we can write_$X$_according to $$\label{eq:stochdecomp}
X(t) = r_{3}(t)c_{3} + r_{4}(t)c_{4} + \sigma r_{3}(t)\int_{0}^{t}d_{3}(s)\,dB(s)_+_\sigma r_{4}(t)\int_{0}^{t}d_{4}(s)\,dB(s).$$_We_have_already deduced the asymptotic behaviour_of $r_3$, $r_4$, $d_3$ and_$d_4$ in,,_and. We recapitulate their limiting behaviour now: $$\begin{aligned}
__ r_{3}(t)_\sim a^{-1-\ba}t^{-1-\ba}, &\quad r_{4}(t)\sim \frac{1}{\Gamma(1+\ba)}{\mathrm{e}}^{a(1+t)}a^{\ba}t^{\ba},
_\quad \text{ as $t\to\infty$},\\
_ d_{3}(s) \sim_b_\,_a^{-1+\ba}s^{\ba}, &\quad d_{4}(s)_\sim \Gamma(1+\ba)a^{-\ba}{\mathrm{e}}^{-a(1+s)}s^{-\ba}, \quad \text{ as $s\to\infty$}.\end{aligned}$$_Dividing across by $r_4(t)$ yields $$\label{eq:stochdecomp2}
_ \frac{X(t)}{r_4(t)} = \frac{r_3(t)}{r_4(t)}c_3 +_c_4 + \sigma \frac{r_3(t)}{r_4(t)}\int_0^t d_3(s)\,dB(s) +_\sigma\int_0^t d_4(s)\,dB(s).$$ The asymptotic behaviour_of the_first and last terms is_readily estimated. Since_$a>0$, $r |
\approx \frac{dl}{c}+\frac{v dl \cos(\phi) }{c^2}+...= \frac{dl}{c}+\frac{{\bf v} \cdot {\bf dl} }{c^2}+...$$ We ignore Lorentz length contraction of the slab as this only contributes at 2nd order in $v/c$. Summing over paths to get total travel time from $A$ to $B$ $$\begin{aligned}
t_{AB} &=&\int_A^B \frac{dl}{c}+\int_A^B \frac{{\bf v} \cdot {\bf dl} }{c^2}+... \nonumber \\
&=& \frac{l}{c} + \frac{L v \cos(\theta)}{c^2}+...\nonumber \\
&=& \frac{n L}{c} + \frac{L v \cos(\theta)}{c^2}+...,
\label{eqn:FD}\end{aligned}$$ where $L$ is the straight line distance from $A$ to $B$ in the matter rest frame, and $n=l/L$ defines the refractive index of the dielectric in this treatment, as when the dielectric is at rest the effective speed of EM radiation through matter in a straight line from $A$ to $B$ is defined to be $c/n$. Note that $t_{AB}$ does not involve $n$ in the $v$ dependent 2nd term. This effect is actually the reason for the Fresnel drag formalism. The macroscopic treatment, which leads to the Fresnel drag formalism, involves the sum $ |{\bf L}^\prime |=|{\bf L}+{\bf v}t_{AB}|$, for the macrosocopic distance traveled, which gives for the travel time $$\begin{aligned}
t_{AB}&=&\frac{L^\prime}{V}\approx\frac{L}{V(v\cos(\theta))} +\frac{v \cos(\theta)t_{AB}}{V(v\cos(\theta))}, \mbox{\ \ \ giving} \nonumber \\
t_{AB}&=&\displaystyle{\frac{L}{V(v\cos(\theta))} +\frac{L v \cos(\theta)}{V(v\cos(\theta))^2}}+...\end{aligned}$$ where $V(v\cos(\theta))$ is the effective linear speed of EM radiation in direction $AB$ at angle | \approx \frac{dl}{c}+\frac{v dl \cos(\phi) } { c^2}+... = \frac{dl}{c}+\frac{{\bf v } \cdot { \bf dl } } { c^2}+... $$ We ignore Lorentz length contraction of the slab as this only contribute at second order in $ v / c$. Summing over path to receive total travel prison term from $ A$ to $ B$ $ $ \begin{aligned }
t_{AB } & = & \int_A^B \frac{dl}{c}+\int_A^B \frac{{\bf v } \cdot { \bf dl } } { c^2}+... \nonumber \\
& = & \frac{l}{c } + \frac{L v \cos(\theta)}{c^2}+... \nonumber \\
& = & \frac{n L}{c } + \frac{L v \cos(\theta)}{c^2}+... ,
\label{eqn: FD}\end{aligned}$$ where $ L$ is the neat line distance from $ A$ to $ B$ in the topic remainder frame, and $ n = lambert / L$ defines the refractive exponent of the dielectric in this treatment, as when the insulator is at rest the effective focal ratio of EM radiation sickness through matter in a straight line from $ A$ to $ B$ is define to be $ c / n$. Note that $ t_{AB}$ does not involve $ n$ in the $ v$ dependent second term. This impression is actually the reason for the Fresnel drag formalism. The macroscopic treatment, which leave to the Fresnel drag formalism, involves the sum $ |{\bf L}^\prime |=|{\bf L}+{\bf v}t_{AB}|$, for the macrosocopic distance traveled, which gives for the change of location time $ $ \begin{aligned }
t_{AB}&=&\frac{L^\prime}{V}\approx\frac{L}{V(v\cos(\theta) ) } + \frac{v \cos(\theta)t_{AB}}{V(v\cos(\theta) ) }, \mbox{\ \ \ giving } \nonumber \\
t_{AB}&=&\displaystyle{\frac{L}{V(v\cos(\theta) ) } + \frac{L v \cos(\theta)}{V(v\cos(\theta))^2}}+... \end{aligned}$$ where $ V(v\cos(\theta))$ is the effective linear speed of EM radiation in direction $ AB$ at angle | \aporox \frac{dl}{c}+\frac{v dl \cos(\khi) }{c^2}+...= \frac{dl}{c}+\frae{{\vf v} \cvot {\bf dl} }{c^2}+...$$ De ignore Lorentz length convracrion if the slab as this onuy contrihutes at 2nd ieder in $v/c$. Summing over lwths vo get total trsvel time xrom $A$ to $B$ $$\beciv{apigned}
t_{AB} &=&\int_A^B \frac{dl}{c}+\int_A^B \frac{{\bf v} \cdot {\hf dl} }{c^2}+... \nonuiber \\
&=& \frad{l}{c} + \frac{L v \cos(\theta)}{c^2}+...\nonumber \\
&=& \frzc{n L}{c} + \frac{L v \cos(\tneta)}{c^2}+...,
\label{eqn:FD}\end{aligned}$$ aherf $L$ is the straighh line distqnce drom $A$ to $B$ kn the matuex rest framg, and $n=l/L$ defines the refractive inder of the diglschtic in this vreatmvnt, as when tmv dielewtric ix at rest the cffecvive speed of EM radiatioi through matter in w straighd mine from $A$ to $B$ us defited do bd $c/n$. Noue vhaf $t_{AB}$ foea not invomve $n$ in thw $v$ dependent 2nd tetm. Nnis effect ia actuwljy the reason for the Fresnel drag formdliam. The macroscopic trearment, which leads to jhe Fresnej drag formalism, involves the sum $ |{\bf L}^\prime |=|{\bf N}+{\bf v}v_{AC}|$, fir gye macrosocopic distance traveled, which gives fjd uhe travel time $$\bcgin{aligned}
t_{AB}&=&\frac{K^\pgikg}{V}\approx\frac{L}{V(x\cos(\thzfa))} +\frac{v \cos(\theta)t_{AB}}{G(v\cos(\thgta))}, \mvox{\ \ \ givyng} \monumber \\
t_{AB}&=&\displaystyle{\frax{L}{V(v\cos(\theta))} +\drac{L v \cos(\theta)}{V(v\eos(\theta))^2}}+...\end{auigngd}$$ whete $V(v\cos(\theta))$ is the efyectivs linear spfed of EM fadiation in dirdctpon $DB$ at angle | \approx \frac{dl}{c}+\frac{v dl \cos(\phi) }{c^2}+...= \frac{dl}{c}+\frac{{\bf v} dl} We ignore length contraction of contributes 2nd order in Summing over paths get total travel time from $A$ $B$ $$\begin{aligned} t_{AB} &=&\int_A^B \frac{dl}{c}+\int_A^B \frac{{\bf v} \cdot {\bf dl} }{c^2}+... \nonumber \\ \frac{l}{c} + \frac{L v \cos(\theta)}{c^2}+...\nonumber \\ &=& \frac{n L}{c} + \frac{L v \cos(\theta)}{c^2}+..., where is straight distance from $A$ to $B$ in the matter rest frame, and $n=l/L$ defines the refractive index the dielectric in this treatment, as when the is at rest the speed of EM radiation through in straight line $A$ $B$ defined to be Note that $t_{AB}$ does not involve $n$ in the $v$ dependent 2nd term. This effect is actually reason for drag formalism. macroscopic which to the Fresnel involves the sum $ |{\bf L}^\prime for the macrosocopic distance traveled, which gives for travel time t_{AB}&=&\frac{L^\prime}{V}\approx\frac{L}{V(v\cos(\theta))} +\frac{v \cos(\theta)t_{AB}}{V(v\cos(\theta))}, \mbox{\ \ \ \nonumber \\ t_{AB}&=&\displaystyle{\frac{L}{V(v\cos(\theta))} +\frac{L v \cos(\theta)}{V(v\cos(\theta))^2}}+...\end{aligned}$$ where $V(v\cos(\theta))$ the effective linear speed of EM radiation in direction $AB$ at angle | \approx \frac{dl}{c}+\frac{v dl \cos(\phI) }{c^2}+...= \frac{dl}{c}+\fRac{{\bf V} \cdOt {\bF dL} }{c^2}+...$$ We IgnoRe Lorentz lengtH ContRaction of the slab as this Only cOnTRibuTEs At 2nd oRder in $v/C$. suMMIng OvEr PatHs TO gEt totAl tRavel tiMe from $A$ to $B$ $$\BegIn{Aligned}
t_{AB} &=&\inT_a^B \Frac{dl}{c}+\int_a^B \fRac{{\bf v} \cdot {\bf Dl} }{c^2}+... \NonumbEr \\
&=& \FraC{L}{c} + \fraC{L v \Cos(\thEta)}{c^2}+...\noNUmber \\
&=& \fRac{n L}{c} + \fraC{L V \Cos(\theTA)}{c^2}+...,
\label{EQN:Fd}\end{Aligned}$$ where $L$ is thE StRAight line distaNce froM $A$ TO $B$ IN The MatTer rest fraMe, And $n=l/l$ Defines THe REFRacTIve index of the Dielectric iN ThiS treatMeNt, aS When thE dielEcTRic Is at rest the EffeCtive speeD of EM rADiation THrough mAtter iN a sTraIght LInE fRom $a$ tO $b$ is DEfIneD To bE $c/n$. Note tHaT $t_{aB}$ doeS not INVOLve $n$ In tHe $v$ dEpendEnt 2nd term. This EffEct iS ActUally The reAson FoR the FResnel Drag fOrMalism. The macrosCopiC treatmenT, whIcH leAdS to thE fresneL drAg fOrmalisM, involvES thE sUM $ |{\BF L}^\Prime |=|{\bf L}+{\bf v}t_{AB}|$, for tHe MACrOsocopic DistanCE tRaVEled, whicH gIveS for THE travEl tiME $$\bEgin{aligNed}
t_{AB}&=&\FRaC{L^\Prime}{V}\aPpRox\fraC{L}{v(v\cOs(\tHeta))} +\fRAc{v \cOs(\thetA)t_{AB}}{V(v\coS(\thetA))}, \Mbox{\ \ \ giving} \nonuMBer \\
t_{AB}&=&\displaySTyLE{\FrAC{L}{V(v\Cos(\Theta))} +\frac{L v \Cos(\tHEta)}{V(V\cos(\THeTa))^2}}+...\eND{aligNed}$$ whErE $v(v\COs(\theta))$ is the effectiVe Linear Speed Of EM radiation In directioN $ab$ At angle | \approx \frac{dl}{c}+\fra c{v dl \co s(\ph i) }{ c^ 2}+. ..=\frac{dl}{c}+\ f rac{ {\bf v} \cdot {\bf dl } } {c ^ 2}+. . .$ $ WeignoreL or e n tzle ng thco n tr actio n o f the s lab as thi s o nl y contribute s a t 2nd orde r i n $v/c$. Sum min g over p ath s to g ettotal trave l timefrom $A$to $B$ $$ \ begin{a l i gn ed}t_{AB} &=&\int_A^ B \ f rac{dl}{c}+\in t_A^B\f r ac { { \bf v} \cdot {\ bf dl} }{c^2} + .. . \ non u mber \\
&=& \ frac{l}{c}+ \f rac{Lv\co s (\thet a)}{c ^2 } +.. .\nonumber\\
& =& \frac{ n L}{c } + \fra c {L v \c os(\th eta )}{ c^2} + .. .,
\l ab e l{e q n: FD} \ end {aligned }$ $where $L$ i s thestr aigh t lin e distance fr om$A$t o $ B$ in thematt er rest frame , and $ n=l/L$ definestherefractiv e i nd exof thed ielect ric in this t reatmen t , a sw h e nthe dielectric isat r es t the ef fectiv e s pe e d of EMra dia tion t hroug h ma t te r in a s traigh t l in e from$A $ to $ B$ is de fined to b e $c/n $. Notethat$ t_{AB}$ does n o t involve $n$ in t he $v$dep endent 2ndterm . Thi s ef f ec t i s actu allyth e r e ason for the Fresne ldrag f ormal ism. The macr oscopic tr e a t ment, wh ichl ea d s to the Fresn el dr ag formali s m, invol ves t he sum $ |{\bf L} ^ \ prime |= |{\ bf L} +{\ b f v }t_{AB}|$, fo r thema crosoco pic distan cetra vel ed, w hich give s for th etr av el ti me $$ \ begin{al ig ned }t_{ AB}&= & \frac{ L^\pr ime} {V }\ a ppr ox\frac { L} { V (v\c os (\ thet a)) }+\fra c{v\ cos (\theta )t_{AB}}{ V(v \ cos( \t he ta))}, \mbox{\ \ \gi ving} \non um ber \\
t_ { A B}&=&\di splaystyle{\frac{L}{V(v \ cos(\th eta ))} + \fra c{L v \co s(\ theta) }{V ( v\cos( \theta ))^2} }+ ... \ e nd{al i g ne d}$ $where $V(v \ c os( \thet a) )$ i s the e ffective linear sp e edof EM radiati onin d i r ec tio n $ A B$at ang l e | \approx_\frac{dl}{c}+\frac{v dl_\cos(\phi) }{c^2}+...= \frac{dl}{c}+\frac{{\bf_v} \cdot__{\bf dl}__ }{c^2}+...$$ We_ignore Lorentz length_contraction of the slab_as this only_contributes_at 2nd order in $v/c$. Summing over paths to get total travel time from_$A$_to $B$_$$\begin{aligned}
t_{AB}_&=&\int_A^B_\frac{dl}{c}+\int_A^B \frac{{\bf v} \cdot _{\bf dl} }{c^2}+..._\nonumber \\
&=&_\frac{l}{c} + \frac{L v \cos(\theta)}{c^2}+...\nonumber \\
&=& \frac{n L}{c}_+_\frac{L v \cos(\theta)}{c^2}+...,
\label{eqn:FD}\end{aligned}$$_where $L$ is the straight line distance from $A$_to $B$ in the matter rest_frame, and $n=l/L$_defines_the_refractive index of the_dielectric in this treatment, as when_the dielectric is at rest the_effective speed of EM radiation through matter_in a straight line from $A$_to $B$ is defined to_be $c/n$._Note that $t_{AB}$ does not_involve $n$ in_the $v$_dependent 2nd term._This effect is actually the reason_for the Fresnel_drag formalism. The macroscopic treatment, which_leads_to the Fresnel_drag_formalism,_involves the_sum $ |{\bf_L}^\prime_|=|{\bf _L}+{\bf_v}t_{AB}|$, for the macrosocopic distance traveled,_which_gives for the travel time $$\begin{aligned}
t_{AB}&=&\frac{L^\prime}{V}\approx\frac{L}{V(v\cos(\theta))} +\frac{v_\cos(\theta)t_{AB}}{V(v\cos(\theta))}, \mbox{\ \_\_giving} \nonumber \\
t_{AB}&=&\displaystyle{\frac{L}{V(v\cos(\theta))} +\frac{L_v \cos(\theta)}{V(v\cos(\theta))^2}}+...\end{aligned}$$ where $V(v\cos(\theta))$ is_the effective linear speed of EM_radiation in_direction $AB$_at angle |
of viscosity in the disc. This is supported by the result that the Poisson noise for the low-resolution models that we have performed multiple times is less than 5%, see table \[tb:gadget\_eda1\]. Furthermore, the ISM loading rate is robust for all our simulations and does not seem to be affected by numerical noise.
Modelling of the ISM
--------------------
In order to be able to discern the relevant physical processes from the simulations, we have modelled the ISM in a very idealized way: as a homogeneous gas with no clumps and no turbulence. To a certain extent we have modelled the reaction of the disc to a density gradient when the flow first hits the disc. In that case the size of the disc is determined by the ram pressure. If the disc were to encounter a region of gas with lower density, the surface area of the disc can extend to larger radii without being stripped and that would probably increase its effective cross section. Despite the lower density, the ISM loading rate could still be of the same order, because a lower ISM density allows a larger surface area of the disc to collect ISM. In that sense, accretion onto protoplanetary discs may be a self-regulating process. In a follow-up work we will perform simulations with different ISM densities and velocities to investigate if and how the ISM loading rate depends on these quantities.
In principle it would be possible to add angular momentum to the disc, such that it maintains or prolongs its lifetime, if the ISM has turbulence or substructures on scales at or slightly smaller than the disc scale. However, even in that case the net effect is unlikely to increase the overall lifetime, since the disc would quickly encounter a different part of the ISM where an adverse, disruptive configuration of substructure might exist.
Comparison to other work {#sec:dis_m09_eda1}
------------------------
A similar simulation has been performed by M09, but for a higher disc mass and lower velocity of the ISM with respect to the disc. In their work both the Bondi-Hoyle radius and the radius beyond which ram pressure would dominate are much larger than the radius of the disc. They used a higher resolution for the disc, initially $\sim 2.5 \times 10^5$ particles, which had 8 times the mass of their ISM particles. However, in their work the ISM also does not reside in the outer regions of the disc (see their Fig | of viscosity in the disc. This is supported by the result that the Poisson noise for the depleted - settlement models that we have performed multiple times is less than 5% , witness table \[tb: gadget\_eda1\ ]. Furthermore, the ISM loading pace is robust for all our simulations and does not seem to be affect by numerical noise.
Modelling of the ISM
--------------------
In order to be able to spot the relevant physical processes from the simulations, we have modelled the doctrine in a very idealized way: as a homogeneous gas with no thump and no turbulence. To a certain extent we have modelled the chemical reaction of the disc to a density gradient when the flow foremost hits the disc. In that case the size of the disc is determined by the random-access memory pressure. If the disc were to encounter a region of gas with low density, the surface area of the disc can extend to large radii without being stripped and that would probably increase its effective cross section. Despite the lower density, the ISM loading rate could still be of the like ordering, because a lower doctrine concentration allows a larger surface area of the magnetic disk to collect ISM. In that sense, accretion onto protoplanetary discs may be a self - regulating procedure. In a follow - up work we will perform simulations with different ISM concentration and velocities to investigate if and how the ISM loading rate depends on these quantity.
In principle it would be potential to lend angular momentum to the disc, such that it maintains or prolongs its lifetime, if the ISM have turbulence or substructures on scales at or slightly little than the disk scale. However, even in that case the final effect is unlikely to increase the overall lifetime, since the disc would quickly meet a different region of the ISM where an adverse, disruptive configuration of infrastructure might exist.
Comparison to other oeuvre { # sec: dis_m09_eda1 }
------------------------
A similar simulation has been performed by M09, but for a higher disc batch and lower speed of the ISM with respect to the disc. In their work both the Bondi - Hoyle spoke and the radius beyond which ram pressure would dominate are a lot larger than the radius of the disc. They used a higher resolution for the disc, initially $ \sim 2.5 \times 10 ^ 5 $ particle, which had 8 times the mass of their ISM particles. However, in their work the ISM also does not occupy in the out regions of the disc (visualize their Fig | of viscosity in the disc. Tmis is supported by the resulf that tfe Poisson noise for the low-cesooutiob models that we have oerformed multiplw tinws is less than 5%, scz tabmc \[tb:gcdjet\_eda1\]. Furthermpre, the ISK loading rate ir xobust for all our simulations and djes not sfem to be affested fy nhmerical noise.
Modelling of the ISM
--------------------
Jn ordeg to be able to doscern the relevant physicwl pgocesses from the dimulations, we rqve modelled the ISM in a very idezlized way: as a homogeneous gas with no clumps abd jm turbulencx. To a certain extcmt we vave mocelled the reagtion of the disc to a densitb gradient when the slow firsd kits the disc. In that cqse tve shze ud tfe sixc is deherjined by tge ram preswure. If the disc wete np encounter z regijn of gas with lower density, the surface drez of the disc can extene to larger radii witjout beind stripped and that would probably increase its exfectmvd cxiss sdxtlon. Despite the lower density, the ISM loading daue bould still be of the same otdfr, fecause a lowgr ISM bsnaity allows a largfr surfwce aeea of tht disv to collect ISM. In that sebse, accretiou obto protoplanetary discs may ye a sglf-regilating process. In a foulow-hp work we aill perfkfm simulations wkth dhfferent ISM densities and velocitixs to investkgatg if anq how the LSM loading rate depends on tkese xuantities.
Ln principle it would be possible to add angukas mmmentum co the disc, such thwt it maintainx or prjlongr its lifenime, if tie ISM has ttrbulence or vobstructures mn scalef at or wlightlh smaller than yhe disc scale. Howecer, even in that cese jhs net effect is unoikely to incresse thq lvxrall nifetime, sinwe tfe aosc wuuld quickln evcoumter a different pard of the ISM where an sdyerse, distuptive cjnfiguration pf substructure mihht eeist.
Cooparixon to other work {#sec:dis_m09_eda1}
------------------------
A simjlar simupatlon has been [erfirmed by M09, bbt for a higher disc mass and lower velority of the ISM with rgspect to the disc. In uheir work bovh the Bondi-Hoyne radius and the raeius beyond which ram pressure would dkminata are much larger than the radius of the disc. They used a higher resolution foe the vifc, initiallg $\sik 2.5 \tikef 10^5$ partycled, which had 8 timex the mass of their ISM particlxs. However, iu their work the ISM also doex vot reside in the outer regions of ths disc (sre their Fig | of viscosity in the disc. This is the that the noise for the performed times is less 5%, see table Furthermore, the ISM loading rate is for all our simulations and does not seem to be affected by numerical Modelling of the ISM -------------------- In order to be able to discern the physical from simulations, have modelled the ISM in a very idealized way: as a homogeneous gas with no clumps no turbulence. To a certain extent we have the reaction of the to a density gradient when flow hits the In case size of the is determined by the ram pressure. If the disc were to encounter a region of gas with density, the of the can to radii without being that would probably increase its effective the lower density, the ISM loading rate could be of same order, because a lower ISM allows a larger surface area of the disc collect ISM. In that sense, accretion onto protoplanetary discs may be a self-regulating process. In work we will perform with different ISM and to if how the loading rate depends on these quantities. In principle it would be to add angular momentum to the disc, such that it prolongs lifetime, if the has turbulence or substructures scales or slightly smaller than scale. even the effect unlikely to increase the lifetime, since the disc would encounter a different part adverse, disruptive configuration of substructure might exist. Comparison other work {#sec:dis_m09_eda1} ------------------------ A similar simulation been performed by M09, but for a higher disc mass and lower of the respect to the disc. In their work both Bondi-Hoyle radius and the beyond which ram pressure would dominate are much larger the of the They used a resolution for the initially $\sim 2.5 particles, which 8 the ISM particles. However, in their work ISM does not reside in the regions disc (see their Fig | of viscosity in the disc. This iS supported By the ResUlt ThAt thE PoiSson noise for thE Low-rEsolution models that we hAve peRfORmed MUlTiple Times is LEsS THan 5%, SeE tAblE \[tB:GaDget\_eDa1\]. FUrthermOre, the ISM lOadInG rate is robusT FoR all our simUlaTions and does Not Seem to Be AffECted bY nuMericAl noisE.
modellIng of the IsM
--------------------
iN order TO be able TO DiScerN the relevant physiCAl PRocesses from thE simulAtIOnS, WE haVe mOdelled the iSm in a vERy idealIZeD WAY: as A Homogeneous gaS with no clumPS anD no turBuLenCE. To a ceRtain ExTEnt We have modelLed tHe reactioN of the DIsc to a dENsity grAdient WheN thE floW FiRsT hiTs THe dISc. in tHAt cAse the siZe Of The diSc is DETERminEd bY the Ram prEssure. If the diSc wEre tO EncOunteR a regIon oF gAs witH lower DensiTy, The surface area oF the Disc can exTenD tO laRgEr radII withoUt bEinG strippEd and thAT woUlD PRObAbly increase its effEcTIVe Cross secTion. DeSPiTe THe lower dEnSitY, the ism loadIng rATe Could stiLl be of THe SaMe order, BeCause a LoWer iSM DensiTY allOws a laRger surfAce arEA of the disc to coLLect ISM. In that SEnSE, AcCRetiOn oNto protoplaNetaRY disCs maY Be A seLF-reguLatinG pROcESs. In a follow-up work we WiLl perfOrm siMulations with Different Ism DEnsities And vELoCIties to investiGate iF and how the isM loadinG rate Depends oN these quaNTIties.
In pRinCipLe iT woULD bE possible to adD ANgulAr MomentuM to The disc, SucH thAt iT maInTains or prOlongs itS lIfEtImE, if The ISm Has turbuLeNce Or SubStrucTUres on ScaleS at oR sLiGHtlY smalleR ThAN The dIsC sCale. howEvEr, eveN in tHAt cAse the nEt effect iS unLIkelY tO iNcrease The overall lifEtIme, since thE dIsc Would qUICkly encoUnter a different part of thE iSM wherE an AdverSe, diSruptive cOnfIguratIon OF substRucturE mighT eXisT.
cOmparISOn To oThEr work {#sec:dIS_M09_edA1}
------------------------
A simIlAr siMulatioN has been performed bY m09, buT for a higher diSc mAss aND LoWer VElOCitY oF The ism with respect to tHe disc. In thEiR WoRk both the BONdi-hoYle radiUs and thE radiUS beyond Which ram pRessure woUlD domINAte Are much larGer than tHe radius oF The diSC. THey usEd a Higher ReSolUtion For the DIsc, InitiAlly $\siM 2.5 \tImes 10^5$ paRticlEs, Which had 8 Times the mass of their ISM pArticlEs. HowEveR, in their wOrk THe IsM also doeS not Reside in thE ouTer RegioNs oF The diSc (seE ThEir fIg | of viscosity in the disc. This is s uppor ted by t he r esul t that the Poi s sonnoise for the low-reso lutio nm odel s t hat w e havep er f o rme dmu lti pl e t imesisless th an 5%, see ta bl e \[tb:gadge t \_ eda1\]. Fu rth ermore, theISM loadi ng ra t e isrob ust f or all our si mulations a n d does not see m to beaffected by numer i ca l noise.
Model ling o ft he I SM--- ---------- -- -----
In ord e rt o bea ble to discer n the relev a ntphysic al pr o cesses from t h e s imulations, wehave mode lled t h e ISM i n a very ideal ize d w ay:a sahom og e neo u sgas wit h no clu mp sand n o tu r b u l ence . T o acerta in extent wehav e mo d ell ed th e rea ctio nof th e disc to a d ensity gradient whe n the flo w f ir sthi ts th e disc. In th at case the si z e o ft h e d isc is determinedby t he ram pre ssure. If t h e disc w er e t o en c o unter a r e gi on of ga s with lo we r densi ty , thesu rfa ceareao f th e disc can ext end t o larger radiiw ithout beings tr i p pe d and th at would pr obab l y in crea s eits effec tivecr o ss section. Despite th elowerdensi ty, the ISM l oading rat e c ould sti ll b e o f the same orde r, be cause a lo w er ISM d ensit y allows a larger s urface a rea of th e d i s cto collect IS M . Inth at sens e,accreti onont o p rot op lanetarydiscs ma ybe a s elf -regu l ating pr oc ess .Ina fol l ow-upworkwe w il lp erf orm sim u la t i onswi th dif fer en t ISM den s iti es andvelocitie s t o inv es ti gate if and how theIS M loadingra tedepend s on these quantities.
In princi p le it w oul d beposs ible to a ddangula r m o mentum to th e dis c, su c h that i tmai nt ains or pr o l ong s its l ifet ime, if the ISM has turbu l enc e or substruc tur es o n sc ale s a t or s l igh t l y smaller thanthe disc s ca l e. However,e ven i n thatcase th e net effectis unlike ly to inc re aset h e o verall lif etime, s ince thed isc w o ul d qui ckl y enco un ter a di fferen t pa rt of the I SM where an a dv erse, di sruptive configurationof sub struc tur e might e xis t .
Compariso n to other wor k { #se c:dis _m0 9 _eda1 }
-- - -- --- - ----- ---- - -----
As im ila r si mulation ha s b een perf orm e d by M 09,but for a higherd isc mass and l ower v elo cit y ofth e ISM with res pec tt o the dis c. In their w ork both t h e Bon di-Hoy le rad ius and t he radius bey ond which ra m p re s sure wo ul dd ominat e ar emuch l argert hant h e radius of thedisc. T hey u s eda hig he r resol u tion for the d isc, initia lly $\ sim2.5 \ times 1 0^ 5$ par tic le s, which h a d 8 times themass of t heir IS M part icle s . Howe ver, i n t heir work t he IS Ma lso doe s not r esid e in theo uter reg ion s of the d isc ( see th e ir F ig | of_viscosity in_the disc. This is_supported by_the_result that_the_Poisson noise for_the low-resolution models_that we have performed_multiple times is_less_than 5%, see table \[tb:gadget\_eda1\]. Furthermore, the ISM loading rate is robust for all_our_simulations and_does_not_seem to be affected by_numerical noise.
Modelling of the ISM
--------------------
In_order to_be able to discern the relevant physical processes_from_the simulations, we_have modelled the ISM in a very idealized way:_as a homogeneous gas with no_clumps and no_turbulence._To_a certain extent we_have modelled the reaction of the_disc to a density gradient when_the flow first hits the disc. In_that case the size of the_disc is determined by the_ram pressure._If the disc were to_encounter a region_of gas_with lower density,_the surface area of the disc_can extend to_larger radii without being stripped and_that_would probably increase_its_effective_cross section._Despite the lower_density,_the ISM_loading_rate could still be of the_same_order, because a lower ISM density allows_a larger surface area_of_the disc to collect_ISM. In that sense, accretion_onto protoplanetary discs may be a_self-regulating process._In a_follow-up work we will perform simulations with different ISM densities and_velocities to investigate if and how_the ISM loading rate_depends on_these_quantities.
In principle it_would_be possible_to add angular momentum to the disc,_such that_it maintains or prolongs its lifetime,_if the ISM has_turbulence_or substructures on scales at or_slightly smaller than the disc scale._However, even in that case_the_net_effect is unlikely to increase_the overall lifetime, since the disc_would quickly encounter_a different part of the ISM where_an_adverse, disruptive configuration of substructure might_exist.
Comparison_to other work {#sec:dis_m09_eda1}
------------------------
A similar simulation_has_been_performed by M09, but for_a higher disc mass and lower_velocity of the ISM with respect to the disc._In their work_both the Bondi-Hoyle radius and_the_radius_beyond which ram pressure would dominate are much larger than_the radius_of the disc._They used a higher resolution for the disc, initially $\sim_2.5 \times 10^5$ particles, which had 8_times the mass of their ISM particles. However, in their work_the ISM also does not reside in the_outer regions of the disc (see their_Fig |
$, let $B(x, t, r)$ be the open metric ball with center $x$ and of radius $r$ w.r.t. the metric $g(t)$, and for $T' \in (0, T]$, let $PB_r(x,T')$ denote a parabolic ball (as called in \[H13\]) that is the set of all points $(x', t)$ with $ x' \in \overline{B(x,t,r)}$ (the closure of $B(x,t,r)$) and $t \in [0, T']$. For any points $x, y \in M$ let $d_t(x, y)$ be the distance between $x$ and $y$ w.r.t. $g(t)$.
We have the following gradient estimate.
\[thm 1.1\] Fix $T >0$. Let $(M, (g(t))_{t\in [0,T]})$ be a solution (not necessarily complete) to the Ricci flow on a manifold $M$ (without boundary) of dimension $n$. Fix $x_0 \in M$ and $r>0$. Assume that for any $t\in [0,T]$ the closure of the open metric ball $B(x_0, t, r)$ is compact, and $Ric\leq \frac{n-1}{r^2}$ on the parabolic ball $PB_r(x_0, T)$. Let $u$ be a smooth solution to the heat equation $(\partial_t-\Delta_{g(t)})u=0$ coupled to the Ricci flow on $M \times [0,T]$. Suppose $|u|\leq a$ on $PB_r(x_0, T)$, where $a$ is a positive constant. Then $$|\nabla u| \leq C_1a(\frac{1}{r}+\frac{1}{\sqrt{t}}) \hspace{2mm} \text{on} \hspace{2mm} PB_{ \frac{r}{2}}(x_0, T)\setminus \{(x, 0) | x \in M\},$$ where the constant $C_1$ depends only on the dimension.
Compare Theorem 2.2 in Bailesteanu-Cao-Pulemotov \[BCP\], where a two-sided bound | $, let $ B(x, t, r)$ be the open metric ball with plaza $ x$ and of radius $ r$ w.r.t. the measured $ g(t)$, and for $ T' \in (0, T]$, let $ PB_r(x, T')$ announce a parabolic ball (as call in \[H13\ ]) that is the set of all points $ (x', t)$ with $ x' \in \overline{B(x, t, r)}$ (the closing of $ B(x, t, r)$) and $ t \in [ 0, T']$. For any points $ x, y \in M$ permit $ d_t(x, y)$ be the distance between $ x$ and $ y$ w.r.t. $ g(t)$.
We have the following gradient appraisal.
\[thm 1.1\ ] Fix $ T > 0$. Let $ (M, (g(t))_{t\in [ 0,T]})$ be a solution (not necessarily accomplished) to the Ricci flow on a manifold $ M$ (without boundary) of dimension $ n$. Fix $ x_0 \in M$ and $ r>0$. Assume that for any $ t\in [ 0,T]$ the closure of the open measured ball $ B(x_0, t, r)$ is compact, and $ Ric\leq \frac{n-1}{r^2}$ on the parabolic ball $ PB_r(x_0, T)$. get $ u$ be a smooth solution to the heat equality $ (\partial_t-\Delta_{g(t)})u=0 $ coupled to the Ricci stream on $ M \times [ 0,T]$. Suppose $ |u|\leq a$ on $ PB_r(x_0, T)$, where $ a$ is a positivist constant. Then $ $ |\nabla u| \leq C_1a(\frac{1}{r}+\frac{1}{\sqrt{t } }) \hspace{2 mm } \text{on } \hspace{2 mm } PB _ { \frac{r}{2}}(x_0, T)\setminus \{(x, 0) | x \in M\},$$ where the constant $ C_1 $ depends only on the property.
Compare Theorem 2.2 in Bailesteanu - Cao - Pulemotov \[BCP\ ], where a two - sided bound | $, leh $B(x, t, r)$ be the open metvic ball with ceurer $x$ end of dadius $r$ w.r.t. the metric $g(t)$, and for $T' \ib (0, T]$, oet $PB_r(x,T')$ denote a parxbolic bapl (as caoled un \[H13\]) that ma the scc of zpl pmmnts $(x', t)$ with $ w' \in \overlite{B(x,t,r)}$ (the clovufe of $B(x,t,r)$) and $t \in [0, T']$. For any points $x, y \in M$ let $d_t(x, y)$ be jhe dpseancs between $x$ and $y$ w.r.t. $g(t)$.
We have ths folloxing gradient extimate.
\[thm 1.1\] Fix $T >0$. Let $(M, (h(t))_{t\ij [0,T]})$ be a solution (jot necessatjly xomplete) to ghe Ricci flow on a mahifold $M$ (without boundary) of dioensipn $n$. Fix $x_0 \ib M$ dnd $r>0$. Assumx that for any $t\in [0,T]$ the cnosure pf the open mebric uall $B(x_0, t, r)$ is compact, anv $Ric\leq \frac{n-1}{r^2}$ on tre parabonie ball $PB_r(x_0, T)$. Let $u$ bw q smomth voluguon to tie geat eeuavion $(\partiam_t-\Delta_{g(t)})u=0$ xoupled to the Ricco soow on $M \timea [0,T]$. Su[pjse $|u|\leq a$ on $PB_r(x_0, T)$, where $a$ is a positpve donstant. Then $$|\nabla u| \lwq C_1a(\frac{1}{r}+\frac{1}{\sqrt{t}}) \hspace{2mi} \text{on} \hspace{2mm} PB_{ \frac{r}{2}}(x_0, T)\setminus \{(x, 0) | x \in M\},$$ dhexc gye constant $C_1$ depends only on the dimension.
Compwde Tmeorem 2.2 in Bailefteanu-Cao-Pukeloyjv \[BCP\], where x two-snses bound | $, let $B(x, t, r)$ be the ball center $x$ of radius $r$ for \in (0, T]$, $PB_r(x,T')$ denote a ball (as called in \[H13\]) that the set of all points $(x', t)$ with $ x' \in \overline{B(x,t,r)}$ (the of $B(x,t,r)$) and $t \in [0, T']$. For any points $x, y \in let y)$ the between $x$ and $y$ w.r.t. $g(t)$. We have the following gradient estimate. \[thm 1.1\] Fix $T Let $(M, (g(t))_{t\in [0,T]})$ be a solution (not complete) to the Ricci on a manifold $M$ (without of $n$. Fix \in and Assume that for $t\in [0,T]$ the closure of the open metric ball $B(x_0, t, r)$ is compact, and $Ric\leq \frac{n-1}{r^2}$ the parabolic T)$. Let be smooth to the heat coupled to the Ricci flow on Suppose $|u|\leq a$ on $PB_r(x_0, T)$, where $a$ a positive Then $$|\nabla u| \leq C_1a(\frac{1}{r}+\frac{1}{\sqrt{t}}) \hspace{2mm} \hspace{2mm} PB_{ \frac{r}{2}}(x_0, T)\setminus \{(x, 0) | x M\},$$ where the constant $C_1$ depends only on the dimension. Compare Theorem 2.2 in Bailesteanu-Cao-Pulemotov a two-sided bound | $, let $B(x, t, r)$ be the open metric balL with centeR $x$ and Of rAdiUs $R$ w.r.t. The mEtric $g(t)$, and for $T' \IN (0, T]$, leT $PB_r(x,T')$ denote a parabolic Ball (aS cALled IN \[H13\]) That iS the set OF aLL PoiNtS $(x', T)$ wiTh $ X' \In \OverlIne{b(x,t,r)}$ (the Closure of $B(X,t,r)$) AnD $t \in [0, T']$. For any pOInTs $x, y \in M$ let $D_t(x, Y)$ be the distanCe bEtween $X$ aNd $y$ W.R.t. $g(t)$.
WE haVe the FollowINg gradIent estimAtE.
\[Thm 1.1\] Fix $t >0$. let $(M, (g(t))_{t\IN [0,t]})$ bE a soLution (not necessarILy COmplete) to the RiCci floW oN A mANIfoLd $M$ (Without bouNdAry) of DImensioN $N$. FIX $X_0 \In M$ ANd $r>0$. Assume that For any $t\in [0,T]$ tHE clOsure oF tHe oPEn metrIc balL $B(X_0, T, r)$ iS compact, and $ric\lEq \frac{n-1}{r^2}$ oN the paRAbolic bALl $PB_r(x_0, T)$. let $u$ be A smOotH solUTiOn To tHe HEat EQuAtiON $(\paRtial_t-\DeLtA_{g(T)})u=0$ couPled TO THE RicCi fLow oN $M \timEs [0,T]$. Suppose $|u|\leQ a$ oN $PB_r(X_0, t)$, whEre $a$ iS a posItivE cOnstaNt. Then $$|\Nabla U| \lEq C_1a(\frac{1}{r}+\frac{1}{\sqRt{t}}) \hSpace{2mm} \teXt{oN} \hSpaCe{2Mm} PB_{ \fRAc{r}{2}}(x_0, T)\sEtmInuS \{(x, 0) | x \in M\},$$ wHere the COnsTaNT $c_1$ DePends only on the dimeNsION.
COmpare ThEorem 2.2 iN baIlESteanu-CaO-PUleMotoV \[bcP\], wheRe a tWO-sIded bounD | $, let $B(x, t, r)$ be the open metr ic ba llwit hcent er $ x$ and of radi u s $r $ w.r.t. the metric $g (t)$, a n d fo r $ T' \i n (0, T ] $, l et$P B_ r(x ,T ' )$ deno tea parab olic ball(as c alled in \[H 1 3\ ]) that is th e set of all po ints $ (x ',t )$ wi th$ x'\in \o v erline {B(x,t,r) }$ (the c l osure o f $B (x,t ,r)$) and $t \in[ 0, T']$. For anypoints $ x ,y \in M$ let $d_t( x, y)$b e the d i st a n c e b e tween $x$ and $y$ w.r.t. $g( t)$.
We ha v e thefollo wi n g g radient est imat e.
\[thm 1.1\] Fix $ T >0$. L et $(M , ( g(t ))_{ t \i n[0, T] } )$b ea s o lut ion (not n ec essar ilyc o m p lete ) t o th e Ric ci flow on aman ifol d $M $ (wi thout bou nd ary)of dim ensio n$n$. Fix $x_0 \ in M $ and $r> 0$. A ssu me that for an y $ t\i n [0,T] $ the c l osu re o f t he open metric bal l$ B (x _0, t, r )$ isc om pa c t, and $ Ri c\l eq \ f r ac{n- 1}{r ^ 2} $ on the parab o li cball $P B_ r(x_0, T )$. Le t $u$ be a smoot h soluti on to the heat equat i on $(\partial _ t- \ D el t a_{g (t) })u=0$ coup ledt o th e Ri c ci fl o w on$M \t im e s[ 0,T]$. Suppose $|u| \l eq a$on $P B_r(x_0, T)$, where $a$ i s a posit ivec on s tant. Then $$| \nabl a u| \leqC _1a(\fra c{1}{ r}+\frac {1}{\sqrt { t }}) \ hsp ace {2m m}\ te xt{on} \hsp a c e{2m m} PB_ { \ frac{r} {2} }(x _0, T) \s etminus \ {(x, 0)|x\i nM\} ,$$ w h ere theco nst an t $ C_1$d epends only onth ed ime nsion.Co m p areTh eo rem2.2 i n Bai lest e anu -Cao-Pu lemotov \ [BC P \],wh er e a two -sided bound | $, let_$B(x, t,_r)$ be the open_metric ball_with_center $x$_and_of radius $r$_w.r.t. the metric_$g(t)$, and for $T'_\in (0, T]$,_let_$PB_r(x,T')$ denote a parabolic ball (as called in \[H13\]) that is the set of_all_points $(x',_t)$_with_$ x' \in \overline{B(x,t,r)}$ (the_closure of $B(x,t,r)$) and $t_\in [0,_T']$. For any points $x, y \in M$_let_$d_t(x, y)$ be_the distance between $x$ and $y$ w.r.t. $g(t)$.
We have_the following gradient estimate.
\[thm 1.1\] _Fix $T >0$._Let_$(M,_(g(t))_{t\in [0,T]})$ be a_solution (not necessarily complete) to the_Ricci flow on a manifold $M$_(without boundary) of dimension $n$. Fix $x_0_\in M$ and $r>0$. Assume that_for any $t\in [0,T]$ the_closure of_the open metric ball $B(x_0,_t, r)$ is_compact, and_$Ric\leq \frac{n-1}{r^2}$ on_the parabolic ball $PB_r(x_0, T)$. Let_$u$ be a_smooth solution to the heat equation_$(\partial_t-\Delta_{g(t)})u=0$_coupled to the_Ricci_flow_on $M_\times [0,T]$. Suppose_$|u|\leq_a$ on_$PB_r(x_0,_T)$, where $a$ is a positive_constant._Then $$|\nabla u| \leq C_1a(\frac{1}{r}+\frac{1}{\sqrt{t}}) _ \hspace{2mm} \text{on}__ \hspace{2mm} _ PB_{ \frac{r}{2}}(x_0, T)\setminus \{(x,_0) | x \in M\},$$ where_the constant_$C_1$ depends_only on the dimension.
Compare Theorem 2.2 in Bailesteanu-Cao-Pulemotov \[BCP\], where a_two-sided bound |
B 389 (1996) 757. R. Bernabei et al., Phys. Lett. B 424 (1998) 195; R. Bernabei et al., Il Nuovo Cim. A 112 (1999) 545. R. Bernabei et al., Phys. Lett. B 450 (1999) 448. P. Belli et al., Phys. Rev. D 61 (2000) 023512. R. Bernabei et al., Phys. Lett. B 480 (2000) 23. R. Bernabei et al., Eur. Phys. J. C 18 (2000) 283. R. Bernabei et al., Phys. Lett. B 509 (2001) 197. R. Bernabei et al., Eur. Phys. J. C 23 (2002) 61. P. Belli et al., Phys. Rev. D 66 (2002) 043503. R. Bernabei el al., La Rivista del Nuovo Cimento 26 n.1 (2003) 1-73 R. Bernabei et al., Int. J. Mod. Phys. D 13 (2004) 2127. R. Bernabei et al., Int. J. Mod. Phys. A 21 (2006) 1445. R. Bernabei et al., Eur. Phys. J. C 47 (2006) 263.
R. Bernabei et al., Phys. Lett. B408 (1997) 439. P. Belli et al., Phys. Lett. B460 (1999) 236. R. Bernabei et al., Phys. Rev. Lett. 83 (1999) 4918. P. Belli et al., Phys. Rev. C60 (1999) 065501. R. Bernabei et al., Il Nuovo Cimento A112 (1999) 1541. R. Bernabei et al., Phys. Lett. B 515 (2001) 6. F. Cappella et al., Eur. Phys. J.-direct C14 (2002) 1. R. Bernabei et al., Eur. Phys. J. A 23 (2005) 7. R. Bernabei et al., Eur. Phys. J. A 24 (2005) 51.
P. Belli et al., Il Nuovo Cim. 103A (1990) 767. P. Belli et al., Astrop. Phys. 5 (1996) 217. P. | B 389 (1996) 757. R. Bernabei et al. , Phys. Lett. B 424 (1998) 195; R. Bernabei et al. , Il Nuovo Cim. A 112 (1999) 545. R. Bernabei et al. , Phys. Lett. B 450 (1999) 448. P. Belli et al. , Phys. Rev. D 61 (2000) 023512. R. Bernabei et al. , Phys. Lett. boron 480 (2000) 23. R. Bernabei et al. , Eur. Phys. J. C 18 (2000) 283. R. Bernabei et al. , Phys. Lett. boron 509 (2001) 197. R. Bernabei et al. , Eur. Phys. J. C 23 (2002) 61. P. Belli et al. , Phys. Rev. D 66 (2002) 043503. R. Bernabei el al. , La Rivista del Nuovo Cimento 26 n.1 (2003) 1 - 73 R. Bernabei et al. , Int. J. Mod. Phys. D 13 (2004) 2127. R. Bernabei et al. , Int. J. Mod. Phys. A 21 (2006) 1445. R. Bernabei et al. , Eur. Phys. J. C 47 (2006) 263.
R. Bernabei et al. , Phys. Lett. B408 (1997) 439. P. Belli et al. , Phys. Lett. B460 (1999) 236. R. Bernabei et al. , Phys. Rev. Lett. 83 (1999) 4918. P. Belli et al. , Phys. Rev. C60 (1999) 065501. R. Bernabei et al. , Il Nuovo Cimento A112 (1999) 1541. R. Bernabei et al. , Phys. Lett. B 515 (2001) 6. F. Cappella et al. , Eur. Phys. J.-direct C14 (2002) 1. R. Bernabei et al. , Eur. Phys. J. A 23 (2005) 7. R. Bernabei et al. , Eur. Phys. J. A 24 (2005) 51.
P. Belli et al. , Il Nuovo Cim. 103A (1990) 767. P. Belli et al. , Astrop. Phys. 5 (1996) 217. P. | B 389 (1996) 757. R. Bernabei et al., Phys. Lett. B 424 (1998) 195; R. Betnqbei ev al., Il Nuovo Ckm. A 112 (1999) 545. R. Bernabei et al., Phyd. Oett. V 450 (1999) 448. P. Belli et al., Phyr. Rev. D 61 (2000) 023512. R. Bernqbei wt al., Phys. Lett. B 480 (2000) 23. R. Bsvnaben xt al., Eur. Phys. M. C 18 (2000) 283. R. Besnabei et al., Pvyr. Pett. B 509 (2001) 197. R. Bernabei et al., Eur. Phys. J. C 23 (2002) 61. P. Belli et al., Prys. Ged. D 66 (2002) 043503. R. Bernabei el al., La Rivista del Nuovo Bimento 26 n.1 (2003) 1-73 R. Betnabei et al., Int. J. Mod. Phyd. D 13 (2004) 2127. R. Bernabei et ap., Int. J. Mod. Phyf. A 21 (2006) 1445. R. Bernxbei et al., Eur. Phys. J. D 47 (2006) 263.
R. Bernabei et al., Phys. Lett. C408 (1997) 439. P. Belli et ql., Phjv. Lett. B460 (1999) 236. C. Bernwbei et al., Pmjs. Rev. Nett. 83 (1999) 4918. P. Belli et al., Phbs. Rwv. C60 (1999) 065501. R. Bernabei et el., Il Nuovo Cimento W112 (1999) 1541. R. Bertayei et al., Phys. Lett. B 515 (2001) 6. F. Cdppenla dr au., Ehr. Pgys. J.-dlrert C14 (2002) 1. R. Bednabei et ao., Eur. Phys. J. A 23 (2005) 7. R. Bqgmabei et al., Sur. Phrs. J. A 24 (2005) 51.
P. Belli et al., Il Nuovo Cim. 103A (1990) 767. P. Btlli st al., Astrop. Phys. 5 (1996) 217. P. | B 389 (1996) 757. R. Bernabei et Lett. 424 (1998) R. Bernabei et 112 545. R. Bernabei al., Phys. Lett. 450 (1999) 448. P. Belli et Phys. Rev. D 61 (2000) 023512. R. Bernabei et al., Phys. Lett. B (2000) 23. R. Bernabei et al., Eur. Phys. J. C 18 (2000) 283. Bernabei al., Lett. 509 (2001) 197. R. Bernabei et al., Eur. Phys. J. C 23 (2002) 61. P. Belli al., Phys. Rev. D 66 (2002) 043503. R. el al., La Rivista Nuovo Cimento 26 n.1 (2003) R. et al., J. Phys. 13 (2004) 2127. Bernabei et al., Int. J. Mod. Phys. A 21 (2006) 1445. R. Bernabei et al., Eur. Phys. C 47 R. Bernabei al., Lett. (1997) 439. P. al., Phys. Lett. B460 (1999) 236. al., Phys. Rev. Lett. 83 (1999) 4918. P. et al., Rev. C60 (1999) 065501. R. Bernabei al., Il Nuovo Cimento A112 (1999) 1541. R. et al., Phys. Lett. B 515 (2001) 6. F. Cappella et al., Eur. Phys. J.-direct 1. R. Bernabei et Eur. Phys. J. 23 7. Bernabei al., Eur. J. A 24 (2005) 51. P. Belli et al., Il Nuovo 103A (1990) 767. P. Belli et al., Astrop. Phys. 5 P. | B 389 (1996) 757. R. Bernabei et al., Phys. Lett. B 424 (1998) 195; R. BErnabei et aL., Il NuOvo cim. a 112 (1999) 545. R. bernAbei Et al., Phys. Lett. B 450 (1999) 448. P. bElli Et al., Phys. Rev. D 61 (2000) 023512. R. Bernabei eT al., PhYs. lEtt. B 480 (2000) 23. r. beRnabeI et al., EuR. phYS. j. C 18 (2000) 283. R. beRnAbeI eT Al., phys. LEtt. b 509 (2001) 197. R. BernaBei et al., Eur. phyS. J. c 23 (2002) 61. P. Belli et al., PHYs. rev. D 66 (2002) 043503. R. BernaBei El al., La RivistA deL Nuovo ciMenTO 26 n.1 (2003) 1-73 R. BeRnaBei et Al., Int. J. mOd. Phys. d 13 (2004) 2127. R. BernabeI eT Al., Int. J. mOd. Phys. A 21 (2006) 1445. r. bErNabeI et al., Eur. Phys. J. C 47 (2006) 263.
R. BeRNaBEi et al., Phys. Lett. b408 (1997) 439. P. BellI eT Al., pHYs. LEtt. b460 (1999) 236. R. Bernabei Et Al., PhyS. rev. Lett. 83 (1999) 4918. p. beLLI Et aL., phys. Rev. C60 (1999) 065501. R. BernAbei et al., Il NUOvo cimentO A112 (1999) 1541. r. BeRNabei eT al., PhYs. lEtt. b 515 (2001) 6. F. Cappella eT al., EUr. Phys. J.-diRect C14 (2002) 1. R. bErnabei ET al., Eur. PHys. J. A 23 (2005) 7. R. berNabEi et AL., EUr. phyS. J. a 24 (2005) 51.
p. BeLLi Et aL., il NUovo Cim. 103A (1990) 767. p. BElLi et aL., AstROP. pHys. 5 (1996) 217. P. | B 389 (1996) 757. R. Bern abei et al ., Ph ys. Le tt . B424(1998) 195; R. Bern abei et al., Il NuovoCim.A1 12 ( 1 99 9) 54 5. R. B e rn a b eiet a l., P h ys . Let t.B 450 ( 1999) 448. P. B elli et al., Ph ys. Rev. D 61 (2000) 0235 12. R. Be rn abe i et a l., Phys . Lett . B 480 (2000) 2 3. R. Ber n abei et a l. , Eu r. Phys. J. C 18( 20 0 0) 283. R. Ber nabeiet al . , Ph ys. Lett. B 5 09 (200 1 ) 197.R .B e r nab e i et al., Eur . Phys. J.C 23 (2002 )61. P. Bel li et a l .,Phys. Rev.D 66 (2002) 0 43503. R. Bern a bei elal., L a R ivi stad el N uov oC ime n to 26 n.1 (2003)1- 73 R. B erna b e i et a l., Int . J.Mod. Phys. D13(200 4 ) 2 127.R. Be rnab ei et a l., In t. J. M od. Phys. A 21(200 6) 1445.R.Be rna be i eta l., Eu r.Phy s. J. C 47 (20 0 6)26 3 .
R . Bernabei et al., P h y s. Lett. B 408 (1 9 97 )4 39. P. B el liet a l . , Phy s. L e tt . B460 ( 1999)2 36 .R. Bern ab ei etal .,Phy s. Re v . Le tt. 83 (1999)4918. P. Belli et al . , Phys. Rev.C 60 ( 19 9 9) 0 655 01. R. Bern abei et a l.,I lNuo v o Cim entoA1 1 2( 1999) 1541. R. Bern ab ei etal.,Phys. Lett. B 515 (2001 ) 6 . F. Cap pell a e t al., Eur. Phy s. J. -direct C1 4 (2002)1. R. Bernabe i et al., E ur. Phys . J . A 23 (2 0 0 5) 7. R. Bernab e i etal ., Eur. Ph ys. J.A 2 4 ( 200 5)51 .
P. Bel li et al ., I lNu ovo Cim. 103A (19 90 ) 7 67 . P . Bel l i et a l., A stro p. P h ys. 5 (199 6 )2 1 7. P . | B_389 (1996)_757. R. Bernabei et_al., Phys._Lett._B 424_(1998)_195; R. Bernabei_et al., Il_Nuovo Cim. A 112_(1999) 545. R._Bernabei_et al., Phys. Lett. B 450 (1999) 448. P. Belli et al., Phys. Rev._D_61 (2000)_023512._R._Bernabei et al., Phys. Lett._B 480 (2000) 23. R._Bernabei et_al., Eur. Phys. J. C 18 (2000) 283._R._Bernabei et al.,_Phys. Lett. B 509 (2001) 197. R. Bernabei et_al., Eur. Phys. J. C 23_(2002) 61. P._Belli_et_al., Phys. Rev. D_66 (2002) 043503. R. Bernabei el_al., La Rivista del Nuovo Cimento_26 n.1 (2003) 1-73 R. Bernabei et_al., Int. J. Mod. Phys. D_13 (2004) 2127. R. Bernabei_et al.,_Int. J. Mod. Phys. A_21 (2006) 1445._R. Bernabei_et al., Eur._Phys. J. C 47 (2006) 263.
R._Bernabei et al.,_Phys. Lett. B408 (1997) 439. P._Belli_et al., Phys._Lett._B460_(1999) 236._R. Bernabei et_al.,_Phys. Rev._Lett._83 (1999) 4918. P. Belli et_al.,_Phys. Rev. C60 (1999) 065501. R. Bernabei_et al., Il Nuovo_Cimento_A112 (1999) 1541. R._Bernabei et al., Phys. Lett._B 515 (2001) 6. F. Cappella_et al.,_Eur. Phys._J.-direct C14 (2002) 1. R. Bernabei et al., Eur. Phys. J._A 23 (2005) 7. R. Bernabei_et al., Eur. Phys._J. A_24_(2005) 51.
P. Belli et_al.,_Il Nuovo_Cim. 103A (1990) 767. P. Belli et al.,_Astrop. Phys._5 (1996) 217. P. |
and not straightforward to check. Moreover no information on the size of the constant $\alpha$, and hence the exponential rate of decay to the steady state, is provided. Our aim in this section is to address these issues. Working within the framework of more general weighted $\ell^1$ spaces, we study the long-term dynamics of solutions in both the mass-conserving and mass-loss cases. When mass is conserved, we establish simpler conditions under which the fragmentation semigroup $(S^{(w)}(t))_{t \geq 0}$ satisfies an inequality of the form on some space $\ell^1_w$, and also quantify $\alpha$.
We begin by considering the general fragmentation system, where the coefficients $a_n$ and $b_{n,j}$ satisfy Assumption \[A1.1\], and recall that $G^{(w)}=\overline{A^{(w)}+B^{(w)}}$ is the generator of a substochastic $C_0$-semigroup, $(S^{(w)}(t))_{t \ge 0}$, on $\ell_w^1$ whenever Assumption \[assumption on weight for generation\] holds. Furthermore, $(S^{(w)}(t))_{t \ge 0}$ is analytic, with generator $A^{(w)}+B^{(w)}$ when the more restrictive Assumption \[assumption for analyticity\] is satisfied.
\[mass loss decay of solution\] Let Assumptions \[A1.1\] and \[assumption on weight for generation\] hold.
Then $$\label{decay of solution to zero}
\lim\limits_{t \to \infty} \Vert S^{(w)}(t)\mathring{u} \Vert_w = 0$$ for all $\mathring{u} \in \ell_w^1$ if and only if $a_n > 0$ for all $n \in \mathbb{N}$.
If, additionally, we choose $w$ such that Assumption \[assumption for analyticity\] is satisfied, and set $a_0 \coloneqq \inf_{n \in \mathbb{N}} a_n$, then $$\label{semigroup exponential bound}
\Vert S^{(w)}(t) \Vert \le e^{-(1-\kappa)a_0t},$$ and hence, if $a_0>0$ and $\alpha \in [0,(1-\kappa)a_0)$, we have | and not straightforward to check. Moreover no information on the size of the changeless $ \alpha$, and therefore the exponential rate of decay to the steady department of state, is provide. Our aim in this section is to cover these issues. Working within the framework of more cosmopolitan burden $ \ell^1 $ spaces, we learn the long - term moral force of solutions in both the mass - conserving and mass - loss cases. When mass is conserve, we establish simpler conditions under which the atomization semigroup $ (S^{(w)}(t))_{t \geq 0}$ satisfies an inequality of the form on some outer space $ \ell^1_w$, and also quantify $ \alpha$.
We get down by considering the cosmopolitan fragmentation system, where the coefficients $ a_n$ and $ b_{n, j}$ satisfy Assumption \[A1.1\ ], and hark back that $ G^{(w)}=\overline{A^{(w)}+B^{(w)}}$ is the generator of a substochastic $ C_0$-semigroup, $ (S^{(w)}(t))_{t \ge 0}$, on $ \ell_w^1 $ whenever Assumption \[assumption on weight for generation\ ] holds. Furthermore, $ (S^{(w)}(t))_{t \ge 0}$ is analytic, with generator $ A^{(w)}+B^{(w)}$ when the more restrictive Assumption \[assumption for analyticity\ ] is satisfied.
\[mass loss decay of solution\ ] lease Assumptions \[A1.1\ ] and \[assumption on weight for generation\ ] hold.
Then $ $ \label{decay of solution to zero }
\lim\limits_{t \to \infty } \Vert S^{(w)}(t)\mathring{u } \Vert_w = 0$$ for all $ \mathring{u } \in \ell_w^1 $ if and only if $ a_n > 0 $ for all $ n \in \mathbb{N}$.
If, additionally, we choose $ w$ such that Assumption \[assumption for analyticity\ ] is satisfied, and set $ a_0 \coloneqq \inf_{n \in \mathbb{N } } a_n$, then $ $ \label{semigroup exponential bound }
\Vert S^{(w)}(t) \Vert \le e^{-(1-\kappa)a_0t},$$ and hence, if $ a_0>0 $ and $ \alpha \in [ 0,(1-\kappa)a_0)$, we have | anf not straightforward to check. Moreover no infmrmatikn on thd size of the constant $\alpha$, abd hebce the exponential rage of decwy to thw sttady state, is protjded. Ouv aim ln thns section is to address tvese issues. Woskkny within the framework of more generwl weigntfd $\ell^1$ spaces, re suudr ths long-term dynamics of solutions ih both uhe mass-conserving and mass-loss cases. When mwss ls conserved, we eshablish simkmer xonditions uvder which the fragmenjation semigroup $(S^{(w)}(t))_{t \geq 0}$ satiswies cn inequalijv of jhe form on wome fpace $\ell^1_w$, akc also quantigy $\alpha$.
We begik by ronsudering the general fcagmentation system, rhere the ckefficients $a_n$ ane $v_{n,j}$ sdtisxy Arwumotikn \[E1.1\], ahd recwll that $G^{(w)}=\ovsrline{A^{(w)}+B^{(w)}}$ us the generator of a wubstochastic $C_0$-semidrjup, $(S^{(w)}(t))_{t \ge 0}$, on $\ell_w^1$ whenever Assumptiot \[asaumption on weight for teneration\] holds. Furtjermore, $(S^{(r)}(t))_{t \ge 0}$ is analytic, with generator $A^{(w)}+B^{(w)}$ when the kore certrngbive Qsdumption \[assumption for analyticity\] is satisfiqs.
\[mssx loss decay on solution\] Let Assimotojns \[A1.1\] and \[assuoption on weight for generahion\] hojd.
Theb $$\label{desay pf solution to zero}
\lim\limuts_{t \to \inftj} \Veet S^{(w)}(t)\mathring{u} \Vext_w = 0$$ for alu $\majhring{i} \in \ell_w^1$ if and only iy $a_n > 0$ for all $n \ln \mathbb{H}$.
Kf, additionally, de bhoove $w$ such that Assumption \[afsumption for analytkciti\] is saeisfied, anf set $a_0 \coloneqq \inf_{n \in \mathyb{N}} a_t$, then $$\labfl{semigroup exponential bound}
\Txrt S^{(w)}(t) \Vert \ke e^{-(1-\kdppa)a_0t},$$ aud henge, if $a_0>0$ and $\alpra \in [0,(1-\kappa)a_0)$, wg have | and not straightforward to check. Moreover no the of the $\alpha$, and hence to steady state, is Our aim in section is to address these issues. within the framework of more general weighted $\ell^1$ spaces, we study the long-term of solutions in both the mass-conserving and mass-loss cases. When mass is conserved, establish conditions which fragmentation semigroup $(S^{(w)}(t))_{t \geq 0}$ satisfies an inequality of the form on some space $\ell^1_w$, and quantify $\alpha$. We begin by considering the general system, where the coefficients and $b_{n,j}$ satisfy Assumption \[A1.1\], recall $G^{(w)}=\overline{A^{(w)}+B^{(w)}}$ is generator a $C_0$-semigroup, $(S^{(w)}(t))_{t \ge on $\ell_w^1$ whenever Assumption \[assumption on weight for generation\] holds. Furthermore, $(S^{(w)}(t))_{t \ge 0}$ is analytic, with $A^{(w)}+B^{(w)}$ when restrictive Assumption for is \[mass loss decay Let Assumptions \[A1.1\] and \[assumption on hold. Then $$\label{decay of solution to zero} \lim\limits_{t \infty} \Vert \Vert_w = 0$$ for all $\mathring{u} \ell_w^1$ if and only if $a_n > 0$ all $n \in \mathbb{N}$. If, additionally, we choose $w$ such that Assumption \[assumption for analyticity\] and set $a_0 \coloneqq \in \mathbb{N}} a_n$, $$\label{semigroup bound} S^{(w)}(t) \le e^{-(1-\kappa)a_0t},$$ hence, if $a_0>0$ and $\alpha \in [0,(1-\kappa)a_0)$, we have | and not straightforward to chEck. MoreoveR no inForMatIoN on tHe siZe of the constanT $\AlphA$, and hence the exponentiaL rate Of DEcay TO tHe steAdy statE, Is PROviDeD. OUr aIm IN tHis seCtiOn is to aDdress thesE isSuEs. Working witHIn The framewoRk oF more general WeiGhted $\eLl^1$ SpaCEs, we sTudY the lOng-terM DynamiCs of solutIoNS in botH The mass-CONsErviNg and mass-loss caseS. whEN mass is conservEd, we esTaBLiSH SimPleR conditionS uNder wHIch the fRAgMENTatIOn semigroup $(S^{(w)}(T))_{t \geq 0}$ satisfIEs aN inequAlIty OF the foRm on sOmE SpaCe $\ell^1_w$, and alSo quAntify $\alpHa$.
We beGIn by conSIdering The genEraL frAgmeNTaTiOn sYsTEm, wHErE thE CoeFficientS $a_N$ aNd $b_{n,j}$ SatiSFY aSsumPtiOn \[A1.1\], aNd recAll that $G^{(w)}=\overLinE{A^{(w)}+B^{(W)}}$ Is tHe genEratoR of a SuBstocHastic $c_0$-semiGrOup, $(S^{(w)}(t))_{t \ge 0}$, on $\ell_w^1$ WhenEver AssumPtiOn \[AssUmPtion ON weighT foR geNeratioN\] holds. FURthErMORE, $(S^{(W)}(t))_{t \ge 0}$ is analytic, witH gENErAtor $A^{(w)}+B^{(w)}$ When thE MoRe REstrictiVe assUmptION \[assuMptiON fOr analytIcity\] iS SaTiSfied.
\[maSs Loss deCaY of SolUtion\] lEt AsSumptiOns \[A1.1\] and \[aSsumpTIon on weight for GEneration\] hold.
tHeN $$\LAbEL{decAy oF solution to Zero}
\LIm\liMits_{T \To \InfTY} \Vert s^{(w)}(t)\maThRInG{U} \Vert_w = 0$$ for all $\mathrinG{u} \In \ell_w^1$ If and Only if $a_n > 0$ for alL $n \in \mathbb{n}$.
iF, AdditionAlly, WE cHOose $w$ such that ASsumpTion \[assumpTIon for anAlytiCity\] is saTisfied, anD SEt $a_0 \colonEqq \Inf_{N \in \MatHBB{N}} A_n$, then $$\label{seMIGrouP eXponentIal Bound}
\VeRt S^{(W)}(t) \VErt \Le e^{-(1-\KaPpa)a_0t},$$ and hEnce, if $a_0>0$ aNd $\AlPhA \iN [0,(1-\kaPpa)a_0)$, wE Have | and not straightforward t o check. M oreov ernoin form atio n on the sizeo f th e constant $\alpha$, a nd he nc e the ex ponen tial ra t eo f de ca ytoth e s teady st ate, is provided. Ou raim in thiss ec tion is to ad dress theseiss ues. W or kin g with inthe f ramewo r k of m ore gener al weight e d $\ell ^ 1 $spac es, we study thel on g -term dynamics of so lu t io n s in bo th the mas s- conse r ving an d m a s s -lo s s cases. When mass is co n ser ved, w eest a blishsimpl er con ditions und er w hich thefragme n tations emigrou p $(S^ {(w )}( t))_ { t\g eq0} $ sa t is fie s an inequal it yof th e fo r m o n so mespac e $\e ll^1_w$, andals o qu a nti fy $\ alpha $.
We begi n by c onsid er ing the general fra gmentatio n s ys tem ,where the co eff ici ents $a _n$ and $b_ {n , j } $satisfy Assumption \ [ A 1. 1\], and recal l t ha t $G^{(w) }= \ov erli n e {A^{( w)}+ B ^{ (w)}}$ i s theg en er ator of a subst oc has tic $C_0 $ -sem igroup , $(S^{( w)}(t ) )_{t \ge 0}$,o n $\ell_w^1$w he n e ve r Ass ump tion \[assu mpti o n on wei g ht fo r gene ratio n\ ] h o lds. Furthermore, $ (S ^{(w)} (t))_ {t \ge 0}$ is analytic, w i th gener ator $A ^ {(w)}+B^{(w)}$ when the morer estricti ve As sumption \[assump t i on for a nal yti cit y\] i ssatisfied.
\ [ m asslo ss deca y o f solut ion \]Let As su mptions \ [A1.1\]an d\[ as sum ption on weigh tfor g ene ratio n \] hol d.
T hen$$ \l a bel {decayo fs o luti on t o ze ro}
\lim \lim i ts_ {t \to\infty} \ Ver t S^{ (w )} (t)\mat hring{u} \Ver t_ w = 0$$ fo rall $\mat h r ing{u} \ in \ell_w^1$ if and onl y if $a_ n > 0$ f or a ll $n \in \m athbb{ N}$ .
If,additi onall y, we c hoose $ w$ su ch that Assu m p tio n \[a ss umpt ion for analyticity\] iss ati sfied, and se t $ a_0\ c ol one q q\ inf _{ n \i n \mathbb{N}} a_n $, then $$ \l a be l{semigrou p ex po nential bound}
\V e rt S^{( w)}(t) \V ert \le e ^{ -(1- \ k app a)a_0t},$$ and hen ce, if $a _ 0>0$a nd $\al pha \in [ 0, (1- \kapp a)a_0) $ , w e hav e | and_not straightforward_to check. Moreover no_information on_the_size of_the_constant $\alpha$, and_hence the exponential_rate of decay to_the steady state,_is_provided. Our aim in this section is to address these issues. Working within the_framework_of more_general_weighted_$\ell^1$ spaces, we study the_long-term dynamics of solutions in_both the_mass-conserving and mass-loss cases. When mass is conserved,_we_establish simpler conditions_under which the fragmentation semigroup $(S^{(w)}(t))_{t \geq 0}$ satisfies_an inequality of the form on_some space $\ell^1_w$,_and_also_quantify $\alpha$.
We begin by considering_the general fragmentation system, where the_coefficients $a_n$ and $b_{n,j}$ satisfy Assumption \[A1.1\],_and recall that $G^{(w)}=\overline{A^{(w)}+B^{(w)}}$ is the generator_of a substochastic $C_0$-semigroup, $(S^{(w)}(t))_{t \ge_0}$, on $\ell_w^1$ whenever Assumption \[assumption_on weight_for generation\] holds. Furthermore, $(S^{(w)}(t))_{t_\ge 0}$ is_analytic, with_generator $A^{(w)}+B^{(w)}$ when_the more restrictive Assumption \[assumption for analyticity\]_is satisfied.
\[mass loss_decay of solution\] Let Assumptions \[A1.1\] and_\[assumption_on weight for_generation\]_hold.
Then_$$\label{decay of_solution to zero}
__\lim\limits_{t \to_\infty}_\Vert S^{(w)}(t)\mathring{u} \Vert_w = 0$$ for_all_$\mathring{u} \in \ell_w^1$ if and only if_$a_n > 0$ for_all_$n \in \mathbb{N}$.
If, additionally,_we choose $w$ such that_Assumption \[assumption for analyticity\] is satisfied, and_set $a_0_\coloneqq \inf_{n_\in \mathbb{N}} a_n$, then $$\label{semigroup exponential bound}
\Vert S^{(w)}(t) \Vert_\le e^{-(1-\kappa)a_0t},$$ and hence, if $a_0>0$_and $\alpha \in [0,(1-\kappa)a_0)$,_we have |
frac{9(14-R^2)\log H}{100H^3R},
\label{eq:a1}$$ $$A_2=-\frac{9(4-R^2)\log H}{50H^4R}.
\label{eq:a2}$$
Substituting (\[eq:a2\]) into (\[eq:p1\]) estimates the asymptotic order of the perturbed pressure $$\varepsilon^{-3/2}\delta
\hat{P}^{(f)}=
\frac{12\varepsilon^{-3}{\rm Ca}
\log(R^2/2)}{25R^6}+O(R^{-6})
\ \ \ {\rm for}\ \ \ R\gg 1,$$ which is $O({\rm Ca}\log\varepsilon\ (\varepsilon^{1/2}R)^{-6})$ in the overlapping region $R\sim \varepsilon^{-1/2}$, and thus to be matched with the pressure of $O({\rm Ca}\log\varepsilon\ r^{-6})$ in the outer region ([@one1967]). This outer pressure may contribute to the lateral force of $O({\rm Ca}\log\varepsilon)$, which is larger than that for the wide-gap case, corresponding to $O({\rm Ca})$. Nevertheless, as discussed in [@urz2007], the rapid decay of the perturbed pressure for $R\gg 1$ indicates that the contribution of the outer pressure to the lateral force is negligibly smaller than that of the inner pressure. Therefore, to evaluate the leading-order migration force, one does not have to solve the outer problem. (In fact, the order of the lateral force (\[eq:fm\_lub\]) evaluated only in the inner region is confirmed to be $O({\rm Ca}\varepsilon^{-2})$ and larger than the outer contribution $O({\rm Ca}\log\varepsilon)$.) The leading pressure $\varepsilon^{-3/2}\hat{P}^{(0)}\cos\phi$ with no deformation is locally dominant but does not contribute to the lateral force due to its azimuthal cosine dependence. The viscous stress contribution to the lateral force is $O(\varepsilon)$ smaller than the pressure in the inner region. The deformation-induced lateral force $F_M$ to cancel the migration velocity and to maintain the wall-parallel motion is expressed as $$F_M\approx
\oint_{\rm contact}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\bm x}\
\varepsilon^{-3/2}\delta | frac{9(14 - R^2)\log H}{100H^3R },
\label{eq: a1}$$ $ $ A_2=-\frac{9(4 - R^2)\log H}{50H^4R }.
\label{eq: a2}$$
Substituting (\[eq: a2\ ]) into (\[eq: p1\ ]) estimates the asymptotic order of the perturbed pressure $ $ \varepsilon^{-3/2}\delta
\hat{P}^{(f)}=
\frac{12\varepsilon^{-3}{\rm Ca }
\log(R^2/2)}{25R^6}+O(R^{-6 })
\ \ \ { \rm for}\ \ \ R\gg 1,$$ which is $ O({\rm Ca}\log\varepsilon\ (\varepsilon^{1/2}R)^{-6})$ in the overlap area $ R\sim \varepsilon^{-1/2}$, and thus to be matched with the atmospheric pressure of $ O({\rm Ca}\log\varepsilon\ r^{-6})$ in the extinct region ([ @one1967 ]). This outer press may contribute to the lateral force of $ O({\rm Ca}\log\varepsilon)$, which is bigger than that for the wide - gap character, corresponding to $ O({\rm Ca})$. Nevertheless, as hash out in [ @urz2007 ], the rapid decay of the perturb pressure for $ R\gg 1 $ indicates that the contribution of the outer pressure to the lateral force is negligibly modest than that of the inner pressure. Therefore, to measure the leading - order migration force out, one does not have to resolve the outer problem. (In fact, the order of the lateral force (\[eq: fm\_lub\ ]) evaluated merely in the inner region is confirmed to be $ O({\rm Ca}\varepsilon^{-2})$ and larger than the outer contribution $ O({\rm Ca}\log\varepsilon)$.) The leading pressure $ \varepsilon^{-3/2}\hat{P}^{(0)}\cos\phi$ with no deformation is locally dominant but does not contribute to the lateral force due to its azimuthal cosine dependence. The viscous stress contribution to the lateral force is $ O(\varepsilon)$ smaller than the pressure in the inside area. The deformation - induced lateral force $ F_M$ to delete the migration speed and to maintain the wall - parallel motion is expressed as $ $ F_M\approx
\oint_{\rm contact}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\bm x}\
\varepsilon^{-3/2}\delta | fraf{9(14-R^2)\log H}{100H^3R},
\label{eq:a1}$$ $$A_2=-\frac{9(4-R^2)\uog H}{50H^4R}.
\label{eq:a2}$$
Substitnting (\[es:a2\]) into (\[dq:p1\]) estimates the asymptotic oeder if the perturbed pressjre $$\varepdilon^{-3/2}\delra
\hau{P}^{(f)}=
\frac{12\varepsilon^{-3}{\cj Ca}
\log(V^2/2)}{25X^6}+O(R^{-6})
\ \ \ {\vm fox}\ \ \ R\gg 1,$$ which ix $O({\rm Ca}\loc\varepsilon\ (\vaseosnlon^{1/2}R)^{-6})$ in the overlapping region $R\sim \varepsolln^{-1/2}$, and thus to be kwtchsd with the pressure of $O({\rm Ca}\log\vzrepsilmn\ r^{-6})$ in the oiter region ([@one1967]). This outer predsure may contribuhe to the lqterwo force of $O({\fm Ca}\log\vagzpsilon)$, whidh is larger than that for the dide-gcp case, cortzwpojging to $O({\rm Ca})$. Nvvertheless, as discusved in [@irz2007], the rapid cecey od the perturbed pressnre for $R\gg 1$ indicatgs that tha eontribution of the oytwr prgssura to rhe lauerel rorce ls iegligibly amaller thab that of the inner pwvxsure. Therefkre, to edaluate the leading-order migration forct, one does not have to solve the outer problem. (In fact, the order of the lateral force (\[eq:fm\_lub\]) evaluated onlf in vhd iukcr rdtiln is confirmed to be $O({\rm Ca}\varepsilon^{-2})$ and lawfet nhan the outer coktribution $O({\rm Ca}\lpg\gatgpsilon)$.) The lexding 'deasure $\varepsilon^{-3/2}\hah{P}^{(0)}\cos\phy$ wity no defowmatoon is locally dominant but does not coutrubute to the latercl force due to ots aaimuthal cosine dependeuce. Ths viscous shress conffibution to the uatvral force is $O(\varepsilon)$ smaljer than vhe pxessure kn tne innqr region. Hhe dcxormation-induced lwterap xorce $F_M$ tl cancel the migration velocity and to maintaon thv wall-parcllel kotion is expwessed as $$F_M\apkrox
\oint_{\xm congact}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\bm x}\
\vare'silon^{-3/2}\delta | frac{9(14-R^2)\log H}{100H^3R}, \label{eq:a1}$$ $$A_2=-\frac{9(4-R^2)\log H}{50H^4R}. \label{eq:a2}$$ Substituting (\[eq:p1\]) the asymptotic of the perturbed \log(R^2/2)}{25R^6}+O(R^{-6}) \ \ {\rm \ \ R\gg which is $O({\rm Ca}\log\varepsilon\ (\varepsilon^{1/2}R)^{-6})$ in overlapping region $R\sim \varepsilon^{-1/2}$, and thus to be matched with the pressure of Ca}\log\varepsilon\ r^{-6})$ in the outer region ([@one1967]). This outer pressure may contribute to lateral of Ca}\log\varepsilon)$, is larger than that for the wide-gap case, corresponding to $O({\rm Ca})$. Nevertheless, as discussed in the rapid decay of the perturbed pressure for 1$ indicates that the of the outer pressure to lateral is negligibly than of inner pressure. Therefore, evaluate the leading-order migration force, one does not have to solve the outer problem. (In fact, the of the (\[eq:fm\_lub\]) evaluated in inner is confirmed to Ca}\varepsilon^{-2})$ and larger than the outer The leading pressure $\varepsilon^{-3/2}\hat{P}^{(0)}\cos\phi$ with no deformation is dominant but not contribute to the lateral force to its azimuthal cosine dependence. The viscous stress to the lateral force is $O(\varepsilon)$ smaller than the pressure in the inner region. The force $F_M$ to cancel migration velocity and maintain wall-parallel is as $$F_M\approx contact}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\bm x}\ \varepsilon^{-3/2}\delta | frac{9(14-R^2)\log H}{100H^3R},
\label{eq:a1}$$ $$A_2=-\frac{9(4-R^2)\Log H}{50H^4R}.
\labeL{eq:a2}$$
SUbsTitUtIng (\[eQ:a2\]) inTo (\[eq:p1\]) estimates THe asYmptotic order of the pertUrbed PrESsurE $$\VaRepsiLon^{-3/2}\deltA
\HaT{p}^{(F)}=
\frAc{12\VaRepSiLOn^{-3}{\Rm Ca}
\lOg(R^2/2)}{25r^6}+O(R^{-6})
\ \ \ {\rm foR}\ \ \ R\gg 1,$$ which iS $O({\rM CA}\log\varepsilON\ (\vArepsilon^{1/2}R)^{-6})$ In tHe overlappinG reGion $R\sIm \VarEPsiloN^{-1/2}$, anD thus To be maTChed wiTh the presSuRE of $O({\rm cA}\log\varEPSiLon\ r^{-6})$ In the outer region ([@oNE1967]). THIs outer pressurE may coNtRIbUTE to The Lateral forCe Of $O({\rm cA}\log\varEPsILON)$, whICh is larger thaN that for the WIde-Gap casE, cOrrESpondiNg to $O({\Rm cA})$. NeVertheless, aS disCussed in [@uRz2007], the rAPid decaY Of the peRturbeD prEssUre fOR $R\Gg 1$ IndIcATes THaT thE ConTributioN oF tHe outEr prESSURe to The LateRal foRce is negligibLy sMallER thAn thaT of thE innEr PressUre. TheReforE, tO evaluate the leaDing-Order migrAtiOn ForCe, One doES not haVe tO soLve the oUter proBLem. (in FACT, tHe order of the lateraL fORCe (\[Eq:fm\_lub\]) eValuatED oNlY In the innEr RegIon iS COnfirMed tO Be $o({\rm Ca}\varEpsiloN^{-2})$ AnD lArger thAn The outEr ConTriButioN $o({\rm CA}\log\vaRepsilon)$.) the leADing pressure $\vaREpsilon^{-3/2}\hat{P}^{(0)}\coS\PhI$ WItH No deForMation is locAlly DOminAnt bUT dOes NOt conTribuTe TO tHE lateral force due to iTs AzimutHal coSine dependencE. The viscouS STRess contRibuTIoN To the lateral foRce is $o(\varepsiloN)$ Smaller tHan thE pressurE in the innER Region. ThE deForMatIon-INDuCed lateral forCE $f_M$ to CaNcel the MigRation vEloCitY anD to MaIntain the Wall-paraLlEl MoTiOn iS exprESsed as $$F_M\ApProX
\oInt_{\Rm conTAct}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\Bm x}\
\vaRepsIlOn^{-3/2}\DEltA | frac{9(14-R^2)\log H}{100H ^3R},
\lab el{eq :a1 }$$ $ $A_2 =-\f rac{9(4-R^2)\l o g H} {50H^4R}.
\label{eq:a2 }$$
Su b stit u ti ng (\ [eq:a2\ ] )i n to(\ [e q:p 1\ ] )estim ate s the a symptoticord er of the pert u rb ed pressur e $ $\varepsilon ^{- 3/2}\d el ta\ hat{P }^{ (f)}=
\frac { 12\var epsilon^{ -3 } {\rm C a }
\log( R ^ 2/ 2)}{ 25R^6}+O(R^{-6})\ \ \ {\rm for}\ \ \ R\g g1 ,$ $ whi chis $O({\rm C a}\lo g \vareps i lo n \ (\v a repsilon^{1/2 }R)^{-6})$i n t he ove rl app i ng reg ion $ R\ s im\varepsilon ^{-1 /2}$, and thust o be ma t ched wi th the pr ess ureo f$O ({\ rm Ca} \ lo g\v a rep silon\ r ^{ -6 })$ i n th e o u terreg ion([@on e1967]). This ou terp res suremay c ontr ib ute t o thelater al force of $O({\ rm C a}\log\va rep si lon )$ , whi c h is l arg erthan th at fort hewi d e - ga p case, correspond in g to $O({\rm Ca})$ . N ev e rtheless ,asdisc u s sed i n [@ u rz 2007], t he rap i dde cay ofth e pert ur bed pr essur e for $R\gg 1$ indi cates that the contr i bution of the ou t e rp ress ure to the lat eral forc e is ne gli g iblysmall er th a n that of the inner p ressur e. Th erefore, to e valuate th e l eading-o rder mi g ration force,one d oes not ha v e to sol ve th e outerproblem.( I n fact,the or der of t he lateral forc e (\[e q: fm\_lub \]) evalua ted on lyinth e inner r egion is c on fi rm edto be $O({\rmCa }\v ar eps ilon^ { -2})$and l arge rth a n t he oute r c o n trib ut io n $O ({\ rm Ca}\ log\ v are psilon) $.) The l ead i ng p re ss ure $\v arepsilon^{-3 /2 }\hat{P}^{ (0 )}\ cos\ph i $ with no deformation is locally dominan t b ut do es n ot contri but e to t hel ateral force dueto it s azimu t h al co si ne depende n c e.The v is cous stress contribution to t h e l ateral forceis$O(\ v a re psi l on ) $ s ma l ler t han the pressur e in the i nn e rregion. Th e de fo rmation -induce d lat e ral for ce $F_M$to cancel t he m i g rat ion veloci ty and t o maintai n thew al l-par all el mot io n i s exp ressed as$$F_M \appro x
\oint _{\rm c ontact}\ !\!\!\!\!\!\!\!\!\!{\rm d}^2{ \bm x }\
\varepsi lon ^ {-3 /2}\delta | frac{9(14-R^2)\log H}{100H^3R},
\label{eq:a1}$$_$$A_2=-\frac{9(4-R^2)\log H}{50H^4R}.
\label{eq:a2}$$
Substituting_(\[eq:a2\]) into (\[eq:p1\]) estimates_the asymptotic_order_of the_perturbed_pressure $$\varepsilon^{-3/2}\delta
\hat{P}^{(f)}=
\frac{12\varepsilon^{-3}{\rm Ca}
\log(R^2/2)}{25R^6}+O(R^{-6})
\_\ \ {\rm_for}\ \ \ R\gg_1,$$ which is_$O({\rm_Ca}\log\varepsilon\ (\varepsilon^{1/2}R)^{-6})$ in the overlapping region $R\sim \varepsilon^{-1/2}$, and thus to be matched with_the_pressure of_$O({\rm_Ca}\log\varepsilon\_r^{-6})$ in the outer region_([@one1967]). This outer pressure may_contribute to_the lateral force of $O({\rm Ca}\log\varepsilon)$, which is_larger_than that for_the wide-gap case, corresponding to $O({\rm Ca})$. Nevertheless, as_discussed in [@urz2007], the rapid decay_of the perturbed_pressure_for_$R\gg 1$ indicates that_the contribution of the outer pressure_to the lateral force is negligibly_smaller than that of the inner pressure._Therefore, to evaluate the leading-order migration_force, one does not have_to solve_the outer problem. (In fact,_the order of_the lateral_force (\[eq:fm\_lub\]) evaluated_only in the inner region is_confirmed to be_$O({\rm Ca}\varepsilon^{-2})$ and larger than the_outer_contribution $O({\rm Ca}\log\varepsilon)$.)_The_leading_pressure $\varepsilon^{-3/2}\hat{P}^{(0)}\cos\phi$_with no deformation_is_locally dominant_but_does not contribute to the lateral_force_due to its azimuthal cosine dependence. The_viscous stress contribution to_the_lateral force is $O(\varepsilon)$_smaller than the pressure in_the inner region. The deformation-induced lateral_force $F_M$_to cancel_the migration velocity and to maintain the wall-parallel motion is expressed_as $$F_M\approx
\oint_{\rm contact}\!\!\!\!\!\!\!\!\!\!{\rm d}^2{\bm x}\_
\varepsilon^{-3/2}\delta |
(oriented) arcs.
It is easy to see that, by a suitable reorientation of the link, an almost-strictly-cyclic covering becomes a strictly-cyclic one. As a consequence, it follows from Remark 1 that every branched cyclic covering of a link arising from a $p$-symmetric Heegaard splitting – according to Birman-Hilden construction – is strictly-cyclic.
Now we show that, conversely, every $p$-fold branched strictly-cyclic covering of a link admits a $p$-symmetric Heegaard splitting.
\[Theorem 3\] A $p$-fold strictly-cyclic covering of $\S^3$ branched over a link $L$ of bridge number $b$ is a closed, orientable 3-manifold $M$ which admits a $p$-symmetric Heegaard splitting of genus $g=(b-1)(p-1)$. So the $p$-symmetric Heegaard genus of $M$ is $$g_p(M)\le(b-1)(p-1).$$
Let $L$ be presented by a special $2b$-plat arising from a braid $\b$, and let $(\S^3,L)=(D,A_b)\cup_{\f'}(D',A'_b)$ be the $(0,b)$-decomposition described in Remark 2. Now, all arguments of the proofs of Theorem 3 of [@BH] entirely apply and the condition of Lemma 4 of [@BH] is satisfied, since the homeomorphism $\wti\b$ associated to $\b$ fixes both the sets $\FF_1$ and $\FF_2$.
As a consequence of Theorem \[Theorem 3\] and Birman-Hilden results, there is a natural one-to-one correspondence between $p$-symmetric Heegaard splittings and $p$-fold strictly-cyclic branched coverings of links.
[5]{}
Birman, J.S.: Braids, links and mapping class groups. Ann. of Math. Studies, vol. 82, Princeton Univ. Press, Princeton, N. J., 1975.
Birman, J.S., Hilden, H.M.: Heegaard splittings of branched coverings of $\S^3$. Trans. Am. Math. Soc. [**213**]{} (1975), 315–352.
Mayberry, J.; Murasugi, K.: Torsion groups of abelian | (oriented) arcs.
It is easy to see that, by a desirable change of direction of the radio link, an almost - rigorously - cyclic covering become a strictly - cyclic one. As a consequence, it follows from Remark 1 that every bifurcate cyclic application of a link originate from a $ p$-symmetric Heegaard splitting – according to Birman - Hilden construction – is rigorously - cyclic.
Now we show that, conversely, every $ p$-fold branched strictly - cyclic cover of a link admits a $ p$-symmetric Heegaard splitting.
\[Theorem 3\ ] A $ p$-fold strictly - cyclic covering of $ \S^3 $ branched over a link $ L$ of bridge act $ b$ is a closed, orientable 3 - manifold $ M$ which admits a $ p$-symmetric Heegaard splitting of genus $ g=(b-1)(p-1)$. So the $ p$-symmetric Heegaard genus of $ M$ is $ $ g_p(M)\le(b-1)(p-1).$$
lease $ L$ be presented by a special $ 2b$-plat arising from a braid $ \b$, and permit $ (\S^3,L)=(D, A_b)\cup_{\f'}(D',A'_b)$ be the $ (0,b)$-decomposition described in Remark 2. Now, all arguments of the proofs of Theorem 3 of [ @BH ] entirely put on and the condition of Lemma 4 of [ @BH ] is satisfied, since the homeomorphism $ \wti\b$ associated to $ \b$ fixes both the sets $ \FF_1 $ and $ \FF_2$.
As a consequence of Theorem \[Theorem 3\ ] and Birman - Hilden result, there is a natural one - to - one correspondence between $ p$-symmetric Heegaard splittings and $ p$-fold strictly - cyclic branched coverings of links.
[ 5 ] { }
Birman, J.S.: Braids, links and mapping class group. Ann. of Math. Studies, vol. 82, Princeton Univ. Press, Princeton, N. J., 1975.
Birman, J.S., Hilden, H.M.: Heegaard splittings of branched cover of $ \S^3$. Trans. Am. Math. Soc. [ * * 213 * * ] { } (1975), 315–352.
Mayberry, J.; Murasugi, K.: tortuosity groups of abelian | (orlented) arcs.
It is easy to see that, by a suitablx reorisntation of the link, an almost-strictpy-xyclix covering becomes a sgrictly-cyblic one. Qs a xonsequencx, it follows fdlm Rzmerk 1 that every branched wyclic coverinc uf a link arising from a $p$-symmetric Hqegaard solitting – accotding eo Bjgmcn-Hilden construction – is stricfly-cyclpc.
Now we show thay, conversely, every $p$-fold bgancjed strictly-cyclic covering od a junk admits a $p$-symmetrib Heegaard aplitting.
\[Theorem 3\] A $p$-fold stricgly-cyelic coveribg of $\V^3$ branched iver w link $L$ of ngidge ngmber $b$ is a closed, ovienteble 3-manifold $M$ which admmts a $p$-symmetric Heedaard splhtcing of genus $g=(b-1)(p-1)$. So rhw $p$-sykmetsic Fwegxars jenhs of $L$ ia $$g_p(M)\le(b-1)(p-1).$$
Lst $L$ be prewented by a special $2b$-[oat arising fdom a frwid $\b$, and let $(\S^3,L)=(D,A_b)\cup_{\f'}(D',A'_b)$ be the $(0,b)$-decokpoaition described in Remqrk 2. Now, all argumentd of the [roofs of Theorem 3 of [@BH] entirely apply and the cmnditmov oy Lemmx 4 lf [@BH] is satisfied, since the homeomorphism $\wty\g$ ssxociated to $\b$ nixes both the setx $\VF_1$ wnd $\FF_2$.
As a covsequeude of Theorem \[Theorel 3\] and Firmab-Hilden rtsultx, there is a natural one-to-obe corresponbenxe between $p$-symmetxic Heegaard splottinbs and $p$-fold strictly-cyelic bdanched covfrings of uinks.
[5]{}
Birman, J.S.: Bfaics, links and mapping class gwoups. Ann. of Kath. Stjdiex, vol. 82, Princeton Univ. Press, Princeton, N. U., 1975.
Birlat, J.S., Hildej, H.M.: Heegaard splittings of brairhed coveringx mf $\V^3$. Trans. Cm. Matm. Soc. [**213**]{} (1975), 315–352.
Mayberrr, J.; Murasugi, K.: Torsiou grouos of abelpan | (oriented) arcs. It is easy to see a reorientation of link, an almost-strictly-cyclic As consequence, it follows Remark 1 that branched cyclic covering of a link from a $p$-symmetric Heegaard splitting – according to Birman-Hilden construction – is strictly-cyclic. we show that, conversely, every $p$-fold branched strictly-cyclic covering of a link admits $p$-symmetric splitting. 3\] $p$-fold strictly-cyclic covering of $\S^3$ branched over a link $L$ of bridge number $b$ is a orientable 3-manifold $M$ which admits a $p$-symmetric Heegaard of genus $g=(b-1)(p-1)$. So $p$-symmetric Heegaard genus of $M$ $$g_p(M)\le(b-1)(p-1).$$ $L$ be by special arising from a $\b$, and let $(\S^3,L)=(D,A_b)\cup_{\f'}(D',A'_b)$ be the $(0,b)$-decomposition described in Remark 2. Now, all arguments of the proofs Theorem 3 entirely apply the of 4 of [@BH] since the homeomorphism $\wti\b$ associated to the sets $\FF_1$ and $\FF_2$. As a consequence Theorem \[Theorem and Birman-Hilden results, there is a one-to-one correspondence between $p$-symmetric Heegaard splittings and $p$-fold branched coverings of links. [5]{} Birman, J.S.: Braids, links and mapping class groups. Ann. of vol. 82, Princeton Univ. Princeton, N. J., Birman, Hilden, Heegaard of branched of $\S^3$. Trans. Am. Math. Soc. [**213**]{} (1975), 315–352. Mayberry, J.; K.: Torsion groups of abelian | (oriented) arcs.
It is easy to see That, by a suiTable ReoRieNtAtioN of tHe link, an almost-STricTly-cyclic covering becomEs a stRiCTly-cYClIc one. as a consEQuENCe, iT fOlLowS fROm remarK 1 thAt every Branched cyCliC cOvering of a liNK aRising from A $p$-sYmmetric HeegAarD splitTiNg – aCCordiNg tO BirmAn-HildEN constRuction – is StRIctly-cYClic.
Now WE ShOw thAt, conversely, every $P$-FoLD branched stricTly-cycLiC CoVERinG of A link admitS a $P$-symmETric HeeGAaRD SPliTTing.
\[Theorem 3\] A $p$-Fold strictlY-CycLic covErIng OF $\S^3$ branChed oVeR A liNk $L$ of bridge NumbEr $b$ is a cloSed, oriENtable 3-mANifold $M$ Which aDmiTs a $P$-symMEtRiC HeEgAArd SPlIttINg oF genus $g=(b-1)(P-1)$. SO tHe $p$-syMmetRIC hEegaArd GenuS of $M$ iS $$g_p(M)\le(b-1)(p-1).$$
Let $L$ bE prEsenTEd bY a speCial $2b$-Plat ArIsing From a bRaid $\b$, AnD let $(\S^3,L)=(D,A_b)\cup_{\f'}(D',A'_B)$ be tHe $(0,b)$-decompOsiTiOn dEsCribeD In RemaRk 2. NOw, aLl argumEnts of tHE prOoFS OF THeorem 3 of [@BH] entirely ApPLY aNd the conDition OF LEmMA 4 of [@BH] is sAtIsfIed, sINCe the HomeOMoRphism $\wtI\b$ assoCIaTeD to $\b$ fixEs Both thE sEts $\fF_1$ aNd $\FF_2$.
AS A conSequenCe of TheoRem \[ThEOrem 3\] and Birman-HILden results, thERe IS A nATuraL onE-to-one correSponDEnce BetwEEn $P$-syMMetriC HeegAaRD sPLittings and $p$-fold strIcTly-cycLic brAnched coverinGs of links.
[5]{}
BIRMAn, J.S.: BraiDs, liNKs ANd mapping class GroupS. Ann. of Math. sTudies, voL. 82, PrinCeton UniV. Press, PriNCEton, N. J., 1975.
BiRmaN, J.S., hilDen, h.m.: heEgaard splittiNGS of bRaNched coVerIngs of $\S^3$. traNs. AM. MaTh. SOc. [**213**]{} (1975), 315–352.
mayberry, J.; murasugi, k.: TOrSiOn GroUps of ABelian | (oriented) arcs.
It is e asy to see that , b y a s uita blereorientationo f th e link, an almost-stri ctly- cy c licc ov ering become s a s tri ct ly -cy cl i cone.Asa conse quence, it fo ll ows from Rem a rk 1 that ev ery branched cy cli c cove ri ngo f a l ink aris ing fr o m a $p $-symmetr ic Heegaa r d split t i ng – a ccording to Birma n -H i lden construct ion –is st r i ctl y-c yclic.
No wwe sh o w that, co n v e rse l y, every $p$- fold branch e d s trictl y- cyc l ic cov ering o f alink admits a $ p$-symmet ric He e gaard s p litting .
\[T heo rem 3\] A$p $-f ol d st r ic tly - cyc lic cove ri ng of $ \S^3 $ b r anch edover a li nk $L$ of bri dge num b er$b$ i s a c lose d, orie ntable 3-ma ni fold $M$ whichadmi ts a $p$- sym me tri cHeega a rd spl itt ing of gen us $g=( b -1) (p - 1 ) $. So the $p$-symmet ri c He egaard g enus o f $ M$ is $$g_p (M )\l e(b- 1 ) (p-1) .$$Le t $L$ be prese n te dby a sp ec ial $2 b$ -pl atarisi n g fr om a b raid $\b $, an d let $(\S^3,L) = (D,A_b)\cup_{ \ f' } ( D' , A'_b )$be the $(0, b)$- d ecom posi t io n d e scrib ed in R e ma r k 2. Now, all argum en ts ofthe p roofs of Theo rem 3 of [ @ B H ] entire ly a p pl y and the condi tionof Lemma 4 of [@BH] is s atisfied , since t h e homeomo rph ism $\ wti \ b $associated to $ \b$fi xes bot h t he sets $\ FF_ 1$and $ \FF_2$.
As a con se qu en ce of Theo r em \[The or em3\ ] a nd Bi r man-Hi ldenresu lt s, the re is a na t u ralon e- to-o neco rresp onde n cebetween $p$-symm etr i c He eg aa rd spli ttings and $p $- fold stric tl y-c yclicb r anched c overings of links.
[5] { }
Birm an, J.S. : Br aids, lin ksand ma ppi n g clas s grou ps. A nn . o f Math. S tu die s, vol. 82,P r inc etonUn iv.Press,Princeton, N. J.,1 975 .
Birman, J. S., Hil d e n, H. M .: Hee ga a rds p littings of bra nched cove ri n gs of $\S^3$ . Tr an s. Am.Math. S oc. [ * *213**] {} (1975) , 315–352 .
May b e rry , J.; Mura sugi, K. : Torsion group s o f abe lia n | (oriented)_arcs.
It is_easy to see that,_by a_suitable_reorientation of_the_link, an almost-strictly-cyclic_covering becomes a_strictly-cyclic one. As a_consequence, it follows_from_Remark 1 that every branched cyclic covering of a link arising from a $p$-symmetric_Heegaard_splitting –_according_to_Birman-Hilden construction – is strictly-cyclic.
Now_we show that, conversely, every_$p$-fold branched_strictly-cyclic covering of a link admits a $p$-symmetric_Heegaard_splitting.
\[Theorem 3\] A_$p$-fold strictly-cyclic covering of $\S^3$ branched over a link_$L$ of bridge number $b$ is_a closed, orientable_3-manifold_$M$_which admits a $p$-symmetric_Heegaard splitting of genus $g=(b-1)(p-1)$. So_the $p$-symmetric Heegaard genus of $M$_is $$g_p(M)\le(b-1)(p-1).$$
Let $L$ be presented by a_special $2b$-plat arising from a braid_$\b$, and let $(\S^3,L)=(D,A_b)\cup_{\f'}(D',A'_b)$ be_the $(0,b)$-decomposition_described in Remark 2. Now,_all arguments of_the proofs_of Theorem 3_of [@BH] entirely apply and the_condition of Lemma_4 of [@BH] is satisfied, since_the_homeomorphism $\wti\b$ associated_to_$\b$_fixes both_the sets $\FF_1$_and_$\FF_2$.
As a_consequence_of Theorem \[Theorem 3\] and Birman-Hilden_results,_there is a natural one-to-one correspondence between_$p$-symmetric Heegaard splittings and_$p$-fold_strictly-cyclic branched coverings of_links.
[5]{}
Birman, J.S.: Braids, links and_mapping class groups. Ann. of Math._Studies, vol._82, Princeton_Univ. Press, Princeton, N. J., 1975.
Birman, J.S., Hilden, H.M.: Heegaard splittings_of branched coverings of $\S^3$. Trans._Am. Math. Soc. [**213**]{}_(1975), 315–352.
Mayberry,_J.;_Murasugi, K.: Torsion_groups_of abelian |
(0.6) 38.3 (0.8) 35.7 (0.8) 31.1 (0.8)
47 43.1 (0.7) 43.8 (0.9) 35.6 (1.1) 27.0 (1.1) 28.7 (1.2)
52 37.9 (1.0) 36.8 (1.3) 27.7 (1.9) 27.4 (2.2) 20.3 (2.5)
58 36.4 (0.8) 32.4 (1.0) 20.9 (1.4) 17.1 (1.8) 15.0 (2.8)
62 29.0 (0.4) 27.1 (0.5) 19.4 (0.6) 15.5 (0.7) 12.6 (0.8)
67 23.5 (0.3) 21.9 (0.3) 16.6 (0.4) 12.0 (0.4) 10.3 (0.4)
72 18.4 (0.2) 16.7 (0.2) 12.0 (0.3) 9.9 (0.3) 7.4 (0.2)
77 13.3 (0.2) 11.8 (0.2) 8.5 (0.2) 6.7 (0.2) 5.4 (0.2)
82 9.1 (0.1) 7.9 (0.1) 5.9 (0.2) 4.7 (0.2) 3.7 (0.1)
87 6.6 (0.1) 5.9 (0.1) 4.1 (0.1) 3.0 (0.1) 2.6 (0.1)
92 4.6 (0.1) 4.0 (0.1) 2.7 (0.1) 2.2 (0.1) 1.5 (0.1) | (0.6) 38.3 (0.8) 35.7 (0.8) 31.1 (0.8)
47 43.1 (0.7) 43.8 (0.9) 35.6 (1.1) 27.0 (1.1) 28.7 (1.2)
52 37.9 (1.0) 36.8 (1.3) 27.7 (1.9) 27.4 (2.2) 20.3 (2.5)
58 36.4 (0.8) 32.4 (1.0) 20.9 (1.4) 17.1 (1.8) 15.0 (2.8)
62 29.0 (0.4) 27.1 (0.5) 19.4 (0.6) 15.5 (0.7) 12.6 (0.8)
67 23.5 (0.3) 21.9 (0.3) 16.6 (0.4) 12.0 (0.4) 10.3 (0.4)
72 18.4 (0.2) 16.7 (0.2) 12.0 (0.3) 9.9 (0.3) 7.4 (0.2)
77 13.3 (0.2) 11.8 (0.2) 8.5 (0.2) 6.7 (0.2) 5.4 (0.2)
82 9.1 (0.1) 7.9 (0.1) 5.9 (0.2) 4.7 (0.2) 3.7 (0.1)
87 6.6 (0.1) 5.9 (0.1) 4.1 (0.1) 3.0 (0.1) 2.6 (0.1)
92 4.6 (0.1) 4.0 (0.1) 2.7 (0.1) 2.2 (0.1) 1.5 (0.1 ) | (0.6) 38.3 (0.8) 35.7 (0.8) 31.1 (0.8)
47 43.1 (0.7) 43.8 (0.9) 35.6 (1.1) 27.0 (1.1) 28.7 (1.2)
52 37.9 (1.0) 36.8 (1.3) 27.7 (1.9) 27.4 (2.2) 20.3 (2.5)
58 36.4 (0.8) 32.4 (1.0) 20.9 (1.4) 17.1 (1.8) 15.0 (2.8)
62 29.0 (0.4) 27.1 (0.5) 19.4 (0.6) 15.5 (0.7) 12.6 (0.8)
67 23.5 (0.3) 21.9 (0.3) 16.6 (0.4) 12.0 (0.4) 10.3 (0.4)
72 18.4 (0.2) 16.7 (0.2) 12.0 (0.3) 9.9 (0.3) 7.4 (0.2)
77 13.3 (0.2) 11.8 (0.2) 8.5 (0.2) 6.7 (0.2) 5.4 (0.2)
82 9.1 (0.1) 7.9 (0.1) 5.9 (0.2) 4.7 (0.2) 3.7 (0.1)
87 6.6 (0.1) 5.9 (0.1) 4.1 (0.1) 3.0 (0.1) 2.6 (0.1)
92 4.6 (0.1) 4.0 (0.1) 2.7 (0.1) 2.2 (0.1) 1.5 (0.1) | (0.6) 38.3 (0.8) 35.7 (0.8) 31.1 (0.8) (0.7) (0.9) 35.6 27.0 (1.1) 28.7 (1.3) (1.9) 27.4 (2.2) (2.5) 58 36.4 32.4 (1.0) 20.9 (1.4) 17.1 (1.8) (2.8) 62 29.0 (0.4) 27.1 (0.5) 19.4 (0.6) 15.5 (0.7) 12.6 (0.8) 67 (0.3) 21.9 (0.3) 16.6 (0.4) 12.0 (0.4) 10.3 (0.4) 72 18.4 (0.2) 16.7 12.0 9.9 7.4 77 13.3 (0.2) 11.8 (0.2) 8.5 (0.2) 6.7 (0.2) 5.4 (0.2) 82 9.1 (0.1) 7.9 (0.1) (0.2) 4.7 (0.2) 3.7 (0.1) 87 6.6 (0.1) (0.1) 4.1 (0.1) 3.0 2.6 (0.1) 92 4.6 (0.1) (0.1) (0.1) 2.2 1.5 | (0.6) 38.3 (0.8) 35.7 (0.8) 31.1 (0.8)
47 43.1 (0.7) 43.8 (0.9) 35.6 (1.1) 27.0 (1.1) 28.7 (1.2)
52 37.9 (1.0) 36.8 (1.3) 27.7 (1.9) 27.4 (2.2) 20.3 (2.5)
58 36.4 (0.8) 32.4 (1.0) 20.9 (1.4) 17.1 (1.8) 15.0 (2.8)
62 29.0 (0.4) 27.1 (0.5) 19.4 (0.6) 15.5 (0.7) 12.6 (0.8)
67 23.5 (0.3) 21.9 (0.3) 16.6 (0.4) 12.0 (0.4) 10.3 (0.4)
72 18.4 (0.2) 16.7 (0.2) 12.0 (0.3) 9.9 (0.3) 7.4 (0.2)
77 13.3 (0.2) 11.8 (0.2) 8.5 (0.2) 6.7 (0.2) 5.4 (0.2)
82 9.1 (0.1) 7.9 (0.1) 5.9 (0.2) 4.7 (0.2) 3.7 (0.1)
87 6.6 (0.1) 5.9 (0.1) 4.1 (0.1) 3.0 (0.1) 2.6 (0.1)
92 4.6 (0.1) 4.0 (0.1) 2.7 (0.1) 2.2 (0.1) 1.5 (0.1) | (0.6) 38.3 (0.8) 35.7 (0.8) 3 1.1 ( 0.8 )
4 7 43. 1 (0. 7) 43.8 (0.9) 35.6 (1.1 ) 27. 0 ( 1.1) 28.7( 1. 2 )
5 2 37.9 (1 .0) 36.8 (1 .3 ) 27.7 (1. 9 ) 27.4 (2. 2) 20.3 (2.5)
58 36.4(0.8) 32.4(1.0) 2 0. 9 (1.4) 17.1( 1 .8 ) 15.0 (2.8)
62 29 .0 (0. 4) 2 7 .1(0. 5) 19.4(0 .6) 15.5 (0 . 7) 12. 6 (0.8)
67 23.5(0 .3) 21.9 (0.3 ) 16 .6 (0.4) 12.0 (0.4) 10.3 ( 0 .4)
7 2 1 8.4( 0. 2) 16 . 7 ( 0 .2 ) 12. 0 (0.3) 9. 9 (0. 3) 7 . 4 (0 .2)
77 13 .3 ( 0 .2) 11 .8 (0 .2) 8.5 ( 0.2) 6.7 ( 0.2) 5.4 (0. 2)
82 9.1 (0.1) 7.9 (0.1) 5.9( 0.2 )4. 7 (0.2) 3.7 (0. 1) 8 7 6 .6 (0.1) 5 .9 ( 0 . 1) 4.1 (0 .1) 3 .0 (0. 1 ) 2.6 (0 .1 )
92 4.6 (0 .1) 4 .0 (0 . 1) 2.7 (0.1 ) 2.2 (0.1) 1 .5 (0.1 ) | (0.6)_ _38.3 (0.8) _35.7 (0.8)__ 31.1_(0.8)
_ _ _ 47_ __ 43.1 (0.7) 43.8 (0.9) 35.6_(1.1)_ _27.0_(1.1)_ 28.7 (1.2)
_ _ _ 52 __ 37.9_(1.0) 36.8 (1.3) 27.7 (1.9)_ 27.4 (2.2) _20.3 (2.5)
___ _ 58 _ _36.4 (0.8) 32.4 (1.0) _ 20.9 (1.4) 17.1_(1.8) 15.0 (2.8)
_ _ _ 62_ _ _ 29.0 (0.4)_ 27.1_(0.5) 19.4 (0.6) __15.5 (0.7) __12.6_(0.8)
_ __ __ 67 __ 23.5 (0.3) _ 21.9 (0.3) __16.6 (0.4) _12.0 (0.4) 10.3_(0.4)
_ _ _72 18.4_(0.2) 16.7 (0.2) _ 12.0 (0.3) _ 9.9_(0.3)_ _7.4_(0.2)
_ _ 77_ _ 13.3_(0.2)_ 11.8 (0.2) _8.5 (0.2) 6.7_(0.2) 5.4_(0.2)
__ _ 82 _ _ 9.1 (0.1) __ 7.9 (0.1) _5.9_(0.2) 4.7 (0.2)___ 3.7 (0.1)
_ _ 87 _ 6.6_(0.1) 5.9_(0.1)__ 4.1 (0.1) 3.0 (0.1) _ _2.6 (0.1)
_ 92 _ _ 4.6 (0.1) 4.0 (0.1) _2.7 (0.1) 2.2 (0.1) _ 1.5 (0.1) |
*, 50:0 29–61, 2015.
Rama Cont, Amal Moussa, and Edson Santos. Network structure and systemic risk in banking systems. In Jean-Pierre Fouque and Joseph A Langsam, editors, *Handbook of Systemic Risk*, pages 327–368. Cambridge University Press, 2013.
Stefano Battiston, Michelangelo Puliga, Rahul Kaushik, Paolo Tasca, and Guido Caldarelli. Debtrank: Too central to fail? financial networks, the [FED]{} and systemic risk. *Scientific reports*, 20 (541), 2012.
Hideaki Aoyama, Stefano Battiston, and Yoshi Fujiwara.. Discussion papers 13087, Research Institute of Economy, Trade and Industry (RIETI), October 2013.
[^1]: <http://www.crisis-economics.eu>
---
abstract: 'In the first part of the paper we introduce some geometric tools needed to describe slow-fast Hamiltonian systems on smooth manifolds. We start with a smooth Poisson bundle $p: M\to B$ of a regular (i.e. of constant rank) Poisson manifold $(M,\omega)$ over a smooth symplectic manifold $(B,\lambda)$, the foliation into leaves of the bundle coincides with the symplectic foliation generated by the Poisson structure on $M$. This defines a singular symplectic structure $\Omega_{\varepsilon}=$ $\omega + \varepsilon^{-1}p^*\lambda$ on $M$ for any positive small $\varepsilon$, where $p^*\lambda$ is a lift of 2-form $\lambda$ on $M$. Given a smooth Hamiltonian $H$ on $M$ one gets a slow-fast Hamiltonian system w.r.t. $\Omega_{\varepsilon}$. We define a slow manifold $SM$ of this system. Assuming $SM$ to be a smooth submanifold, we define a slow Hamiltonian flow on $SM$. The second part of the paper deals with singularities of the restriction of $p$ on $SM$ and their relations with the description of the system near them. It appears, if $\dim M = 4,$ $\dim B = 2$ and Hamilton function $H$ is generic, then behavior of the system near singularities of the fold type is described in the principal approximation by the equation Painlevé-I, but if a singular point is a cusp, then | *, 50:0 29–61, 2015.
Rama Cont, Amal Moussa, and Edson Santos. Network structure and systemic risk in banking systems. In Jean - Pierre Fouque and Joseph A Langsam, editor, * Handbook of Systemic Risk *, page 327–368. Cambridge University Press, 2013.
Stefano Battiston, Michelangelo Puliga, Rahul Kaushik, Paolo Tasca, and Guido Caldarelli. Debtrank: Too central to fail? financial network, the [ FED ] { } and systemic risk. * Scientific reports *, 20 (541), 2012.
Hideaki Aoyama, Stefano Battiston, and Yoshi Fujiwara.. Discussion newspaper 13087, Research Institute of Economy, Trade and Industry (RIETI), October 2013.
[ ^1 ]: < http://www.crisis-economics.eu >
---
abstract:' In the beginning part of the composition we introduce some geometric tools need to describe slow - firm Hamiltonian systems on smooth manifolds. We start with a smooth Poisson bunch $ p: M\to B$ of a regular (i.e. of constant membership) Poisson manifold $ (M,\omega)$ over a smooth symplectic manifold $ (B,\lambda)$, the foliation into leaves of the package coincides with the symplectic foliation generated by the Poisson structure on $ M$. This defines a singular symplectic social organization $ \Omega_{\varepsilon}=$ $ \omega + \varepsilon^{-1}p^*\lambda$ on $ M$ for any positive small $ \varepsilon$, where $ p^*\lambda$ is a lift of 2 - form $ \lambda$ on $ M$. Given a smooth Hamiltonian $ H$ on $ M$ one catch a slow - fast Hamiltonian system w.r.t. $ \Omega_{\varepsilon}$. We define a slow manifold $ SM$ of this system. Assuming $ SM$ to be a smooth submanifold, we define a slow Hamiltonian stream on $ SM$. The second part of the paper deal with singularity of the restriction of $ p$ on $ SM$ and their relations with the description of the system near them. It appear, if $ \dim M = 4,$ $ \dim B = 2 $ and Hamilton function $ H$ is generic, then behavior of the system near singularities of the fold character is described in the principal approximation by the equation Painlevé - I, but if a singular point is a cusp, then | *, 50:0 29–61, 2015.
Gama Cont, Amal Moussa, ana Edson Santos. Uwtwork strucfure and systemic risk in banking sydtwms. Ib Jean-Pierre Fouque ana Joseph A Langsam, ediuors, *Handbook of Systemic Risk*, lwges 327–368. Cambridge Uniyersity Prevs, 2013.
Stefano Batdirtln, Michelangelo Puliga, Rahul Kaushih, Paolo Twsca, and Guido Calcwrelmp. Bebtrank: Too central to fail? fihancial networks, the [FED]{} and systemic risk. *Scifntivic reports*, 20 (541), 2012.
Hidewki Aoyama, Wtefwbo Battiston, and Yoshi Fujiwara.. Diacussion papers 13087, Research Instigute pf Economy, Teadf and Industcy (RIENI), October 2013.
[^1]: <hbnp://www.crhsis-ecomomics.eu>
---
abstrsct: 'In the first part of thx paper we introduce some geokecric tools needed to eewcribg slof-fasg Haoiluonman systels kn smooth janifolds. Ww start with a smoouh [iisson bundle $p: M\to B$ of a regular (i.e. of constant rank) Poissmn janifold $(M,\omega)$ over a wmooth symplectic manlfold $(B,\laibda)$, the foliation into leaves of the bundle coinwides ditk the rtmolectic foliation generated by the Poisson stwhcuurv on $M$. This defincs a singular sympkeftos structure $\Ooega_{\vaxspailon}=$ $\omega + \varepdilon^{-1}p^*\lwmbda$ on $M$ for any positive small $\varepsilon$, qhere $p^*\lambdc$ iw a lift of 2-form $\lcmbda$ on $M$. Gnven a smooyh Hamiltonian $H$ on $M$ oue geta a slow-fash Hamiltohkan system w.r.t. $\Ooegs_{\vdrepsilon}$. We define a slow manifold $SM$ pf this sysjem. Asstming $SM$ tl be a smooth submanifolf, we befina a slow Hwmiltonian flow on $SM$. The seconv part of the pdpeg deals wnth sikgularities of ehe restrictiou of $p$ ou $SM$ avd their rvlations xith the dessription of tvg system near them. It appwars, if $\dim O = 4,$ $\dim B = 2$ anc Hamiltou funxtion $H$ is generic, thgn behavior of thz wystem near sinbulxrieivs pf dhe fold typa is dervribea in tht pvinzipak approximation by tve esuation Painlevé-I, bit if a sibgular pjint is a cusl, then | *, 50:0 29–61, 2015. Rama Cont, Amal Edson Network structure systemic risk in and A Langsam, editors, of Systemic Risk*, 327–368. Cambridge University Press, 2013. Stefano Michelangelo Puliga, Rahul Kaushik, Paolo Tasca, and Guido Caldarelli. Debtrank: Too central to financial networks, the [FED]{} and systemic risk. *Scientific reports*, 20 (541), 2012. Hideaki Stefano and Fujiwara.. papers 13087, Research Institute of Economy, Trade and Industry (RIETI), October 2013. [^1]: <http://www.crisis-economics.eu> --- abstract: the first part of the paper we introduce geometric tools needed to slow-fast Hamiltonian systems on smooth We with a Poisson $p: B$ of a (i.e. of constant rank) Poisson manifold $(M,\omega)$ over a smooth symplectic manifold $(B,\lambda)$, the foliation into leaves the bundle the symplectic generated the structure on $M$. a singular symplectic structure $\Omega_{\varepsilon}=$ $\omega $M$ for any positive small $\varepsilon$, where $p^*\lambda$ a lift 2-form $\lambda$ on $M$. Given a Hamiltonian $H$ on $M$ one gets a slow-fast system w.r.t. $\Omega_{\varepsilon}$. We define a slow manifold $SM$ of this system. Assuming $SM$ to smooth submanifold, we define slow Hamiltonian flow $SM$. second of paper deals singularities of the restriction of $p$ on $SM$ and their relations the description of the system near them. It appears, if = $\dim B = and Hamilton function $H$ generic, behavior of the system of fold in principal by the equation Painlevé-I, if a singular point is cusp, then | *, 50:0 29–61, 2015.
Rama Cont, Amal Moussa, and EdsoN Santos. NetWork sTruCtuRe And sYsteMic risk in bankiNG sysTems. In Jean-Pierre Fouque And JoSePH A LaNGsAm, ediTors, *HanDBoOK Of SYsTeMic riSK*, pAges 327–368. CAmbRidge UnIversity PrEss, 2013.
stEfano BattistON, MIchelangelO PuLiga, Rahul KauShiK, Paolo taSca, ANd GuiDo CAldarElli. DeBTrank: TOo central To FAil? finANcial neTWOrKs, thE [FED]{} and systemic riSK. *SCIentific reportS*, 20 (541), 2012.
HideaKi aOyAMA, StEfaNo BattistoN, aNd YosHI FujiwaRA.. DISCUssIOn papers 13087, ReseaRch InstitutE Of EConomy, trAde ANd InduStry (RiEti), OcTober 2013.
[^1]: <http://wwW.criSis-economIcs.eu>
---
aBStract: 'IN The firsT part oF thE paPer wE InTrOduCe SOme GEoMetRIc tOols needEd To DescrIbe sLOW-FAst HAmiLtonIan syStems on smooth ManIfolDS. We Start With a SmooTh poissOn bundLe $p: M\tO B$ Of a regular (i.e. of cOnstAnt rank) PoIssOn ManIfOld $(M,\oMEga)$ oveR a sMooTh symplEctic maNIfoLd $(b,\LAMbDa)$, the foliation into LeAVEs Of the bunDle coiNCiDeS With the sYmPleCtic FOLiatiOn geNErAted by thE PoissON sTrUcture oN $M$. this deFiNes A siNgulaR SympLectic StructurE $\OmegA_{\Varepsilon}=$ $\omegA + \Varepsilon^{-1}p^*\laMBdA$ ON $M$ FOr anY poSitive small $\VarePSiloN$, wheRE $p^*\LamBDa$ is a Lift oF 2-fORm $\LAmbda$ on $M$. Given a smootH HAmiltoNian $H$ On $M$ one gets a slOw-fast HamiLTONian systEm w.r.T. $\omEGa_{\varepsilon}$. We DefinE a slow maniFOld $SM$ of tHis syStem. AssuMing $SM$ to bE A Smooth suBmaNifOld, We dEFInE a slow HamiltoNIAn flOw On $SM$. The SecOnd part Of tHe pApeR deAlS with singUlaritieS oF tHe ReStrIctioN Of $p$ on $SM$ aNd TheIr RelAtionS With thE descRiptIoN oF The System nEAr THEm. It ApPeArs, iF $\diM M = 4,$ $\Dim B = 2$ aNd HaMIltOn functIon $H$ is genEriC, Then BeHaVior of tHe system near sInGularities Of The Fold tyPE Is descriBed in the principal approxIMation bY thE equaTion painlevé-I, But If a sinGulAR point Is a cusP, then | *, 50:0 29–61, 2015.
Rama Cont, Ama l Mou ssa , a nd Eds on S antos. Network stru cture and systemic ris k inba n king sy stems . In Je a n- P i err eFo uqu ea nd Jose phA Langs am, editor s,*H andbook of S y st emic Risk* , p ages 327–368 . C ambrid ge Un i versi tyPress , 2013 .
Stef ano Batti st o n, Mic h elangel o Pu liga , Rahul Kaushik,P ao l o Tasca, and G uido C al d ar e l li. De btrank: To ocentr a l to fa i l? f i nan c ial networks, the [FED]{ } an d syst em icr isk. * Scien ti f icreports*, 2 0 (5 41), 2012 .
Hid e aki Aoy a ma, Ste fano B att ist on,a nd Y osh iF uji w ar a.. Dis cussionpa pe rs 13 087, R e s earc h I nsti tuteof Economy, T rad e an d In dustr y (RI ETI) ,Octob er 201 3.
[ ^1 ]: <http://www. cris is-econom ics .e u> - --
ab s tract: 'I n t he firs t parto f t he p a pe r we introduce som eg e om etric to ols ne e de dt o descri be sl ow-f a s t Ham ilto n ia n system s on s m oo th manifo ld s. West art wi th as moot h Pois son bund le $p : M\to B$ of ar egular (i.e.o fc o ns t antran k) Poissonmani f old$(M, \ om ega ) $ ove r a s mo o th symplectic manifold $ (B,\la mbda) $, the foliat ion into l e a v es of th e bu n dl e coincides wit h the symplecti c foliati on ge neratedby the Po i s son stru ctu reon$M$ . Th is defines as i ngul ar symple cti c struc tur e $ \Om ega _{ \varepsil on}=$ $\ om eg a+\va repsi l on^{-1}p ^* \la mb da$ on $ M $ forany p osit iv es mal l $\var e ps i l on$, w he re $ p^* \l ambda $ is a l ift of2-form $\ lam b da$on $ M$. Giv en a smooth H am iltonian $ H$ on $M$ o n e gets aslow-fast Hamiltonian s y stem w. r.t . $\O mega _{\vareps ilo n}$. W e d e fine a slowmanif ol d $ S M $ oft h is sy st em. Assumi n g $S M$ to b e asmoothsubmanifold, we de f ine a slow Hamil ton ianf l ow on $S M $.Th e se c o nd part of thepaper deal sw it h singular i tie sof therestric tiono f $p$ o n $SM$ an d their r el atio n s wi th the des cription of the s y stemn ea r the m.It app ea rs, if $ \dim M = 4 ,$ $\ dim B=2$ and Hami lt on funct ion $H$ is generic, the n beha viorofthe syste m n e arsingulari ties of the fo ldtyp e isdes c ribed int he pr i ncipa l ap p roximatio n b y t h e e quation Pai n l e vé- I, bu t i f a sin gula r point is a cusp , then | *, 50:0_29–61, 2015.
Rama_Cont, Amal Moussa, and_Edson Santos._Network_structure and_systemic_risk in banking_systems. In Jean-Pierre_Fouque and Joseph A Langsam,_editors, *Handbook of_Systemic_Risk*, pages 327–368. Cambridge University Press, 2013.
Stefano Battiston, Michelangelo Puliga, Rahul Kaushik, Paolo Tasca,_and_Guido Caldarelli._Debtrank:_Too_central to fail? financial networks,_the [FED]{} and systemic risk._*Scientific reports*,_20 (541), 2012.
Hideaki Aoyama, Stefano Battiston, and Yoshi_Fujiwara.._Discussion papers 13087,_Research Institute of Economy, Trade and Industry (RIETI), October_2013.
[^1]: <http://www.crisis-economics.eu>
---
abstract: 'In the first_part of the_paper_we_introduce some geometric tools_needed to describe slow-fast Hamiltonian systems_on smooth manifolds. We start with_a smooth Poisson bundle $p: M\to B$_of a regular (i.e. of constant_rank) Poisson manifold $(M,\omega)$ over_a smooth_symplectic manifold $(B,\lambda)$, the foliation_into leaves of_the bundle_coincides with the_symplectic foliation generated by the Poisson_structure on $M$._This defines a singular symplectic structure_$\Omega_{\varepsilon}=$_$\omega + \varepsilon^{-1}p^*\lambda$_on_$M$_for any_positive small $\varepsilon$,_where_$p^*\lambda$ is_a_lift of 2-form $\lambda$ on $M$._Given_a smooth Hamiltonian $H$ on $M$ one_gets a slow-fast Hamiltonian_system_w.r.t. $\Omega_{\varepsilon}$. We define_a slow manifold $SM$ of_this system. Assuming $SM$ to be_a smooth_submanifold, we_define a slow Hamiltonian flow on $SM$. The second part of_the paper deals with singularities of_the restriction of $p$_on $SM$_and_their relations with_the_description of_the system near them. It appears, if_$\dim M_= 4,$ $\dim B = 2$_and Hamilton function $H$_is_generic, then behavior of the system_near singularities of the fold type_is described in the principal_approximation_by_the equation Painlevé-I, but if_a singular point is a cusp,_then |
also worth mentioning that one always has $$\mathrm{cd}_{\mathbb{Q}}(G)\leq \underline{\mathrm{cd}}_{\mathcal{M}}(G),$$ where $\mathrm{cd}_{\mathbb{Q}}(G)$ denotes the rational cohomological dimension of $G$. This follows from the fact that if $F_{\ast}\rightarrow \underline{A}$ is a free resolution in $\mathrm{Mack}_{\mathcal{F}}G$, then $F(G/e)\otimes_{\mathbb{Z}}\mathbb{Q} \rightarrow A(G/e)\otimes_{\mathbb{Z}}\mathbb{Q}=\mathbb{Q} $ is a projective $\mathbb{Q}[G]$-resolution of $\mathbb{Q}$. Note that this inequality can be strict (e.g. see [@DavisBook Ex. 8.5.8]) and that it implies that $\underline{\mathrm{cd}}_{\mathcal{M}}(G)=0$ if and only if $G$ is finite.
Proper $G$-spaces and proper $G$-spectra
========================================
Throughout this section, let $G$ be a discrete group and let $\mathcal{F}$ be the family of finite subgroups of $G$. If $X$ is a $G$-space, then $X_{+}$ denotes the space obtained by adding a disjoint $G$-fixed basepoint $\ast$ to $X$. In what follows we will freely make use of the language of triangulated categories (e.g. see [@Gelfand-Manin]) and model categories (e.g. see [@Hovey]).
Denote by $G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast}$ the model category of compactly generated weak Hausdorff spaces equipped with a continuous $G$-action and $G$-fixed basepoint together with $G$-equivariant based continuous maps, where weak-equivalences and fibrations are required to be weak-equivalences and fibrations on $H$-fixed point spaces, for all $H \in \mathcal{F}$ (see [@Fau Section 2.1-2.3]). The associated homotopy category is denoted by $\mathrm{Ho}(G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast})$. Note that if $H$ is a subgroup of $G$, then $G/H_{+}$ is a cofibrant object in $G\mbox{-}\mathrm{Top}^{\mathcal{ | also worth mentioning that one always take $ $ \mathrm{cd}_{\mathbb{Q}}(G)\leq \underline{\mathrm{cd}}_{\mathcal{M}}(G),$$ where $ \mathrm{cd}_{\mathbb{Q}}(G)$ announce the rational cohomological dimension of $ G$. This follows from the fact that if $ F_{\ast}\rightarrow \underline{A}$ is a detached resolution in $ \mathrm{Mack}_{\mathcal{F}}G$, then $ F(G / e)\otimes_{\mathbb{Z}}\mathbb{Q } \rightarrow A(G / e)\otimes_{\mathbb{Z}}\mathbb{Q}=\mathbb{Q } $ is a projective $ \mathbb{Q}[G]$-resolution of $ \mathbb{Q}$. Note that this inequality can be rigid (for example see [ @DavisBook Ex. 8.5.8 ]) and that it implies that $ \underline{\mathrm{cd}}_{\mathcal{M}}(G)=0 $ if and only if $ G$ is finite.
Proper $ G$-spaces and proper $ G$-spectra
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Throughout this section, let $ G$ be a discrete group and lease $ \mathcal{F}$ be the family of finite subgroups of $ G$. If $ X$ is a $ G$-space, then $ X_{+}$ announce the space obtained by adding a disjoint $ G$-fixed basepoint $ \ast$ to $ X$. In what follows we will freely make function of the language of triangulated categories (for example see [ @Gelfand - Manin ]) and model class (e.g. understand [ @Hovey ]).
Denote by $ G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast}$ the model category of compactly generated weak Hausdorff space equipped with a continuous $ G$-action and $ G$-fixed basepoint together with $ G$-equivariant based continuous maps, where weak - comparison and fibrations are required to be weak - equivalences and fibrations on $ H$-fixed point spaces, for all $ H \in \mathcal{F}$ (see [ @Fau Section 2.1 - 2.3 ]). The associated homotopy category is denote by $ \mathrm{Ho}(G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast})$. Note that if $ H$ is a subgroup of $ G$, then $ G / H_{+}$ is a cofibrant object in $ G\mbox{-}\mathrm{Top}^{\mathcal { | aldo worth mentioning that one always has $$\mathrm{rd}_{\mathbg{Q}}(G)\leq \uvderline{\mathrm{cd}}_{\mathcal{M}}(G),$$ whxre $\nathrn{cd}_{\mathbb{Q}}(G)$ denotes thd rationap cohomooogiral dimension of $G$. This njllosd frmn the fact thaj if $F_{\ast}\rigvtarrow \underlhnd{A}$ is a free resolution in $\mathrm{Mack}_{\iathcal{G}}G$, then $F(G/e)\otimef_{\matnfb{Z}}\mznhnb{Q} \rightarrow A(G/e)\otimes_{\mathbb{Z}}\mzthbb{Q}=\methbb{Q} $ is a prpjective $\mathbb{Q}[G]$-resolutioj of $\mathbb{Q}$. Note that this inequqlitr can be strizt (e.g. see [@DavisBook Ex. 8.5.8]) and that it implies that $\undefline{\kathrm{cd}}_{\majkxal{L}}(C)=0$ if and onoy if $G$ is finite.
Igoper $G$-vpaces snd proper $G$-spcctra
========================================
Vhroyghout this section, lxt $G$ be a discrete gtoup and lat $\mathcal{F}$ be the danily mf fhnitd sucgrku's kf $G$. Iv $X$ is a $G$-spade, then $X_{+}$ dwnotes the space obuaigvc by adding z disjjigt $G$-fixed basepoint $\ast$ to $X$. In what fonloss we will freely make yse of the language ov triangujated categories (e.g. see [@Gelfand-Manin]) and model catagorixs (e.g. wec [@Hoxwy]).
Fenote by $G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast}$ the model dauegpry of compactjy generatec aesh Hausdorff skaces equilped with a continkous $G$-astion and $G$-fixtd baxepoint together with $G$-equicariant baseb cintinuous maps, whexe weak-equivclencex and fibrations are requireb to bs weak-equivwlences aha fibrations on $F$-fiqed [oint spaces, for all $H \in \iathcal{F}$ (wee [@Yau Sectkon 2.1-2.3]). The afsociated jomotopy category is denlted yy $\madhrm{Ho}(G\mbod{-}\mathrm{Top}^{\mathcal{F}}_{\ast})$. Note that if $H$ is a subbrmup of $G$, thzn $G/H_{+}$ ls a cofibrant jbject in $G\mbor{-}\mathrm{Tjp}^{\matfcal{ | also worth mentioning that one always has where denotes the cohomological dimension of fact if $F_{\ast}\rightarrow \underline{A}$ a free resolution $\mathrm{Mack}_{\mathcal{F}}G$, then $F(G/e)\otimes_{\mathbb{Z}}\mathbb{Q} \rightarrow A(G/e)\otimes_{\mathbb{Z}}\mathbb{Q}=\mathbb{Q} $ a projective $\mathbb{Q}[G]$-resolution of $\mathbb{Q}$. Note that this inequality can be strict (e.g. [@DavisBook Ex. 8.5.8]) and that it implies that $\underline{\mathrm{cd}}_{\mathcal{M}}(G)=0$ if and only if is Proper and $G$-spectra ======================================== Throughout this section, let $G$ be a discrete group and let $\mathcal{F}$ be the of finite subgroups of $G$. If $X$ is $G$-space, then $X_{+}$ denotes space obtained by adding a $G$-fixed $\ast$ to In follows will freely make of the language of triangulated categories (e.g. see [@Gelfand-Manin]) and model categories (e.g. see [@Hovey]). Denote by the model compactly generated Hausdorff equipped a continuous $G$-action basepoint together with $G$-equivariant based continuous and fibrations are required to be weak-equivalences and on $H$-fixed spaces, for all $H \in \mathcal{F}$ [@Fau Section 2.1-2.3]). The associated homotopy category is by $\mathrm{Ho}(G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast})$. Note that if $H$ is a subgroup of $G$, then $G/H_{+}$ is a in $G\mbox{-}\mathrm{Top}^{\mathcal{ | also worth mentioning that onE always has $$\MathrM{cd}_{\MatHbB{Q}}(G)\lEq \unDerline{\mathrm{cD}}_{\MathCal{M}}(G),$$ where $\mathrm{cd}_{\mathBb{Q}}(G)$ dEnOTes tHE rAtionAl cohomOLoGICal DiMeNsiOn OF $G$. this fOllOws from The fact thaT if $f_{\aSt}\rightarrow \UNdErline{A}$ is a FreE resolution iN $\maThrm{MaCk}_{\MatHCal{F}}G$, TheN $F(G/e)\oTimes_{\mAThbb{Z}}\mAthbb{Q} \rigHtARrow A(G/E)\Otimes_{\mATHbB{Z}}\maThbb{Q}=\mathbb{Q} $ is a prOJeCTive $\mathbb{Q}[G]$-reSolutiOn OF $\mATHbb{q}$. NoTe that this InEqualITy can be STrICT (E.g. sEE [@DavisBook Ex. 8.5.8]) aNd that it impLIes That $\unDeRliNE{\mathrM{cd}}_{\maThCAl{M}}(g)=0$ if and only iF $G$ is Finite.
ProPer $G$-spACes and pROper $G$-spEctra
========================================
THroUghOut tHIs SeCtiOn, LEt $G$ BE a DisCRetE group anD lEt $\MathcAl{F}$ bE THE FamiLy oF finIte suBgroups of $G$. If $X$ Is a $g$-spaCE, thEn $X_{+}$ deNotes The sPaCe obtAined bY addiNg A disjoint $G$-fixed BasePoint $\ast$ tO $X$. IN wHat FoLlows WE will fReeLy mAke use oF the lanGUagE oF TRIaNgulated categories (E.g. SEE [@GElfand-MaNin]) and MOdEl CAtegorieS (e.G. seE [@HovEY]).
denotE by $G\MBoX{-}\mathrm{TOp}^{\mathCAl{f}}_{\aSt}$ the moDeL categOrY of ComPactlY GeneRated wEak HausdOrff sPAces equipped wiTH a continuous $G$-ACtION aND $G$-fiXed Basepoint toGethER witH $G$-eqUIvAriANt basEd conTiNUoUS maps, where weak-equivAlEnces aNd fibRations are reqUired to be wEAK-EquivaleNces ANd FIbrations on $H$-fiXed poInt spaces, fOR all $H \in \mAthcaL{F}$ (see [@Fau section 2.1-2.3]). ThE ASsociateD hoMotOpy CatEGOrY is denoted by $\mATHrm{HO}(G\Mbox{-}\matHrm{top}^{\mathCal{f}}_{\asT})$. NoTe tHaT if $H$ is a suBgroup of $g$, tHeN $G/h_{+}$ iS a cOfibrANt object In $g\mbOx{-}\MatHrm{ToP}^{\MathcaL{ | also worth mentioning tha t one alwa ys ha s $ $\m at hrm{ cd}_ {\mathbb{Q}}(G ) \leq \underline{\mathrm{cd }}_{\ ma t hcal { M} }(G), $$ wher e $ \ m ath rm {c d}_ {\ m at hbb{Q }}( G)$ den otes the r ati on al cohomolog i ca l dimensio n o f $G$. Thisfol lows f ro m t h e fac t t hat i f $F_{ \ ast}\r ightarrow \ u nderli n e{A}$ i s afree resolution in $\ m at h rm{Mack}_{\mat hcal{F }} G $, t hen $F (G/e)\otim es _{\ma t hbb{Z}} \ ma t h b b{Q } \rightarrowA(G/e)\otim e s_{ \mathb b{ Z}} \ mathbb {Q}=\ ma t hbb {Q} $ is aproj ective $\ mathbb { Q}[G]$- r esoluti on of$\m ath bb{Q } $. N ote t h att hi s i n equ ality ca nbe stri ct ( e . g . see [@ Davi sBook Ex. 8.5.8])and tha t it impl ies t hat$\ under line{\ mathr m{ cd}}_{\mathcal{ M}}( G)=0$ ifand o nly i f $G$ is fin ite .
Proper$G$-spa c esan d p ro per $G$-spectra
== == = = == ======== ====== = == == = ======== == ==
Thr o u ghout thi s s ection,let $G $ b ea discr et e grou pand le t $\m a thca l{F}$be the f amily of finite subg r oups of $G$.I f$ X $i s a$G$ -space, the n $X _ {+}$ den o te s t h e spa ce ob ta i ne d by adding a disjoi nt $G$-f ixedbasepoint $\a st$ to $X$ . I n what f ollo w sw e will freelymakeuse of the language of t riangula ted categ o r ies (e.g . s ee[@G elf a n d- Manin]) and m o d el c at egories (e .g. see [@ Hov ey] ).
D enote by$G\mbox{ -} \m at hr m{T op}^{ \ mathcal{ F} }_{ \a st} $ the modelcateg oryof c o mpa ctly ge n er a t ed w ea kHaus dor ff spac es e q uip ped wit h a conti nuo u s $G $- ac tion an d $G$-fixed b as epoint tog et her with$ G $-equiva riant based continuousm aps, wh ere weak -equ ivalences an d fibr ati o ns are requi red t obew e ak-eq u i va len ce s and fibr a t ion s on$H $-fi xed poi nt spaces, for all $H\in \mathcal{ F}$ (se e [@ Fau Se c tio n2 .1- 2 . 3]). The associ ated homot op y c ategory is den ot ed by $ \mathrm {Ho}( G \mbox{- }\mathrm{ Top}^{\ma th cal{ F } }_{ \ast})$. N ote that if $H$ i s a su b gr oup o f $ G$, th en $G /H_{+ }$ isa co fibra nt obj ec t in $ G\mbo x{ -}\mathr m{Top}^{\mathcal{ | also_worth mentioning_that one always has_$$\mathrm{cd}_{\mathbb{Q}}(G)\leq \underline{\mathrm{cd}}_{\mathcal{M}}(G),$$_where_$\mathrm{cd}_{\mathbb{Q}}(G)$ denotes_the_rational cohomological dimension_of $G$. This_follows from the fact_that if $F_{\ast}\rightarrow_\underline{A}$_is a free resolution in $\mathrm{Mack}_{\mathcal{F}}G$, then $F(G/e)\otimes_{\mathbb{Z}}\mathbb{Q} \rightarrow A(G/e)\otimes_{\mathbb{Z}}\mathbb{Q}=\mathbb{Q} $ is a projective_$\mathbb{Q}[G]$-resolution_of $\mathbb{Q}$._Note_that_this inequality can be strict_(e.g. see [@DavisBook Ex. 8.5.8])_and that_it implies that $\underline{\mathrm{cd}}_{\mathcal{M}}(G)=0$ if and only if_$G$_is finite.
Proper $G$-spaces_and proper $G$-spectra
========================================
Throughout this section, let $G$ be a_discrete group and let $\mathcal{F}$ be_the family of_finite_subgroups_of $G$. If $X$_is a $G$-space, then $X_{+}$ denotes_the space obtained by adding a_disjoint $G$-fixed basepoint $\ast$ to $X$. In_what follows we will freely make_use of the language of_triangulated categories_(e.g. see [@Gelfand-Manin]) and model categories_(e.g. see [@Hovey]).
Denote by_$G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast}$ the_model category of_compactly generated weak Hausdorff spaces equipped_with a continuous_$G$-action and $G$-fixed basepoint together with_$G$-equivariant_based continuous maps,_where_weak-equivalences_and fibrations_are required to_be_weak-equivalences and_fibrations_on $H$-fixed point spaces, for all_$H_\in \mathcal{F}$ (see [@Fau Section 2.1-2.3]). The_associated homotopy category is_denoted_by $\mathrm{Ho}(G\mbox{-}\mathrm{Top}^{\mathcal{F}}_{\ast})$. Note that_if $H$ is a subgroup_of $G$, then $G/H_{+}$ is a_cofibrant object_in $G\mbox{-}\mathrm{Top}^{\mathcal{ |
1$.
Suppose that the word $W(i,j)a_k^\ell$ has property (E). Consider all possible factorizations for $W(i,j)a_k^\ell$ of the form $$\begin{gathered}
\label{eeqq2}
W(i,j)a_k^\ell \equiv U_1 U_2,\end{gathered}$$ where $|U_1|, |U_2| \ge 1$. Let us show that either of the words $U_1, U_2$ can be represented in a form $W[i', j', k',\ell']$ so that $W[i', j', k',\ell'] \prec W[i, j, k,\ell]$.
First suppose $|U_1| \le |W(i,j)| =j$. If $i +|U_1| \le |W|$, we have $$U_1 \overset \star = W[i, |U_1|, 1,0], \quad U_2 \overset \star = W[i +|U_1|, j-|U_1|,k,\ell].$$ On the other hand, if $i +|U_1| = |W|+1$, we have $$U_1 \overset \star = W[i, |U_1|, 1,0], \quad U_2\overset \star = W[|W|, 0,k,\ell].$$
Now suppose $|U_1| > |W(i,j)| =j$. Let $\ell_1, \ell_2$ be integers such that $\ell_1, \ell_2, \ell$ have the same sign, $ \ell = \ell_1+\ell_2$, and $j + |\ell| = |U_1| +\ell_2$. Then we have $$U_1 \overset \star = W[i, |U_1|, k,\ell_1], \quad U_2 \overset \star = W[ 1,0, k, \ell_2].$$
Note that the induction parameter $(j', |\ell'|)$ for indicated representations of $U_1, U_2$ is smaller than the parameter | 1$.
Suppose that the word $ W(i, j)a_k^\ell$ has property (east). think all possible factorizations for $ W(i, j)a_k^\ell$ of the form $ $ \begin{gathered }
\label{eeqq2 }
W(i, j)a_k^\ell \equiv U_1 U_2,\end{gathered}$$ where $ |U_1|, |U_2| \ge 1$. get us show that either of the words $ U_1, U_2 $ can be represent in a shape $ W[i', j', k',\ell']$ so that $ W[i', j', k',\ell' ] \prec W[i, j, k,\ell]$.
First think $ |U_1| \le |W(i, j)| = j$. If $ i + |U_1| \le |W|$, we have $ $ U_1 \overset \star = W[i, |U_1|, 1,0 ], \quad U_2 \overset \star = W[i + |U_1|, j-|U_1|,k,\ell].$$ On the other bridge player, if $ i + |U_1| = |W|+1 $, we have $ $ U_1 \overset \star = W[i, |U_1|, 1,0 ], \quad U_2\overset \star = W[|W|, 0,k,\ell].$$
Now presuppose $ |U_1| > |W(i, j)| = j$. Let $ \ell_1, \ell_2 $ be integers such that $ \ell_1, \ell_2, \ell$ have the same sign, $ \ell = \ell_1+\ell_2 $, and $ joule + |\ell| = |U_1| + \ell_2$. Then we have $ $ U_1 \overset \star = W[i, |U_1|, k,\ell_1 ], \quad U_2 \overset \star = W [ 1,0, k, \ell_2].$$
Note that the induction parameter $ (joule', |\ell'|)$ for indicated representations of $ U_1, U_2 $ is smaller than the parameter | 1$.
Suopose that the word $W(i,j)a_y^\ell$ has properjy (E). Convider zll posskble factorizations for $W(i,j)a_n^\eol$ of the form $$\begin{gatherea}
\label{eeqe2}
W(i,j)a_k^\eol \ewyiv U_1 U_2,\end{jzthered}$$ where $|M_1|, |U_2| \gz 1$. Let us show tmat either mf the words $U_1, U_2$ ccn be represented in a form $W[i', j', k',\elj']$ so thst $W[i', j', k',\ell'] \pres W[i, t, k,\eml]$.
First suppose $|U_1| \le |W(i,j)| =j$. If $i +|U_1| \le |W|$, wt have $$U_1 \overset \syar = W[i, |U_1|, 1,0], \quad U_2 \oversft \shar = W[i +|U_1|, j-|U_1|,k,\ell].$$ On the othgd hwbd, if $i +|U_1| = |W|+1$, we have $$U_1 \overset \stat = W[i, |U_1|, 1,0], \quad U_2\overset \staf = W[|W|, 0,k,\ell].$$
Noq wupomse $|U_1| > |W(i,j)| =j$. Len $\ell_1, \ell_2$ be lmtegerv such yhat $\ell_1, \ell_2, \ekl$ iave the same sign, $ \ell = \xll_1+\ell_2$, and $j + |\ell| = |U_1| +\ell_2$. Then wz have $$U_1 \overset \star = W[i, |U_1|, k,\eln_1], \quxe U_2 \kvxrsst \stag = W[ 1,0, k, \elm_2].$$
Note that rhe induction paramttew $(j', |\ell'|)$ for insicateq wepresentations of $U_1, U_2$ is smaller than uhe pzrameter | 1$. Suppose that the word $W(i,j)a_k^\ell$ has Consider possible factorizations $W(i,j)a_k^\ell$ of the U_1 where $|U_1|, |U_2| 1$. Let us that either of the words $U_1, can be represented in a form $W[i', j', k',\ell']$ so that $W[i', j', \prec W[i, j, k,\ell]$. First suppose $|U_1| \le |W(i,j)| =j$. If $i +|U_1| |W|$, have \overset = W[i, |U_1|, 1,0], \quad U_2 \overset \star = W[i +|U_1|, j-|U_1|,k,\ell].$$ On the other hand, $i +|U_1| = |W|+1$, we have $$U_1 \overset = W[i, |U_1|, 1,0], U_2\overset \star = W[|W|, 0,k,\ell].$$ suppose > |W(i,j)| Let \ell_2$ integers such that \ell_2, \ell$ have the same sign, $ \ell = \ell_1+\ell_2$, and $j + |\ell| = |U_1| +\ell_2$. we have \star = |U_1|, \quad \overset \star = k, \ell_2].$$ Note that the induction for indicated representations of $U_1, U_2$ is smaller the parameter | 1$.
Suppose that the word $W(i,j)a_k^\elL$ has properTy (E). CoNsiDer AlL posSiblE factorizationS For $W(I,j)a_k^\ell$ of the form $$\begin{gAtherEd}
\LAbel{EEqQ2}
W(i,j)a_K^\ell \equIV U_1 u_2,\ENd{gAtHeRed}$$ WhERe $|u_1|, |U_2| \ge 1$. LEt uS show thAt either of The WoRds $U_1, U_2$ can be rePReSented in a fOrm $w[i', j', k',\ell']$ so thaT $W[i', J', k',\ell'] \pReC W[i, J, K,\ell]$.
FIrsT suppOse $|U_1| \le |w(I,j)| =j$. If $i +|u_1| \le |W|$, we havE $$U_1 \OVerset \STar = W[i, |U_1|, 1,0], \qUAD U_2 \OverSet \star = W[i +|U_1|, j-|U_1|,k,\ell].$$ ON ThE Other hand, if $i +|U_1| = |W|+1$, We have $$u_1 \oVErSET \stAr = W[I, |U_1|, 1,0], \quad U_2\oveRsEt \staR = w[|W|, 0,k,\ell].$$
NOW sUPPOse $|u_1| > |w(i,j)| =j$. Let $\ell_1, \ell_2$ Be integers sUCh tHat $\ell_1, \ElL_2, \elL$ Have thE same SiGN, $ \elL = \ell_1+\ell_2$, and $j + |\Ell| = |U_1| +\Ell_2$. Then we Have $$U_1 \oVErset \stAR = W[i, |U_1|, k,\elL_1], \quad U_2 \OveRseT \staR = w[ 1,0, k, \ElL_2].$$
NoTe THat THe IndUCtiOn parameTeR $(j', |\Ell'|)$ foR indICATEd rePreSentAtionS of $U_1, U_2$ is smalleR thAn thE ParAmeteR | 1$.
Suppose that the wor d $W(i,j)a _k^\e ll$ ha sprop erty (E). Consider allpossible factorization s for $ W (i,j ) a_ k^\el l$ of t h ef o rm$$ \b egi n{ g at hered }
\ label{e eqq2}
W(i ,j) a_ k^\ell \equi v U _1 U_2,\en d{g athered}$$ w her e $|U_ 1| , | U _2| \ ge1$. L et uss how th at either o f the w o rds $U_ 1 , U _2$can be represente d i n a form $W[i', j', k ', \ el l ' ]$sothat $W[i' ,j', k ' ,\ell'] \p r e c W[ i , j, k,\ell]$ .
First su p pos e $|U_ 1| \ l e |W(i ,j)|=j $ . I f $i +|U_1| \le |W|$, we have$ $U_1 \o v erset \ star = W[ i,|U_1 | ,1, 0], \ qua d U_ 2 \o verset \ st ar = W[i+ | U _ 1|,j-| U_1| ,k,\e ll].$$ On the ot herh and , if$i +| U_1| = |W|+ 1$, we have $ $U_1 \overset \ star = W[i,|U_ 1| , 1 ,0 ], \ q uad U _2\ overset \star = W [ | W |, 0,k,\ell].$$
Now s u p po se $|U_1 | > | W (i ,j ) | =j$. L et $\ ell_ 1 , \ell _2$b eintegers sucht ha t$\ell_1 ,\ell_2 ,\el l$havet he s ame si gn, $ \e ll =\ ell_1+\ell_2$, and $j + |\el l |= |U _ 1| + \el l_2$. Thenwe h a ve $ $U_1 \o ver s et \s tar =W [i , |U_1|, k,\ell_1],\q uad U_2 \ overset \star = W[ 1 , 0 , k, \ell _2]. $ $Note that theinduc tion param e ter $(j' , |\e ll'|)$ f or indica t e d repres ent ati ons of $ U_ 1, U_2$ is sm a l lerth an thepar ameter | 1$.
Suppose_that the_word $W(i,j)a_k^\ell$ has property_(E). Consider_all_possible factorizations_for_$W(i,j)a_k^\ell$ of the_form $$\begin{gathered}
\label{eeqq2}
W(i,j)a_k^\ell_\equiv U_1 U_2,\end{gathered}$$ where_$|U_1|, |U_2| \ge_1$._Let us show that either of the words $U_1, U_2$ can be represented in_a_form $W[i',_j',_k',\ell']$_so that $W[i', j', k',\ell']_\prec W[i, j, k,\ell]$.
First suppose_$|U_1| _\le |W(i,j)| =j$. If $i +|U_1| \le |W|$,_we_have $$U_1 \overset_\star = W[i, |U_1|, 1,0], \quad _U_2 \overset \star = _W[i +|U_1|, j-|U_1|,k,\ell].$$_On_the_other hand, if $i_+|U_1| = |W|+1$, we have $$U_1_\overset \star = W[i, |U_1|,_1,0], \quad _U_2\overset \star = _W[|W|, 0,k,\ell].$$
Now suppose $|U_1| _> |W(i,j)|_=j$. Let $\ell_1, \ell_2$ be_integers such that_$\ell_1, \ell_2,_\ell$ have the_same sign, $ \ell = \ell_1+\ell_2$,_and $j +_|\ell| = |U_1| +\ell_2$. Then we_have_$$U_1 \overset \star__=_W[i, |U_1|,_k,\ell_1], \quad __U_2 \overset_\star_ = W[ 1,0,_k,_\ell_2].$$
Note that the induction parameter $(j', |\ell'|)$_for indicated representations of_$U_1,_U_2$ is smaller than_the parameter |
$ and a labeling function is constructed for each subformula of $f$. As seen from the algorithm $CheckCTL$ (at the end of Section \[Ideas\]), the algorithm either gives a definite “yes” or “no” answer to the CTL model-checking problem, i.e., $\langle M,X\rangle, s_0\models f$, or it reduces the problem to checking whether the ID expression $\psi$ labeled to $s_0$ can be evaluated true at the state. The evaluation procedure is carried out by the following recursive procedure $TestWG$, after an input sequence $\pi$ has been accepted by the unspecified component $X$.
[**Procedure**]{} $TestWG(X,\pi,s_0,\psi)$
$\psi=\psi_1\vee \psi_2$:
$TestWG(X,\pi,s_0,\psi_1)$ [**Then**]{}
$true$;
$TestWG(X,\pi,s_0,\psi_2)$
$\psi=\neg \psi_1$:
$\neg TestWG(X,\pi,s_0, \psi_1)$
$\psi=1$:
$true$;
$\psi=i$ with $2\le i\le k+1$:
${\cal I}^{-1}(i)$ is an $EX$ graph
$TestEX(X,\pi,s_0,{\cal I}^{-1}(i))$;
${\cal I}^{-1}(i)$ is an $EU$ graph
$TestEU(X,\pi,s_0,{\cal I}^{-1}(i), level=0)$;
${\cal I}^{-1}(i)$ is an $EG$ graph
$TestEG(X,\pi,s_0,{\cal I}^{-1}(i)).$
In $TestWG$, the first three cases are straightforward, which are consistent with the intended meaning of ID expressions. The cases $TestEX, TestEU, TestEG$ for evaluating $EX, EU$, $EG$ graphs are discussed in the following three subsections.
### $TestEX$
The case for checking whether an $EX$ graph $G=\langle N,E, L_g\rangle$ can be evaluated true at a state $s_0$ is simple. We just test whether the system $ | $ and a labeling function is constructed for each subformula of $ f$. As seen from the algorithm $ CheckCTL$ (at the end of incision \[Ideas\ ]), the algorithm either render a definite “ yes ” or “ no ” answer to the CTL model - checking trouble, i.e., $ \langle M, X\rangle, s_0\models f$, or it reduces the problem to checking whether the ID saying $ \psi$ labeled to $ s_0 $ can be evaluated truthful at the state. The evaluation procedure is post out by the following recursive procedure $ TestWG$, after an remark succession $ \pi$ has been accepted by the unspecified component $ X$.
[ * * operation * * ] { } $ TestWG(X,\pi, s_0,\psi)$
$ \psi=\psi_1\vee \psi_2 $:
$ TestWG(X,\pi, s_0,\psi_1)$ [ * * Then * * ] { }
$ true$;
$ TestWG(X,\pi, s_0,\psi_2)$
$ \psi=\neg \psi_1 $:
$ \neg TestWG(X,\pi, s_0, \psi_1)$
$ \psi=1 $:
$ true$;
$ \psi = i$ with $ 2\le i\le k+1 $:
$ { \cal I}^{-1}(i)$ is an $ EX$ graph
$ TestEX(X,\pi, s_0,{\cal I}^{-1}(i))$;
$ { \cal I}^{-1}(i)$ is an $ EU$ graph
$ TestEU(X,\pi, s_0,{\cal I}^{-1}(i), level=0)$;
$ { \cal I}^{-1}(i)$ is an $ EG$ graph
$ TestEG(X,\pi, s_0,{\cal I}^{-1}(i)).$
In $ TestWG$, the first three cases are square, which are consistent with the intended meaning of ID expressions. The case $ TestEX, TestEU, TestEG$ for evaluating $ EX, EU$, $ EG$ graphs are discussed in the following three subsection.
# # # $ TestEX$
The case for control whether an $ EX$ graph $ G=\langle N, E, L_g\rangle$ can be evaluated true at a state $ s_0 $ is childlike. We just test whether the system $ | $ anf a labeling function is constructed fot wach snbformuma of $f$. Xs seen from the algorithm $CieckXTL$ (au the end of Sectiov \[Ideas\]), tje algoruthm wither givxa a defluite “gcs” or “io” answer to thg CTL model-cvecking problek, k.e., $\langle M,X\rangle, s_0\models f$, or it requces tne problem to chgckinb whefher the ID expression $\psi$ labeled to $s_0$ cen be evaluated true at the state. The evapuatlon procedure is cwrried out vy trw following fecursive krmcedure $TeatWG$, after an input sequence $\pi$ has yeen acceptgb by jhe unspecifmed coiponent $X$.
[**Progvdure**]{} $TastWG(X,\po,s_0,\psi)$
$\psi=\psi_1\vee \psm_2$:
$TesrWG(X,\pi,s_0,\psi_1)$ [**Then**]{}
$true$;
$TevtWG(X,\pi,s_0,\psi_2)$
$\psi=\neg \ksi_1$:
$\neg TesdWY(X,\pi,s_0, \psi_1)$
$\psi=1$:
$true$;
$\psi=i$ qirh $2\le i\le k+1$:
${\cau I}^{-1}(i)$ is ai $ES$ grapj
$TeatEX(X,\pi,s_0,{\cam I}^{-1}(i))$;
${\cal I}^{-1}(i)$ us an $EU$ graph
$TestEI(X,\[p,x_0,{\cal I}^{-1}(i), levem=0)$;
${\cal I}^{-1}(y)$ ys an $EG$ graph
$TestEG(X,\pi,s_0,{\cal I}^{-1}(i)).$
In $TestWG$, ths first three cases are straightforward, whicj are confistent with the intended meaning of ID expressiots. Thx zasts $TestDZ, HestEU, TestEG$ for evaluating $EX, EU$, $EG$ graphs wde dpscussed in the fjllowing thtef xobsections.
### $TesjEX$
The ezss for checking whehher an $EX$ geaph $G=\landle M,E, L_g\rangle$ can be evaluatee true at a frate $s_0$ is simple. Wz just test chethet the xystem $ | $ and a labeling function is constructed subformula $f$. As from the algorithm Section the algorithm either a definite “yes” “no” answer to the CTL model-checking i.e., $\langle M,X\rangle, s_0\models f$, or it reduces the problem to checking whether ID expression $\psi$ labeled to $s_0$ can be evaluated true at the state. evaluation is out the following recursive procedure $TestWG$, after an input sequence $\pi$ has been accepted by the unspecified $X$. [**Procedure**]{} $TestWG(X,\pi,s_0,\psi)$ $\psi=\psi_1\vee \psi_2$: $TestWG(X,\pi,s_0,\psi_1)$ [**Then**]{} $true$; $\psi=\neg \psi_1$: $\neg TestWG(X,\pi,s_0, $\psi=1$: $true$; $\psi=i$ with $2\le k+1$: I}^{-1}(i)$ is $EX$ $TestEX(X,\pi,s_0,{\cal ${\cal I}^{-1}(i)$ is $EU$ graph $TestEU(X,\pi,s_0,{\cal I}^{-1}(i), level=0)$; ${\cal I}^{-1}(i)$ is an $EG$ graph $TestEG(X,\pi,s_0,{\cal I}^{-1}(i)).$ In $TestWG$, the first cases are are consistent the meaning ID expressions. The TestEU, TestEG$ for evaluating $EX, EU$, discussed in the following three subsections. ### $TestEX$ case for whether an $EX$ graph $G=\langle N,E, can be evaluated true at a state $s_0$ simple. We just test whether the system $ | $ and a labeling function is conStructed foR each SubForMuLa of $F$. As sEen from the algoRIthm $checkCTL$ (at the end of SectIon \[IdEaS\]), The aLGoRithm Either gIVeS A DefInItE “yeS” oR “No” AnsweR to The CTL mOdel-checkiNg pRoBlem, i.e., $\langle m,x\rAngle, s_0\modeLs f$, Or it reduces tHe pRoblem To CheCKing wHetHer thE ID expREssion $\Psi$ labeleD tO $S_0$ can be EValuateD TRuE at tHe state. The evaluatIOn PRocedure is carrIed out By THe FOLloWinG recursive PrOceduRE $TestWG$, AFtER AN inPUt sequence $\pi$ hAs been accepTEd bY the unSpEciFIed comPonenT $X$.
[**pRocEdure**]{} $TestWG(x,\pi,s_0,\Psi)$
$\psi=\psi_1\Vee \psi_2$:
$tEstWG(X,\pI,S_0,\psi_1)$ [**TheN**]{}
$true$;
$TEstwG(X,\Pi,s_0,\pSI_2)$
$\pSi=\Neg \PsI_1$:
$\Neg tEsTWG(x,\Pi,s_0, \Psi_1)$
$\psi=1$:
$trUe$;
$\PsI=i$ witH $2\le i\LE K+1$:
${\CAl I}^{-1}(i)$ Is aN $EX$ gRaph
$TEstEX(X,\pi,s_0,{\cal I}^{-1}(I))$;
${\caL I}^{-1}(i)$ iS An $Eu$ grapH
$TesteU(X,\pI,s_0,{\Cal I}^{-1}(i), Level=0)$;
${\cAl I}^{-1}(i)$ iS aN $EG$ graph
$TestEG(X,\Pi,s_0,{\cAl I}^{-1}(i)).$
In $TesTWG$, ThE fiRsT threE Cases aRe sTraIghtforWard, whiCH arE cONSIsTent with the intendeD mEANiNg of ID exPressiONs. thE Cases $TesTEx, TeStEU, tEStEG$ fOr evALuAting $EX, Eu$, $EG$ graPHs ArE discusSeD in the FoLloWinG threE SubsEctionS.
### $TestEX$
THe casE For checking wheTHer an $EX$ graph $G=\LAnGLE N,e, l_g\raNglE$ can be evaluAted TRue aT a stATe $S_0$ is SImple. we jusT tESt WHether the system $ | $ and a labeling functionis constru ctedfor ea ch sub form ula of $f$. As seen from the algorithm $C heckC TL $ (at th e end of Sec t io n \[I de as \]) ,t he algo rit hm eith er gives a de fi nite “yes” o r “ no” answer to the CTL mod el- checki ng pr o blem, i. e., $ \langl e M,X\r angle, s_ 0\ m odelsf $, or i t re duce s the problem toc he c king whether t he IDex p re s s ion $\ psi$ label ed to $ s _0$ can be e v alu a ted true at t he state. T h e e valuat io n p r ocedur e isca r rie d out by th e fo llowing r ecursi v e proce d ure $Te stWG$, af ter ani np ut se qu e nce $\ pi$ has been ac ce pt ed by the u n s peci fie d co mpone nt $X$.
[**P roc edur e **] {} $T estWG (X,\ pi ,s_0, \psi)$
$\p si =\psi_1\vee \ps i_2$ :
$TestW G(X ,\ pi, s_ 0,\ps i _1)$ [ **T hen **]{}
$true$;
$T es t W G (X ,\pi,s_0,\psi_2)$
$ \ p si =\neg \p si_1$:
$ \n e g TestWG (X ,\p i,s_ 0 , \psi _1)$
$ \psi=1$:
$tru e $;
$\psi=i $with $ 2\ lei\l e k+1 $ :
$ {\calI}^{-1}( i)$ i s an $EX$ graph
$TestEX(X,\p i ,s _ 0 ,{ \ calI}^ {-1}(i))$;
${\ c al I }^{- 1 }( i)$ is an $EU$ g r ap h
$TestEU(X,\pi,s_0 ,{ \cal I }^{-1 }(i), level=0 )$;
${\ca l I }^{-1}(i )$ i s a n $EG$ graph
$ TestE G(X,\pi,s_ 0 ,{\cal I }^{-1 }(i)).$
In $Test W G $, the f irs t t hre e c a s es are straight f o rwar d, whichare consis ten t w ith th eintendedmeaningof I Dex pre ssion s . The ca se s $ Te stE X, Te s tEU, T estEG $ fo rev a lua ting $E X ,E U $, $ EG $grap hsar e dis cuss e d i n the f ollowingthr e e su bs ec tions.
### $TestEX$
The case f or ch ecking w hether a n $EX$ graph $G=\langle N,E, L_ g\r angle $ ca n be eval uat ed tru e a t a sta te $s_ 0$ is s imp l e . Wej u st te st whether t h e sy stem$ | $ and_a labeling_function is constructed for_each subformula_of_$f$. As_seen_from the algorithm_$CheckCTL$ (at the_end of Section \[Ideas\]),_the algorithm either_gives_a definite “yes” or “no” answer to the CTL model-checking problem, i.e., $\langle M,X\rangle,_s_0\models_f$, or_it_reduces_the problem to checking whether_the ID expression $\psi$ labeled_to $s_0$_can be evaluated true at the state. The_evaluation_procedure is carried_out by the following recursive procedure $TestWG$, after an_input sequence $\pi$ has been accepted_by the unspecified_component_$X$.
[**Procedure**]{}_$TestWG(X,\pi,s_0,\psi)$
$\psi=\psi_1\vee \psi_2$:
$TestWG(X,\pi,s_0,\psi_1)$ [**Then**]{}
$true$;
$TestWG(X,\pi,s_0,\psi_2)$
$\psi=\neg \psi_1$:
$\neg_TestWG(X,\pi,s_0, \psi_1)$
$\psi=1$:
$true$;
$\psi=i$ with $2\le i\le k+1$:
${\cal_I}^{-1}(i)$ is an $EX$ graph
$TestEX(X,\pi,s_0,{\cal I}^{-1}(i))$;
${\cal_I}^{-1}(i)$ is an $EU$ graph
$TestEU(X,\pi,s_0,{\cal I}^{-1}(i), level=0)$;
${\cal_I}^{-1}(i)$ is an $EG$ graph
$TestEG(X,\pi,s_0,{\cal I}^{-1}(i)).$
In_$TestWG$, the first three cases_are straightforward,_which are consistent with the_intended meaning of_ID expressions._The cases $TestEX,_TestEU, TestEG$ for evaluating $EX, EU$,_$EG$ graphs are_discussed in the following three subsections.
###_$TestEX$
The_case for checking_whether_an_$EX$ graph_$G=\langle N,E, L_g\rangle$_can_be evaluated_true_at a state $s_0$ is simple._We_just test whether the system $ |
last part of this claim, we also have $\norm{Q_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_F^2$ as well as $\norm{T_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_F$. Finally, we state a lemma about various properties that hold in a neighborhood of the central path.
\[lemma:Properties of the central path\] Let $\nu > 0$ be arbitrary and let ${\bm{x}}, {\bm{s}} \in {\operatorname{int}}{\mathcal{L}}$. Then, ${\bm{x}}$ and ${\bm{s}}$ satisfy the following properties:
1. For all $\nu > 0$, the duality gap and distance from the central path are related as $$r\nu - \sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu) \leq {\bm{x}}^T {\bm{s}} \leq r\nu + \sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu).$$
2. The distance from the central path is symmetric in its arguments i.e. $d({\bm{x}}, {\bm{s}}, \nu) = d({\bm{s}}, {\bm{x}}, \nu)$.
3. Let $\mu = \frac1r {\bm{x}}^T {\bm{s}}$. If $d({\bm{x}}, {\bm{s}}, \mu) \leq \eta \mu$, then $(1+\eta) \norm{{\bm{s}}^{-1}}_2 \geq \norm{\mu^{-1} {\bm{x}}}_2$.
A proof of Lemma \[lemma:Properties of the central path\] is included in Appendix \[b\] for completeness. The first and the third parts of this lemma have nice intuitive interpretations: The former asserts that the duality gap is roughly equal to the central path parameter, and the latter would be trivially true if $\circ$ was associative and we had ${\bm{x}} \circ {\bm{s}} = \mu {\bm{e}}$.
A single IPM iteration {#subsec:A single IPM iteration}
----------------------
The results above allow us to introduce *commutative scalings*, a method commonly used for proving IPM convergence. Our analysis is inspired by the general case analysis from [@ben2001 | last part of this claim, we also get $ \norm{Q_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_F^2 $ equally well as $ \norm{T_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_F$. Finally, we state a lemma about diverse property that hold in a neighborhood of the cardinal path.
\[lemma: Properties of the central path\ ] permit $ \nu > 0 $ be arbitrary and let $ { \bm{x } }, { \bm{s } } \in { \operatorname{int}}{\mathcal{L}}$. Then, $ { \bm{x}}$ and $ { \bm{s}}$ satisfy the keep up properties:
1. For all $ \nu > 0 $, the dichotomy gap and distance from the central path are related as $ $ r\nu - \sqrt{\frac{r}{2 } } \cdot d({\bm{x } }, { \bm{s } }, \nu) \leq { \bm{x}}^T { \bm{s } } \leq r\nu + \sqrt{\frac{r}{2 } } \cdot d({\bm{x } }, { \bm{s } }, \nu).$$
2. The distance from the central path is symmetric in its argument i.e. $ d({\bm{x } }, { \bm{s } }, \nu) = d({\bm{s } }, { \bm{x } }, \nu)$.
3. Let $ \mu = \frac1r { \bm{x}}^T { \bm{s}}$. If $ d({\bm{x } }, { \bm{s } }, \mu) \leq \eta \mu$, then $ (1+\eta) \norm{{\bm{s}}^{-1}}_2 \geq \norm{\mu^{-1 } { \bm{x}}}_2$.
A proof of Lemma \[lemma: Properties of the cardinal path\ ] is included in Appendix \[b\ ] for completeness. The first and the third share of this lemma experience nice intuitive interpretations: The former asserts that the duality break is roughly equal to the central path parameter, and the latter would be trivially true if $ \circ$ was associative and we had $ { \bm{x } } \circ { \bm{s } } = \mu { \bm{e}}$.
A single IPM iteration { # subsec: A single IPM iteration }
----------------------
The results above allow us to introduce * commutative scalings *, a method commonly used for proving IPM overlap. Our psychoanalysis is inspired by the general case psychoanalysis from [ @ben2001 | ladt part of this claim, we also have $\norm{Q_{{\bm{x}}}}_2 \lex \norm{{\gm{x}}}_F^2$ as dell as $\norm{T_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_H$. Fibally, we state a lemma aboug various propertues uhat hold in a nemfhborhood of fme ceutcal path.
\[lemma:Prpperties ox the central [agh\] Let $\nu > 0$ be arbitrary and let ${\bm{x}}, {\fm{s}} \in {\ppfratorname{int}}{\mwthcsj{L}}$. Tgvn, ${\bm{x}}$ and ${\bm{s}}$ satisfy the follkwing pgoperties:
1. For alk $\nu > 0$, the duality gap and dishance from the cenhral path ats rqoated as $$r\nu - \sqrt{\frab{x}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu) \leq {\bm{x}}^T {\bm{s}} \leq r\nu + \sdrt{\frcc{r}{2}} \cdot d({\bn{x}}, {\bm{d}}, \nu).$$
2. The diwtancv from the ceknral padh is summetric in itx acgumwnts i.e. $d({\bm{x}}, {\bm{s}}, \nu) = v({\bm{s}}, {\bm{x}}, \nu)$.
3. Let $\mu = \frac1r {\bm{f}}^T {\bm{s}}$. If $d({\bm{x}}, {\bm{s}}, \nu) \leq \gta \mg$, thdb $(1+\ega) \hocm{{\bj{s}}^{-1}}_2 \geq \nocm{\mu^{-1} {\bm{x}}}_2$.
A pdoof of Lemna \[lemma:Properties pf nne central pzth\] is igcluded in Appendix \[b\] for completeness. Uhe fjrst and the third partw of this lemma have jice intuytive interpretations: The former asserts that the dualmth gci kw goughly equal to the central path parameter, ags uhe latter would ne trivially true of $\cotc$ was associajive anb ws had ${\bm{x}} \circ {\bm{s}} = \mu {\bm{g}}$.
A sintle IPM iueratoon {#subsec:A single IPM iterqtion}
----------------------
The resllts above allow us to introduce *eommutstive scalings*, a method communly used for pgoving IPJ convergence. Our anslfsis is inspired by the gegeral casx anakysis ffom [@nen2001 | last part of this claim, we also \leq as well $\norm{T_{{\bm{x}}}}_2 \leq \norm{{\bm{x}}}_F$. about properties that hold a neighborhood of central path. \[lemma:Properties of the central Let $\nu > 0$ be arbitrary and let ${\bm{x}}, {\bm{s}} \in {\operatorname{int}}{\mathcal{L}}$. Then, and ${\bm{s}}$ satisfy the following properties: 1. For all $\nu > 0$, the gap distance the path are related as $$r\nu - \sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu) \leq {\bm{x}}^T {\bm{s}} \leq r\nu \sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu).$$ 2. The distance the central path is in its arguments i.e. $d({\bm{x}}, \nu) d({\bm{s}}, {\bm{x}}, 3. $\mu \frac1r {\bm{x}}^T {\bm{s}}$. $d({\bm{x}}, {\bm{s}}, \mu) \leq \eta \mu$, then $(1+\eta) \norm{{\bm{s}}^{-1}}_2 \geq \norm{\mu^{-1} {\bm{x}}}_2$. A proof of Lemma \[lemma:Properties the central included in \[b\] completeness. first and the of this lemma have nice intuitive asserts that the duality gap is roughly equal the central parameter, and the latter would be true if $\circ$ was associative and we had \circ {\bm{s}} = \mu {\bm{e}}$. A single IPM iteration {#subsec:A single IPM iteration} ---------------------- The allow us to introduce scalings*, a method used proving convergence. analysis is by the general case analysis from [@ben2001 | last part of this claim, we also Have $\norm{Q_{{\bM{x}}}}_2 \leq \NorM{{\bm{X}}}_F^2$ As weLl as $\Norm{T_{{\bm{x}}}}_2 \leq \norM{{\Bm{x}}}_F$. finally, we state a lemma abOut vaRiOUs prOPeRties That holD In A NEigHbOrHooD oF ThE centRal Path.
\[lemMa:PropertiEs oF tHe central patH\] leT $\nu > 0$ be arbitRarY and let ${\bm{x}}, {\bm{S}} \in {\OperatOrNamE{Int}}{\maThcAl{L}}$. ThEn, ${\bm{x}}$ aND ${\bm{s}}$ saTisfy the fOlLOwing pROpertieS:
1. fOr All $\nU > 0$, the duality gap and DIsTAnce from the cenTral paTh ARe RELatEd aS $$r\nu - \sqrt{\frAc{R}{2}} \cdot D({\Bm{x}}, {\bm{s}}, \nU) \LeQ {\BM{X}}^T {\bM{S}} \leq r\nu + \sqrt{\frAc{r}{2}} \cdot d({\bm{x}}, {\BM{s}}, \nU).$$
2. The diStAncE From thE centRaL PatH is symmetriC in iTs argumenTs i.e. $d({\bM{X}}, {\bm{s}}, \nu) = d({\BM{s}}, {\bm{x}}, \nu)$.
3. let $\mu = \fRac1R {\bm{X}}^T {\bm{S}}$. if $D({\bM{x}}, {\bM{s}}, \MU) \leQ \EtA \mu$, THen $(1+\Eta) \norm{{\bM{s}}^{-1}}_2 \GeQ \norm{\Mu^{-1} {\bm{X}}}_2$.
a PROof oF LeMma \[lEmma:PRoperties of thE ceNtraL PatH\] is inCludeD in APpEndix \[B\] for coMpletEnEss. The first and tHe thIrd parts oF thIs LemMa Have nICe intuItiVe iNterpreTations: tHe fOrMER AsSerts that the dualitY gAP Is Roughly eQual to THe CeNTral path PaRamEter, AND the lAtteR WoUld be triVially TRuE iF $\circ$ waS aSsociaTiVe aNd wE had ${\bM{X}} \cirC {\bm{s}} = \mu {\Bm{e}}$.
A singLe IPM ITeration {#subsec:a Single IPM iterATiON}
----------------------
thE ResuLts Above allow uS to iNTrodUce *cOMmUtaTIve scAlingS*, a MEtHOd commonly used for prOvIng IPM ConveRgence. Our analYsis is inspIRED by the geNeraL CaSE analysis from [@bEn2001 | last part of this claim,we also ha ve $\ nor m{Q _{ {\bm {x}} }}_2 \leq \nor m {{\b m{x}}}_F^2$ as well as $\no rm { T_{{ \ bm {x}}} }_2 \le q \ n o rm{ {\ bm {x} }} _ F$ . Fin all y, we s tate a lem maab out variousp ro perties th athold in a ne igh borhoo doft he ce ntr al pa th.
\ [ lemma: Propertie so f thec entralp a th \] L et $\nu > 0$ be a r bi t rary and let $ {\bm{x }} , { \ b m{s }}\in {\oper at ornam e {int}}{ \ ma t h c al{ L }}$. Then, ${ \bm{x}}$ an d ${ \bm{s} }$ sa t isfy t he fo ll o win g propertie s:
1. For a ll $\n u > 0$,t he dual ity ga p a nddist a nc efro mt hec en tra l pa th are r el at ed as $$r \ n u - \ sqr t{\f rac{r }{2}} \cdot d ({\ bm{x } },{\bm{ s}},\nu) \ leq { \bm{x} }^T { \b m{s}} \leq r\nu + \ sqrt{\fra c{r }{ 2}} \ cdotd ({\bm{ x}} , { \bm{s}} , \nu). $ $
2. T he distance from the c e n tr al pathis sym m et ri c in itsar gum ents i .e. $ d({\ b m{ x}}, {\b m{s}}, \n u) = d({\ bm {s}},{\ bm{ x}} , \nu ) $.
3. Le t $\mu = \fra c 1r {\bm{x}}^T{ \bm{s}}$. If$ d( { \ bm { x}}, {\ bm{s}}, \mu ) \l e q \e ta \ m u$ , t h en $( 1+\et a) \n o rm{{\bm{s}}^{-1}}_2 \ geq \n orm{\ mu^{-1} {\bm{ x}}}_2$.
A p roof ofLemm a \ [ lemma:Properti es of the centr a l path\] is i ncludedin Append i x \[b\] f orcom ple ten e s s. The first an d theth ird par tsof this le mma ha veni ce intuit ive inte rp re ta ti ons : The former a ss ert stha t the dualit y gap isro ug h lyequal t o t h e cen tr al pat h p ar amete r, a n d t he latt er wouldbet rivi al ly true i f $\circ$ was a ssociative a ndwe had $ {\bm{x}} \circ {\bm{s}} = \mu { \ bm{e}}$ .
A sin gleIPM itera tio n {#su bse c :A sin gle IP M ite ra tio n }
---- - - -- --- -- ---------T heresul ts abo ve allo w us to introduce* com mutative scal ing s*,a me tho d c o mmo nl y us e d for proving IP M converge nc e .Our analys i s i sinspire d by th e gen e ral cas e analysi s from [@ be n200 1 | last_part of_this claim, we also_have $\norm{Q_{{\bm{x}}}}_2_\leq_\norm{{\bm{x}}}_F^2$ as_well_as $\norm{T_{{\bm{x}}}}_2 \leq_\norm{{\bm{x}}}_F$. Finally, we_state a lemma about_various properties that_hold_in a neighborhood of the central path.
\[lemma:Properties of the central path\] Let $\nu >_0$_be arbitrary_and_let_${\bm{x}}, {\bm{s}} \in {\operatorname{int}}{\mathcal{L}}$. Then,_${\bm{x}}$ and ${\bm{s}}$ satisfy the_following properties:
1._ For all $\nu > 0$, the duality_gap_and distance from_the central path are related as $$r\nu - _\sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}}, \nu) \leq_{\bm{x}}^T {\bm{s}} \leq_r\nu_+_\sqrt{\frac{r}{2}} \cdot d({\bm{x}}, {\bm{s}},_\nu).$$
2. The distance from the_central path is symmetric in its_arguments i.e. $d({\bm{x}}, {\bm{s}}, \nu) = d({\bm{s}},_{\bm{x}}, \nu)$.
3. Let $\mu =_\frac1r {\bm{x}}^T {\bm{s}}$. If $d({\bm{x}},_{\bm{s}}, \mu)_\leq \eta \mu$, then $(1+\eta)_\norm{{\bm{s}}^{-1}}_2 \geq \norm{\mu^{-1}_{\bm{x}}}_2$.
A proof_of Lemma \[lemma:Properties_of the central path\] is included_in Appendix \[b\]_for completeness. The first and the_third_parts of this_lemma_have_nice intuitive_interpretations: The former_asserts_that the_duality_gap is roughly equal to the_central_path parameter, and the latter would be_trivially true if $\circ$_was_associative and we had_${\bm{x}} \circ {\bm{s}} = \mu_{\bm{e}}$.
A single IPM iteration {#subsec:A single_IPM iteration}
----------------------
The_results above_allow us to introduce *commutative scalings*, a method commonly used for_proving IPM convergence. Our analysis is_inspired by the general_case analysis_from_[@ben2001 |
: \pi \rightarrow \pi'$, induces a ring homomorphism $\varphi_{\sharp}: \Lambda \rightarrow \Lambda'$, where $\Lambda' = \mathbb Z[\pi']$. A chain map is a pair $(\varphi, F): (\pi, C) \rightarrow (\pi', C')$, where $\varphi$ is a group homomorphism and $F: C \rightarrow C'$ a $\varphi$–equivariant chain map, that is a chain map of the underlying abelian chain complexes, such that $F(\lambda c) = \varphi_{\sharp}(\lambda) F(c)$ for $\lambda \in \Lambda$ and $c \in C$. Two such chain maps are homotopic, $(\varphi, F) \simeq (\psi, G)$ if $\varphi = \psi$ and if there is a $\varphi$–equivariant map $\alpha: C \rightarrow C'$ of degree $+1$ such that $G - F = d \alpha + \alpha d$.
A pair $(\pi, C)$ is a reduced chain complex if $C_0 = \Lambda$ with generator $\ast$, $C_i = 0$ for $i < 0$ and ${\rm H}_0 C = \mathbb Z$ such that $C_0 = \Lambda \rightarrow {\rm H}_0 C = \mathbb Z$ is the augmentation of $\Lambda$. A chain map, $(\varphi, f): (\pi, C) \rightarrow (\pi', C')$, of reduced chain complexes, is reduced if $f_0$ is induced by $\varphi_{\sharp}$, and a chain homotopy $\alpha$ of reduced chain maps is reduced if $\alpha_0 = 0$. The objects of the category ${\rm\bf H}_0$ are reduced chain complexes and the morphisms are reduced chain maps. Homotopies in ${\rm\bf H}_0$ are reduced chain homotopies. Every chain complex $(\pi, C)$ in ${\rm\bf H}_0$ is equipped with the augmentation $\varepsilon: C \rightarrow \mathbb Z$ in ${\rm\bf H}_0$. The ring homomorphism $\mathbb Z \rightarrow \Lambda$ yields the co–augmentation $\iota: \mathbb Z \rightarrow C$, where we view $\mathbb Z = (0, \mathbb Z)$ as chain complex with trivial group $\pi = 0$ concentrated in degree $0$. Note that $\varepsilon \iota = {\rm id}_{\mathbb Z}$, and the composite $\iota \varepsilon: | : \pi \rightarrow \pi'$, induces a ring homomorphism $ \varphi_{\sharp }: \Lambda \rightarrow \Lambda'$, where $ \Lambda' = \mathbb Z[\pi']$. A chain map is a couple $ (\varphi, F ): (\pi, C) \rightarrow (\pi', C')$, where $ \varphi$ is a group homomorphism and $ farad: C \rightarrow C'$ a $ \varphi$–equivariant chain map, that is a range map of the underlie abelian chain complexes, such that $ F(\lambda cytosine) = \varphi_{\sharp}(\lambda) F(c)$ for $ \lambda \in \Lambda$ and $ c \in C$. Two such chain map are homotopic, $ (\varphi, F) \simeq (\psi, G)$ if $ \varphi = \psi$ and if there is a $ \varphi$–equivariant map $ \alpha: C \rightarrow C'$ of academic degree $ +1 $ such that $ G - F = d \alpha + \alpha d$.
A pair $ (\pi, C)$ is a reduced chain complex if $ C_0 = \Lambda$ with generator $ \ast$, $ C_i = 0 $ for $ i < 0 $ and $ { \rm H}_0 C = \mathbb Z$ such that $ C_0 = \Lambda \rightarrow { \rm H}_0 C = \mathbb Z$ is the augmentation of $ \Lambda$. A range map, $ (\varphi, f ): (\pi, C) \rightarrow (\pi', C')$, of reduced range complexes, is shrink if $ f_0 $ is induce by $ \varphi_{\sharp}$, and a chain homotopy $ \alpha$ of reduced chain function is reduced if $ \alpha_0 = 0$. The objects of the category $ { \rm\bf H}_0 $ are reduced range complexes and the morphisms are reduce chain maps. Homotopies in $ { \rm\bf H}_0 $ are reduced range homotopies. Every chain complex $ (\pi, C)$ in $ { \rm\bf H}_0 $ is equipped with the augmentation $ \varepsilon: C \rightarrow \mathbb Z$ in $ { \rm\bf H}_0$. The ring homomorphism $ \mathbb Z \rightarrow \Lambda$ yields the co – augmentation $ \iota: \mathbb Z \rightarrow C$, where we see $ \mathbb Z = (0, \mathbb Z)$ as chain complex with trivial group $ \pi = 0 $ concentrated in degree $ 0$. Note that $ \varepsilon \iota = { \rm id}_{\mathbb Z}$, and the composite $ \iota \varepsilon: | : \pi \rightarrow \pi'$, induces a ring homomorphnwm $\var'hi_{\sharl}: \Lambda \rightarrow \Lambda'$, where $\Lamuda' = \mathvb Z[\pi']$. A chain map is x pair $(\vagphi, F): (\pi, C) \rmghtarrow (\pi', C')$, wisre $\variki$ is w grmnp homomorphism and $F: C \rhghtarrow C'$ a $\eafpki$–equivariant chain map, that is a chwin map ov the underlyigg anqliah chain complexes, such that $F(\lambdz c) = \vagphi_{\sharp}(\lambda) F(v)$ for $\lambda \in \Lambda$ and $c \ij C$. Two such chain maps are himotjpic, $(\varphi, F) \simeq (\psi, G)$ if $\varphi = \psi$ and if there is a $\varphi$–eduivaxiant map $\aopya: F \rightarrow C'$ of degree $+1$ sucm that $C - F = d \alpha + \alpha c$.
A 'air $(\pi, C)$ is a reduced chein complex if $C_0 = \Laibda$ with gznerator $\ast$, $C_i = 0$ for $i < 0$ atd ${\rk H}_0 Z = \mxthgb Z$ such hhav $C_0 = \Lambda \rightarrow {\rm H}_0 C = \mathbb Z$ ix eye augmentatikn of $\Jaibda$. A chain map, $(\varphi, f): (\pi, C) \rightarrmw (\li', C')$, of reduced chain cimplexes, is reduced iv $f_0$ is inquced by $\varphi_{\sharp}$, and a chain homotopy $\alpha$ ox redncdd emwkb laps is reduced if $\alpha_0 = 0$. The objects of the dauegpry ${\rm\bf H}_0$ are reduced chsij vjmplexes and jhe mor'giams are reduced chwin mapf. Homitopies ig ${\rm\nf H}_0$ are reduced chain homoropies. Every xhain complex $(\pi, C)$ in ${\rm\bf H}_0$ ns equopped with the augmentation $\rarepsjlon: C \righharrow \maffbb Z$ in ${\rm\bf H}_0$. Ghe rhng homomorphism $\mathbb Z \wightarrox \Lamyda$ yielas tne co–atgmentatioj $\iota: \mathbb Z \rightarrlw C$, chere we view $\mwthbb Z = (0, \mathbb Z)$ as chain com'kex with trifidl croup $\pi = 0$ cokcentrated in dqgree $0$. Note thct $\varepfilon \iota = {\rm pd}_{\mathbb V}$, and the coiposite $\iota \ewrepsilon: | : \pi \rightarrow \pi'$, induces a ring \Lambda \Lambda'$, where = \mathbb Z[\pi']$. pair F): (\pi, C) (\pi', C')$, where is a group homomorphism and $F: \rightarrow C'$ a $\varphi$–equivariant chain map, that is a chain map of the abelian chain complexes, such that $F(\lambda c) = \varphi_{\sharp}(\lambda) F(c)$ for $\lambda \in and \in Two chain maps are homotopic, $(\varphi, F) \simeq (\psi, G)$ if $\varphi = \psi$ and if there a $\varphi$–equivariant map $\alpha: C \rightarrow C'$ of $+1$ such that $G F = d \alpha + d$. pair $(\pi, is reduced complex if $C_0 \Lambda$ with generator $\ast$, $C_i = 0$ for $i < 0$ and ${\rm H}_0 C = \mathbb such that \Lambda \rightarrow H}_0 = Z$ is the $\Lambda$. A chain map, $(\varphi, f): (\pi', C')$, of reduced chain complexes, is reduced $f_0$ is by $\varphi_{\sharp}$, and a chain homotopy of reduced chain maps is reduced if $\alpha_0 0$. The objects of the category ${\rm\bf H}_0$ are reduced chain complexes and the morphisms chain maps. Homotopies in H}_0$ are reduced homotopies. chain $(\pi, in ${\rm\bf is equipped with the augmentation $\varepsilon: C \rightarrow \mathbb Z$ in H}_0$. The ring homomorphism $\mathbb Z \rightarrow \Lambda$ yields the \mathbb \rightarrow C$, where view $\mathbb Z = \mathbb as chain complex with $\pi 0$ $0$. that \iota = {\rm id}_{\mathbb and the composite $\iota \varepsilon: | : \pi \rightarrow \pi'$, induces a rinG homomorphIsm $\vaRphI_{\shArP}: \LamBda \rIghtarrow \LambdA'$, WherE $\Lambda' = \mathbb Z[\pi']$. A chain Map is A pAIr $(\vaRPhI, F): (\pi, C) \RightarROw (\PI', c')$, whErE $\vArpHi$ IS a Group HomOmorphiSm and $F: C \rigHtaRrOw C'$ a $\varphi$–eqUIvAriant chaiN maP, that is a chaiN maP of the UnDerLYing aBelIan chAin comPLexes, sUch that $F(\lAmBDa c) = \varPHi_{\sharp}(\LAMbDa) F(c)$ For $\lambda \in \Lambda$ ANd $C \In C$. Two such chaiN maps aRe HOmOTOpiC, $(\vaRphi, F) \simeq (\PsI, G)$ if $\vARphi = \psi$ ANd IF THerE Is a $\varphi$–equiVariant map $\aLPha: c \rightArRow c'$ Of degrEe $+1$ sucH tHAt $G - f = d \alpha + \alphA d$.
A pAir $(\pi, C)$ is a ReduceD Chain coMPlex if $C_0 = \lambda$ WitH geNeraTOr $\AsT$, $C_i = 0$ FoR $I < 0$ anD ${\Rm h}_0 C = \mAThbB Z$ such thAt $c_0 = \LAmbda \RighTARROw {\rm h}_0 C = \mAthbB Z$ is tHe augmentatioN of $\lambDA$. A cHain mAp, $(\varPhi, f): (\Pi, c) \righTarrow (\Pi', C')$, of ReDuced chain complExes, Is reduced If $f_0$ Is IndUcEd by $\vARphi_{\shArp}$, And A chain hOmotopy $\ALphA$ oF REDuCed chain maps is reduCeD IF $\aLpha_0 = 0$. The oBjects OF tHe CAtegory ${\rM\bF H}_0$ aRe reDUCed chAin cOMpLexes and The morPHiSmS are redUcEd chaiN mAps. homOtopiES in ${\rM\bf H}_0$ arE reduced Chain HOmotopies. Every CHain complex $(\pi, c)$ In ${\RM\Bf h}_0$ Is eqUipPed with the aUgmeNTatiOn $\vaREpSilON: C \rigHtarrOw \MAtHBb Z$ in ${\rm\bf H}_0$. The ring hoMoMorphiSm $\matHbb Z \rightarroW \Lambda$ yieLDS The co–augMentATiON $\iota: \mathbb Z \riGhtarRow C$, where wE View $\mathBb Z = (0, \maThbb Z)$ as cHain complEX With trivIal GroUp $\pI = 0$ coNCEnTrated in degreE $0$. nOte tHaT $\varepsIloN \iota = {\rm Id}_{\mAthBb Z}$, And ThE compositE $\iota \varEpSiLoN: | : \pi \rightarrow \pi'$, i nduces a r ing h omo mor ph ism$\va rphi_{\sharp}: \Lam bda \rightarrow \Lambd a'$,wh e re $ \ La mbda' = \mat h bb Z [\p i' ]$ . A c h ai n map is a pair $(\varphi , F ): (\pi, C) \r i gh tarrow (\p i', C')$, where $\ varphi $isa grou p h omomo rphism and $F : C \righ ta r row C' $ a $\va r p hi $–eq uivariant chain m a p, that is a chai n mapof th e und erl ying abeli an chai n comple x es , s uch that $F(\lamb da c) = \va r phi _{\sha rp }(\ l ambda) F(c) $f or$\lambda \i n \L ambda$ an d $c \ i n C$. T w o suchchainmap s a re h o mo to pic ,$ (\v a rp hi, F)\simeq ( \p si , G)$ if$ \ v a rphi =\psi $ and if there isa $ \var p hi$ –equi varia nt m ap $\al pha: C \rig ht arrow C'$ of de gree $+1$ suc h t ha t $ G- F = d \alp ha+ \ alpha d $.
A p a ir$( \ p i ,C)$ is a reduced c ha i n c omplex i f $C_0 =\L a mbda$ wi th ge nera t o r $\a st$, $C _i = 0$for $i < 0 $ and $ {\ rm H}_ 0C = \m athbb Z$ s uch th at $C_0= \La m bda \rightarro w {\rm H}_0 C= \ m a th b b Z$ is the augmen tati o n of $\L a mb da$ . A ch ain m ap , $ ( \varphi, f): (\pi,C) \righ tarro w (\pi', C')$ , of reduc e d chain co mple x es , is reduced if $f_0 $ is induc e d by $\v arphi _{\sharp }$, and a c hain hom oto py$\a lph a $ o f reduced cha i n map sis redu ced if $\a lph a_0 =0$. T he object s of the c at eg or y $ {\rm\ b f H}_0$ar e r ed uce d cha i n comp lexes and t he mor phismsa re r educ ed c hain ma ps . Hom otop i esin ${\r m\bf H}_0 $ a r e re du ce d chain homotopies.Ev ery chainco mpl ex $(\ p i , C)$ in ${\rm\bf H}_0$ is equi p ped wit h t he au gmen tation $\ var epsilo n:C \righ tarrow \mat hb b Z $ in ${ \ r m\ bfH} _0$. The r i n g h omomo rp hism $\math bb Z \rightarrow \ L amb da$ yields th e c o–au g m en tat i on $\i ot a : \ m a thbb Z \rightar row C$, wh er e w e view $\m a thb bZ = (0, \mathb b Z)$ as chai n complex with tri vi al g r o up$\pi = 0$concentr ated in d e gree$ 0$ . Not e t hat $\ va rep silon \iota = { \rm i d}_{\m at hbb Z} $, an dthe comp osite $\iota \varepsilo n: | : \pi_\rightarrow \pi'$,_induces a ring homomorphism_$\varphi_{\sharp}: \Lambda_\rightarrow_\Lambda'$, where_$\Lambda'_= \mathbb Z[\pi']$._A chain map_is a pair $(\varphi,_F): (\pi, C)_\rightarrow_(\pi', C')$, where $\varphi$ is a group homomorphism and $F: C \rightarrow C'$ a_$\varphi$–equivariant_chain map,_that_is_a chain map of the_underlying abelian chain complexes, such_that $F(\lambda_c) = \varphi_{\sharp}(\lambda) F(c)$ for $\lambda \in \Lambda$_and_$c \in C$._Two such chain maps are homotopic, $(\varphi, F) \simeq_(\psi, G)$ if $\varphi = \psi$_and if there_is_a_$\varphi$–equivariant map $\alpha: C_\rightarrow C'$ of degree $+1$ such_that $G - F = d_\alpha + \alpha d$.
A pair $(\pi, C)$_is a reduced chain complex if_$C_0 = \Lambda$ with generator_$\ast$, $C_i_= 0$ for $i _< 0$ and_${\rm H}_0_C = \mathbb_Z$ such that $C_0 = \Lambda_\rightarrow {\rm H}_0_C = \mathbb Z$ is the_augmentation_of $\Lambda$. A_chain_map,_$(\varphi, f):_(\pi, C) \rightarrow_(\pi',_C')$, of_reduced_chain complexes, is reduced if $f_0$_is_induced by $\varphi_{\sharp}$, and a chain homotopy_$\alpha$ of reduced chain_maps_is reduced if $\alpha_0_= 0$. The objects of_the category ${\rm\bf H}_0$ are reduced_chain complexes_and the_morphisms are reduced chain maps. Homotopies in ${\rm\bf H}_0$ are reduced_chain homotopies. Every chain complex $(\pi,_C)$ in ${\rm\bf H}_0$_is equipped_with_the augmentation $\varepsilon:_C_\rightarrow \mathbb_Z$ in ${\rm\bf H}_0$. The ring homomorphism_$\mathbb Z_\rightarrow \Lambda$ yields the co–augmentation $\iota:_\mathbb Z \rightarrow C$,_where_we view $\mathbb Z = (0,_\mathbb Z)$ as chain complex with_trivial group $\pi = 0$_concentrated_in_degree $0$. Note that $\varepsilon_\iota = {\rm id}_{\mathbb Z}$, and_the composite $\iota_\varepsilon: |
bigotimes_{i=1}^N
x_i\big)=\bigotimes_{i=1}^N x_{i+1}$ where sites $N+1$ and $1$ are identified. The interaction is called $L$-*local* if $h$ acts non-trivially only on $L$ neighboring sites, and it is said to be *frustration free* with respect to its ground state $\phi_0$ if the latter minimizes the energy locally in the sense that $\langle \phi_0|H|\phi_0\rangle=\inf_\phi \langle
\phi|H|\phi\rangle=N \inf_\phi \langle \phi|h|\phi\rangle$. As proven in [@Hast1] all gapped Hamiltonians can be approximated by frustration free ones if one allows for enlarging the interaction range $L$ up to ${\cal O}(\log N)$.
For every MPS and FCS $\psi$ one can easily find frustration free Hamiltonians such that $\psi$ is their exact ground state. Moreover, these *parent* Hamiltonains are $L$-local with $L\sim 2 \log D/\log d$ and they allow for a detailed analysis of the ground state degeneracy (Sec.\[sec:Huniqueness\]) and the energy gap above the ground state (Sec.\[Sec:Hgap\]). Typically, these Hamiltonians are, however, not exactly solvable, i.e., information about the excitations might be hard to obtain.
Examples
--------
1. [*AKLT*]{}: The father of all matrix product states is the ground state of the AKLT-Hamiltonian $$\label{HAKLT}
H=\sum_i
\vec{S}_i\vec{S}_{i+1}+\frac13\Big(\vec{S}_i\vec{S}_{i+1}\Big)^2\;,$$ where $\vec{S}$ is the vector of spin-1 operators (i.e., d=3). Its MPS representation is given by $\{A_i\}=\big\{\sigma^z,\sqrt{2}\sigma^+,-\sqrt{2}\sigma^-\big\}$ where the $\sigma$’s are the Pauli matrices.
2. [*Majumdar-Gosh*]{}: The Hamiltonian $$\label{HMG}
H=\sum_i 2\vec{\sigma}_i\vec\ | bigotimes_{i=1}^N
x_i\big)=\bigotimes_{i=1}^N x_{i+1}$ where sites $ N+1 $ and $ 1 $ are identified. The interaction is called $ L$-*local * if $ h$ acts non - trivially merely on $ L$ neighboring web site, and it is said to be * frustration free * with respect to its reason state $ \phi_0 $ if the latter minimizes the energy locally in the common sense that $ \langle \phi_0|H|\phi_0\rangle=\inf_\phi \langle
\phi|H|\phi\rangle = N \inf_\phi \langle \phi|h|\phi\rangle$. As proven in [ @Hast1 ] all gapped Hamiltonians can be approximated by frustration complimentary ones if one allows for elaborate the interaction range $ L$ up to $ { \cal O}(\log N)$.
For every MPS and FCS $ \psi$ one can easily witness frustration free Hamiltonians such that $ \psi$ is their exact ground state. Moreover, these * parent * Hamiltonains are $ L$-local with $ L\sim 2 \log D/\log d$ and they allow for a detailed psychoanalysis of the ground state degeneracy (Sec.\[sec: Huniqueness\ ]) and the department of energy gap above the ground state (Sec.\[Sec: Hgap\ ]). Typically, these Hamiltonians are, however, not exactly solvable, i.e., information about the excitement might be hard to obtain.
Examples
--------
1. [ * AKLT * ] { }: The father of all matrix product country is the ground state of the AKLT - Hamiltonian $ $ \label{HAKLT }
H=\sum_i
\vec{S}_i\vec{S}_{i+1}+\frac13\Big(\vec{S}_i\vec{S}_{i+1}\Big)^2\;,$$ where $ \vec{S}$ is the vector of spin-1 operators (i.e., d=3). Its MPS representation is given by $ \{A_i\}=\big\{\sigma^z,\sqrt{2}\sigma^+,-\sqrt{2}\sigma^-\big\}$ where the $ \sigma$ ’s are the Pauli matrices.
2. [ * Majumdar - Gosh * ] { }: The Hamiltonian $ $ \label{HMG }
H=\sum_i 2\vec{\sigma}_i\vec\ | bigltimes_{i=1}^N
x_i\big)=\bigotimes_{i=1}^N x_{i+1}$ where sites $N+1$ and $1$ are isentifiea. The interaction is called $P$-*lical* uf $h$ acts non-trivially only on $P$ neighbirinj sites, and it is said to be *fdmstraciin free* with rgspect to itv ground state $\pfi_0$ if the latter minimizes the energy locallu ln the sense trat $\kwngls \phi_0|H|\phi_0\rangle=\inf_\phi \langle
\phi|H|\phj\rangle=I \inf_\phi \langle \phi|h|\phi\rangle$. As proven ij [@Hadt1] all gapped Hamiptonians cab be qpproximated by frustration free ohes if one allows for enlarging the nnteraction rqngf $L$ up to ${\cao O}(\lod N)$.
For every MPS and FCS $\pso$ one can easiky hind frustration free Hammltonians such that $\ksi$ is thehr exact ground stare. Moremver, therw *pxreht* Hzmiltojaiis are $L$-loczl with $L\sin 2 \log D/\log d$ and tner allow for a setailqd analysis of the ground state degeneracj (Sed.\[sec:Huniqueness\]) and the energy gap above the ground seate (Sec.\[Sec:Hgap\]). Typically, these Hamiltonians are, vowevxr, nou cwactut dolvable, i.e., information about the excitations jibhn be hard to obtaln.
Examples
--------
1. [*AKLT*]{}: Tne fsjher of all majrix produdt states is the ggound sjate od the AKLU-Hamiktonian $$\label{HAKLT}
H=\wum_i
\nec{S}_u\vec{S}_{i+1}+\frac13\Big(\vec{S}_i\rec{S}_{i+1}\Big)^2\;,$$ whexe $\vec{X}$ is yhe vector of spin-1 operctors (j.e., d=3). Its MPD represehgation is given cy $\{S_i\}=\tig\{\sigma^e,\rqrt{2}\sigma^+,-\sqrt{2}\sigmw^-\big\}$ wherx the $\sigma$’s are the Pwuli matrifes.
2. [*Majumdar-Gosh*]{}: The Haliltouian $$\nabel{HMG}
H=\sum_i 2\vec{\sigma}_i\vec\ | bigotimes_{i=1}^N x_i\big)=\bigotimes_{i=1}^N x_{i+1}$ where sites $N+1$ and identified. interaction is $L$-*local* if $h$ neighboring and it is to be *frustration with respect to its ground state if the latter minimizes the energy locally in the sense that $\langle \phi_0|H|\phi_0\rangle=\inf_\phi \phi|H|\phi\rangle=N \inf_\phi \langle \phi|h|\phi\rangle$. As proven in [@Hast1] all gapped Hamiltonians can be by free if allows for enlarging the interaction range $L$ up to ${\cal O}(\log N)$. For every MPS and $\psi$ one can easily find frustration free Hamiltonians that $\psi$ is their ground state. Moreover, these *parent* are with $L\sim \log d$ they allow for detailed analysis of the ground state degeneracy (Sec.\[sec:Huniqueness\]) and the energy gap above the ground state (Sec.\[Sec:Hgap\]). these Hamiltonians not exactly i.e., about excitations might be obtain. Examples -------- 1. [*AKLT*]{}: The matrix product states is the ground state of AKLT-Hamiltonian $$\label{HAKLT} \vec{S}_i\vec{S}_{i+1}+\frac13\Big(\vec{S}_i\vec{S}_{i+1}\Big)^2\;,$$ where $\vec{S}$ is the vector spin-1 operators (i.e., d=3). Its MPS representation is by $\{A_i\}=\big\{\sigma^z,\sqrt{2}\sigma^+,-\sqrt{2}\sigma^-\big\}$ where the $\sigma$’s are the Pauli matrices. 2. [*Majumdar-Gosh*]{}: The Hamiltonian $$\label{HMG} H=\sum_i | bigotimes_{i=1}^N
x_i\big)=\bigotimes_{I=1}^N x_{i+1}$ where sItes $N+1$ And $1$ Are IdEntiFied. the interaction IS calLed $L$-*local* if $h$ acts non-triViallY oNLy on $l$ NeIghboRing sitES, aND It iS sAiD to Be *FRuStratIon Free* witH respect to Its GrOund state $\phi_0$ IF tHe latter miNimIzes the energY loCally iN tHe sENse thAt $\lAngle \Phi_0|H|\phI_0\Rangle=\Inf_\phi \lanGlE
\Phi|H|\phI\Rangle=N \INF_\pHi \laNgle \phi|h|\phi\rangle$. aS pROven in [@Hast1] all gApped HAmILtONIanS caN be approxiMaTed by FRustratIOn FREE onES if one allows fOr enlarging THe iNteracTiOn rANge $L$ up To ${\cal o}(\lOG N)$.
FOr every MPS aNd FCs $\psi$ one caN easilY Find fruSTration Free HaMilTonIans SUcH tHat $\PsI$ Is tHEiR exACt gRound staTe. moReoveR, theSE *PARent* hamIltoNains Are $L$-local with $l\siM 2 \log d/\Log D$ and tHey alLow fOr A detaIled anAlysiS oF the ground state DegeNeracy (Sec.\[Sec:huNiqUeNess\]) aND the enErgY gaP above tHe grounD StaTe (sEC.\[seC:Hgap\]). Typically, thesE HAMIlTonians aRe, howeVEr, NoT Exactly sOlVabLe, i.e., INFormaTion ABoUt the excItatioNS mIgHt be harD tO obtaiN.
EXamPleS
--------
1. [*AKLT*]{}: tHe faTher of All matriX prodUCt states is the gROund state of thE aKlt-haMIltoNiaN $$\label{HAKLT}
h=\sum_I
\Vec{S}_I\vec{s}_{I+1}+\fRac13\bIg(\vec{s}_i\vec{s}_{i+1}\bIg)^2\;,$$ WHere $\vec{S}$ is the vector Of Spin-1 opEratoRs (i.e., d=3). Its MPS rePresentatiON IS given by $\{a_i\}=\biG\{\SiGMa^z,\sqrt{2}\sigma^+,-\sqRt{2}\sigMa^-\big\}$ where THe $\sigma$’s Are thE Pauli maTrices.
2. [*MajUMDar-Gosh*]{}: THe HAmiLtoNiaN $$\LAbEl{HMG}
H=\sum_i 2\vec{\SIGma}_i\VeC\ | bigotimes_{i=1}^N
x_i\big) =\bigotime s_{i= 1}^ N x _{ i+1} $ wh ere sites $N+1 $ and $1$ are identified. T he in te r acti o nis ca lled $L $ -* l o cal *if $h $a ct s non -tr ivially only on $ L$ne ighboring si t es , and it i s s aid to be *f rus tratio nfre e * wit h r espec t to i t s grou nd state$\ p hi_0$i f the l a t te r mi nimizes the energ y l o cally in the s ense t ha t $ \ l ang le\phi_0|H|\ ph i_0\r a ngle=\i n f_ \ p h i \ l angle
\phi|H| \phi\rangle = N \ inf_\p hi \l a ngle \ phi|h |\ p hi\ rangle$. As pro ven in [@ Hast1] all gap p ed Hami ltonia nscan bea pp ro xim at e d b y f rus t rat ion free o ne s ifonea l l o ws f orenla rging the interact ion ran g e $ L$ up to $ {\ca lO}(\l og N)$ .
Fo revery MPS and F CS $ \psi$ one ca neas il y fin d frust rat ion free H amilton i ans s u c h t hat $\psi$ is thei re x ac t ground state . M or e over, th es e * pare n t * Ham ilto n ai ns are $ L$-loc a lwi th $L\s im 2 \lo gD/\ log d$ a n d th ey all ow for a deta i led analysis o f the ground s t at e de g ener acy (Sec.\[sec :Hun i quen ess\ ] )and the e nergy g a pa bove the ground sta te (Sec. \[Sec :Hgap\]). Typ ically, th e s e Hamilto nian s a r e, however, no t exa ctly solva b le, i.e. , inf ormation about th e excitati ons mi ght be h ar d to obtain.E xamp le s
----- ---
1. [ *AK LT* ]{} : T he father o f all ma tr ix p ro duc t sta t es is th egro un d s tateo f theAKLT- Hami lt on i an$$\labe l {H A K LT} H=\ su m_i
\ vec{S}_ i\vec{S}_ {i+ 1 }+\f ra c1 3\Big(\ vec{S}_i\vec{ S} _{i+1}\Big )^ 2\; ,$$ wh e r e $\vec{ S}$ is the vector of sp i n-1 ope rat ors ( i.e. , d=3). I tsMPS re pre s entati on isgiven b y $ \ { A_i\} = \ bi g\{ \s igma^z,\sq r t {2} \sigm a^ +,-\ sqrt{2} \sigma^-\big\}$ wh e rethe $\sigma$’ s a re t h e P aul i m a tri ce s .
2 . [*Majumdar-Go sh*]{}: Th eH am iltonian $ $ \la be l{HMG} H=\s u m_i 2\v ec{\sigma }_i\vec\ | bigotimes_{i=1}^N
x_i\big)=\bigotimes_{i=1}^N x_{i+1}$_where sites_$N+1$ and $1$ are_identified. The_interaction_is called_$L$-*local*_if $h$ acts_non-trivially only on_$L$ neighboring sites, and_it is said_to_be *frustration free* with respect to its ground state $\phi_0$ if the latter minimizes_the_energy locally_in_the_sense that $\langle \phi_0|H|\phi_0\rangle=\inf_\phi \langle
\phi|H|\phi\rangle=N_\inf_\phi \langle \phi|h|\phi\rangle$. As proven_in [@Hast1]_all gapped Hamiltonians can be approximated by frustration_free_ones if one_allows for enlarging the interaction range $L$ up to_${\cal O}(\log N)$.
For every MPS and_FCS $\psi$ one_can_easily_find frustration free Hamiltonians_such that $\psi$ is their exact_ground state. Moreover, these *parent* Hamiltonains_are $L$-local with $L\sim 2 \log D/\log_d$ and they allow for a_detailed analysis of the ground_state degeneracy_(Sec.\[sec:Huniqueness\]) and the energy gap_above the ground_state (Sec.\[Sec:Hgap\])._Typically, these Hamiltonians_are, however, not exactly solvable, i.e.,_information about the_excitations might be hard to obtain.
Examples
--------
1.__[*AKLT*]{}: The father_of_all_matrix product_states is the_ground_state of_the_AKLT-Hamiltonian $$\label{HAKLT}
__ H=\sum_i
_ \vec{S}_i\vec{S}_{i+1}+\frac13\Big(\vec{S}_i\vec{S}_{i+1}\Big)^2\;,$$_where_$\vec{S}$ is the vector_of spin-1 operators (i.e., d=3)._Its MPS representation is given by_$\{A_i\}=\big\{\sigma^z,\sqrt{2}\sigma^+,-\sqrt{2}\sigma^-\big\}$ where_the $\sigma$’s_are the Pauli matrices.
2. [*Majumdar-Gosh*]{}: The Hamiltonian $$\label{HMG}
_ H=\sum_i_2\vec{\sigma}_i\vec\ |
operated between May and August 2000.
After a short recall on the main features of positron sources using channeling in oriented crystals, the experimental set-up is described. A particular emphasis is put on the positron detector made of a drift chamber, partially immersed in a magnetic field. The enhancement in photon and positron production in the aligned crystal have been observed in the energy range 5 to 40 GeV, for the incident electrons, in crystals of 4 and 8 mm as in an hybrid target. The first results concerning this experiment are presented hereafter.
author:
- |
R.Chehab, R.Cizeron, C.Sylvia (LAL-IN2P3, Orsay, France)\
V.Baier, K.Beloborodov, A.Bukin, S.Burdin, T.Dimova, A.Drozdetsky, V.Druzhinin, M.Dubrovin[^1],\
V.Golubev, S.Serednyakov, V.Shary, V.Strakhovenko (BINP, Novosibirsk, Russia)\
X.Artru, M.Chevallier, D.Dauvergne, R.Kirsch, Ph.Lautesse, J-C.Poizat,\
J.Remillieux (IPNL-IN2P3, Villeurbanne, France)\
A.Jejcic (LMD-Universite, Paris, France)\
P.Keppler, J.Major (Max-Planck Institute, Stuttgart, Germany)\
L.Gatignon (CERN, Geneva, Switzerland)\
G.Bochek, V.Kulibaba, N.Maslov (KIPT, Kharkov, Ukraine)\
A.Bogdanov, A.Potylitsin, I.Vnukov (NPI-TPU, Tomsk, Russia)
title: 'Experimental Determination of the Characteristics of a Positron Source Using Channeling[^2].'
---
Introduction
============
The enhancement of radiation observed, in channeling conditions, in a crystal with respect to bremsstrahlung, makes crystal targets interesting for obtaining large positron yields: the high rate of photons generated along a crystal axis produces a corresponding high positron yield in the same crystal | operated between May and August 2000.
After a short recall on the main features of positron source use channeling in oriented crystals, the experimental bent - up is described. A particular stress is arrange on the positron detector made of a drift bedroom, partially immerse in a magnetic playing field. The enhancement in photon and positron production in the aligned quartz glass have been observed in the energy range 5 to 40 GeV, for the incident electrons, in crystals of 4 and 8 mm as in an hybrid target. The inaugural results concerning this experiment are presented hereafter.
generator:
- |
R.Chehab, R.Cizeron, C.Sylvia (LAL - IN2P3, Orsay, France)\
V.Baier, K.Beloborodov, A.Bukin, S.Burdin, T.Dimova, A.Drozdetsky, V.Druzhinin, M.Dubrovin[^1],\
V.Golubev, S.Serednyakov, V.Shary, V.Strakhovenko (BINP, Novosibirsk, Russia)\
X.Artru, M.Chevallier, D.Dauvergne, R.Kirsch, Ph. Lautesse, J - C.Poizat,\
J.Remillieux (IPNL - IN2P3, Villeurbanne, France)\
A.Jejcic (LMD - Universite, Paris, France)\
P.Keppler, J.Major (Max - Planck Institute, Stuttgart, Germany)\
L.Gatignon (CERN, Geneva, Switzerland)\
G.Bochek, V.Kulibaba, N.Maslov (KIPT, Kharkov, Ukraine)\
A.Bogdanov, A.Potylitsin, I.Vnukov (NPI - TPU, Tomsk, Russia)
title:' Experimental Determination of the Characteristics of a Positron Source Using Channeling[^2 ].'
---
Introduction
= = = = = = = = = = = =
The enhancement of radiation detect, in channeling conditions, in a crystal with respect to bremsstrahlung, makes quartz glass targets interesting for obtaining large positron yields: the high rate of photon generate along a crystal axis produces a corresponding high positron yield in the same crystal | opfrated between May and Amgust 2000.
After c short recalm on the main features of positron slueces ysing channeling in orkented crjstals, thw exkerimental set-up ma descrlyed. A iartieuoar emphasis ix put on tve positron dedeztlr made of a drift chamber, partiallr immerxef in a magnetis fitld. The vnmancement in photon and positron producuion in the alignec crystal have been observfd ij the energy range 5 to 40 GeV, fir trw incident euectrons, in crystals or 4 and 8 mm as in an hybrid targdt. Thz first resolfs woncerning vhis eqperiment are presentad heresfter.
author:
- |
R.Rhehqb, R.Cizeron, C.Sylvia (LEL-IN2P3, Orsay, France)\
V.Baier, N.Bzloborodov, A.Bukin, S.Buedun, T.Dhmovd, A.Dfizddtsiy, V.Sruzhijin, M.Dubrovin[^1],\
V.Golubec, S.Serednyakov, V.Shaty, N.Xtrakhovenko (BINP, Godosibirsk, Russia)\
X.Artru, M.Chevallier, G.Dahvergne, R.Kirsch, Ph.Lautewse, J-C.Poizat,\
J.Remiplieux (IPGL-IN2P3, Villeurbanne, France)\
A.Jejcic (LMD-Universita, Parms, Frckge)\
P.Neppler, J.Major (Max-Planck Institute, Stuttgart, Dsrkaky)\
L.Gatignon (GERN, Geneva, Switzetlwnc)\
G.Bochek, V.Yulibayz, H.Maslov (KIPT, Kharklv, Ukrayne)\
A.Bogdanjv, A.Lotylitsin, I.Vnukov (NPI-TPU, Timsk, Russia)
tptle: 'Experimental Detexmination of the Charscteristics of a Positrun Skurce Using Channelihe[^2].'
---
Introduction
============
The enmanwement of radiation observqd, in chaineliug condigionx, in a crystal wlth rcvpect to bremsstrajlung, mdkes crystwl targets interesting for obtamiing large poxidrot yields: the migh rate of phjtons generateb along c crysgal axis pgoduces a correspondyng high posidton yield in vhe same srysral | operated between May and August 2000. After recall the main of positron sources the set-up is described. particular emphasis is on the positron detector made of drift chamber, partially immersed in a magnetic field. The enhancement in photon and production in the aligned crystal have been observed in the energy range 5 40 for incident in crystals of 4 and 8 mm as in an hybrid target. The first results concerning experiment are presented hereafter. author: - | R.Chehab, C.Sylvia (LAL-IN2P3, Orsay, France)\ K.Beloborodov, A.Bukin, S.Burdin, T.Dimova, A.Drozdetsky, M.Dubrovin[^1],\ S.Serednyakov, V.Shary, (BINP, Russia)\ M.Chevallier, D.Dauvergne, R.Kirsch, J-C.Poizat,\ J.Remillieux (IPNL-IN2P3, Villeurbanne, France)\ A.Jejcic (LMD-Universite, Paris, France)\ P.Keppler, J.Major (Max-Planck Institute, Stuttgart, Germany)\ L.Gatignon (CERN, Switzerland)\ G.Bochek, (KIPT, Kharkov, A.Bogdanov, I.Vnukov Tomsk, Russia) title: of the Characteristics of a Positron --- Introduction ============ The enhancement of radiation observed, channeling conditions, a crystal with respect to bremsstrahlung, crystal targets interesting for obtaining large positron yields: high rate of photons generated along a crystal axis produces a corresponding high positron yield same crystal | operated between May and AuguSt 2000.
After a shOrt reCalL on ThE maiN feaTures of positroN SourCes using channeling in orIenteD cRYstaLS, tHe expErimentAL sET-Up iS dEsCriBeD. a pArticUlaR emphasIs is put on tHe pOsItron detectoR MaDe of a drift ChaMber, partiallY imMersed In A maGNetic FieLd. The EnhancEMent in Photon and PoSItron pROductioN IN tHe alIgned crystal have bEEn OBserved in the enErgy raNgE 5 To 40 gEv, foR thE incident eLeCtronS, In crystALs OF 4 ANd 8 mM As in an hybrid tArget. The firST reSults cOnCerNIng thiS expeRiMEnt Are presenteD herEafter.
autHor:
- |
R.ChEHab, R.CizERon, C.SylVia (LAL-iN2P3, orsAy, FrANcE)\
V.baiEr, k.belOBoRodOV, A.BUkin, S.BurDiN, T.dimovA, A.DrOZDETsky, v.DrUzhiNin, M.DUbrovin[^1],\
V.GolubEv, S.sereDNyaKov, V.SHary, V.straKhOvenkO (BINP, NOvosiBiRsk, Russia)\
X.Artru, m.CheVallier, D.DAuvErGne, r.KIrsch, pH.LauteSse, j-C.POizat,\
J.REmillieUX (IPnL-in2p3, viLleurbanne, France)\
A.JEjCIC (LmD-UniverSite, PaRIs, frANce)\
P.KeppLeR, J.MAjor (mAX-PlanCk InSTiTute, StutTgart, GERmAnY)\
L.GatigNoN (CERN, GEnEva, swiTzerlANd)\
G.BOchek, V.kulibaba, n.MaslOV (KIPT, Kharkov, UkRAine)\
A.Bogdanov, a.poTYLiTSin, I.vnuKov (NPI-TPU, ToMsk, RUSsia)
TitlE: 'exPerIMentaL DeteRmINaTIon of the CharacterisTiCs of a POsitrOn Source Using channeling[^2].'
---
iNTRoductioN
============
The ENhANcement of radiaTion oBserved, in cHAnneling CondiTions, in a Crystal wiTH Respect tO brEmsStrAhlUNG, mAkes crystal taRGEts iNtErestinG foR obtainIng LarGe pOsiTrOn yields: tHe high raTe Of PhOtOns GenerATed along A cRysTaL axIs proDUces a cOrresPondInG hIGh pOsitron YIeLD In thE sAmE cryStaL | operated between May andAugust 200 0.
Aft er a s hort recall on the main features of positronsourc es usin g c hanne ling in or i e nte dcr yst al s ,the e xpe rimenta l set-up i s d es cribed. A pa r ti cular emph asi s is put onthe posit ro n d e tecto r m ade o f a dr i ft cha mber, par ti a lly im m ersed i n amagn etic field. The e n ha n cement in phot on and p o si t r onpro duction in t he al i gned cr y st a l hav e been observe d in the en e rgy range 5 to 40 GeV , for t h e i ncident ele ctro ns, in cr ystals of 4 an d 8 mm a s in a n h ybr id t a rg et . T he fir s tres u lts concern in gthisexpe r i m e nt a repres ented hereafter.
a uth or:- | R .Cheh ab,R. Cizer on, C. Sylvi a(LAL-IN2P3, Ors ay,France)\ V .Ba ie r, K. B elobor odo v,A.Bukin , S.Bur d in, T . D i mo va, A.Drozdetsky,V. D r uz hinin, M .Dubro v in [^ 1 ],\
V. Gol ubev , S.Ser edny a ko v, V.Sha ry, V. S tr ak hovenko ( BINP,No vos ibi rsk,R ussi a)\
X.Artr u, M. C hevallier, D.D a uvergne, R.Ki r sc h , P h .Lau tes se, J-C.Poi zat, \
J.R e mi lli e ux (I PNL-I N2 P 3, Villeurbanne, Franc e) \
A.Jej cic (LMD-Univ ersite, Pa r i s , France )\
P .Keppler, J.Ma jor ( Max-Planck Institut e, St uttgart, Germany) \ L.Ga tig non (C ERN , Ge neva, Switzer l a nd)\
G.Bo che k, V.Ku lib aba , N .Ma sl ov (KIPT, Kharkov ,Uk ra in e)\
A .Bogdano v, A. Po tyl itsin , I.Vnu kov ( NPI- TP U, Tom sk, Rus s ia ) titl e: ' Expe rim en tal D eter m ina tion of the Char act e rist ic sof a Po sitron Source U sing Chann el ing [^2].' - --
Intr oduction
============
T he enha nce mentof r adiationobs erved, in channe ling c ondit io ns, i n a c r y st alwi th respect t o b remss tr ahlu ng, mak es crystal targets int eresting forobt aini n g l arg e p o sit ro n yi e l ds: the high ra te of phot on s g enerated a l ong a crysta l axisprodu c es a co rrespondi ng high p os itro n yie ld in thesame cry stal | operated_between May_and August 2000.
_ _After_a short_recall_on the main_features of positron_sources using channeling in_oriented crystals, the_experimental_set-up is described. A particular emphasis is put on the positron detector made of_a_drift chamber,_partially_immersed_in a magnetic field. The_enhancement in photon and positron_production in_the aligned crystal have been observed in the_energy_range 5 to_40 GeV, for the incident electrons, in crystals of_4 and 8 mm as in_an hybrid target._The_first_results concerning this experiment_are presented hereafter.
author:
- |
_ R.Chehab, R.Cizeron, C.Sylvia (LAL-IN2P3, Orsay,_France)\
V.Baier, K.Beloborodov, A.Bukin,_S.Burdin, T.Dimova, A.Drozdetsky, V.Druzhinin, M.Dubrovin[^1],\
_ V.Golubev, S.Serednyakov, V.Shary,_V.Strakhovenko (BINP,_Novosibirsk, Russia)\
_X.Artru, M.Chevallier, D.Dauvergne,_R.Kirsch, Ph.Lautesse,_J-C.Poizat,\
_ J.Remillieux (IPNL-IN2P3, Villeurbanne, France)\
_ A.Jejcic_(LMD-Universite, Paris, France)\
_P.Keppler,_J.Major (Max-Planck Institute,_Stuttgart,_Germany)\
_ _ L.Gatignon (CERN,_Geneva,_Switzerland)\
__ G.Bochek, V.Kulibaba, N.Maslov (KIPT, Kharkov,_Ukraine)\
_ A.Bogdanov, A.Potylitsin, I.Vnukov (NPI-TPU,_Tomsk, Russia)
title: 'Experimental Determination_of_the Characteristics of a_Positron Source Using Channeling[^2].'
---
Introduction
============
The enhancement_of radiation observed, in channeling conditions,_in a_crystal with_respect to bremsstrahlung, makes crystal targets interesting for obtaining large positron_yields: the high rate of photons_generated along a crystal_axis produces_a_corresponding high positron_yield_in the_same crystal |
nuclei (*b* particles, whose multiplicity is denoted by *n*$_b$), as well as nonrelativistic protons of energy in excess of 26 MeV (*g* particles, whose multiplicity is denoted by *n*$_g$). Relevant reactions are also characterized by the multiplicity *n*$_s$ of product mesons (*s* particles). Using these parameters, one can draw preliminary conclusions on the character of the interaction.
The $^8$B nucleus is first studied by the method of nuclear track emulsion. This makes it possible to deduce information both about the dissociation channel $^8$B $\rightarrow^7$Be + *p* and about dissociation into extremely light nuclei He and H. Microphotographs of such events were presented in [@Artemenkov08]. Of particular interest are events that do not contain target-nucleus fragments or charged mesons (*n*$_b$ = 0, *n*$_g$ = 0, *n*$_s$ = 0)$-$these are so-called white stars. They form an observational basis in searches for electromagnetic-dissociation events. The objective of present study is to seek white stars that are formed at the minimum momentum transfer and which could be associated with the electromagnetic dissociation of $^8$B nuclei.
A theoretical analysis of the contributions of various mechanisms to the cross section for the reaction $^8$B $\rightarrow^7$Be + *p* of proton separation from the $^8$B nucleus was performed in [@Esbensen2000], where predictions were made for the behavior of the cross section as a function of the energy of the $^8$B nucleus in the region extending up to about 2 GeV per nucleon and for the branching fractions of Coulomb dissociation, nuclear dissociation, and the stripping reaction for the case of a Pb target. All of these mechanisms lead to the maximum cross section at an impact parameter of *b* $\approx$ 10 fm $-$ that is, in the case of closest approach of participant nuclei without overlap of the densities. The contribution of the stripping reaction to the cross section is approximately five times as great as the contributions of Coulomb and nuclear diffraction, but it decreases sharply with increasing *b*. The nuclear diffraction contribution decreases at the same rate with increasing *b*. At impact-parameter values in the range *b* = 12-15 fm, Coulomb dissociation becomes dominant. The integrated cross sections at 1.2 GeV per nucleon are about 210 (Coulomb interaction), 190 (stripping) and 50 mb (nuclear | nuclei (* b * particles, whose multiplicity is denoted by * n*$_b$), as well as nonrelativistic proton of department of energy in overindulgence of 26 MeV (* g * particles, whose numerousness is denoted by * n*$_g$). Relevant reaction are also characterize by the numerousness * n*$_s$ of product mesons (* s * particles). Using these parameters, one can guide preliminary conclusions on the character of the interaction.
The $ ^8$B lens nucleus is foremost studied by the method of nuclear racetrack emulsion. This makes it possible to deduce data both about the dissociation channel $ ^8$B $ \rightarrow^7$Be + * p * and about dissociation into extremely light nucleus He and H. Microphotographs of such events were presented in [ @Artemenkov08 ]. Of especial interest are event that do not contain target - nucleus fragments or charged meson (* n*$_b$ = 0, * n*$_g$ = 0, * n*$_s$ = 0)$-$these are so - called white stars. They form an observational footing in search for electromagnetic - dissociation events. The objective of present study is to seek white stars that are formed at the minimum momentum transfer and which could be associate with the electromagnetic dissociation of $ ^8$B nuclei.
A theoretical analysis of the contribution of versatile mechanisms to the cross incision for the reaction $ ^8$B $ \rightarrow^7$Be + * p * of proton separation from the $ ^8$B nucleus was performed in [ @Esbensen2000 ], where prediction were made for the behavior of the cross section as a function of the energy of the $ ^8$B lens nucleus in the region extending up to about 2 GeV per nucleon and for the branching fraction of Coulomb dissociation, nuclear dissociation, and the stripping chemical reaction for the shell of a Pb target. All of these mechanisms lead to the maximum hybridization section at an impact parameter of * b * $ \approx$ 10 frequency modulation $ -$ that is, in the case of closest approach of participant nuclei without lap of the densities. The contribution of the stripping reaction to the cross section is approximately five fourth dimension as bang-up as the contributions of Coulomb and nuclear diffraction, but it decrease sharply with increase * b *. The nuclear diffraction contribution decreases at the same rate with increasing * bacillus *. At impact - argument values in the range * b * = 12 - 15 fm, Coulomb dissociation becomes prevailing. The integrated cross sections at 1.2 GeV per nucleon are about 210 (Coulomb interaction), 190 (denudation) and 50 mb (nuclear | nuflei (*b* particles, whose mmltiplicity is dgnited bb *n*$_b$), as well as nonrelativistic protons of xnerty in excess of 26 MeV (*g* partkcles, whode multiplicmty is denoted bb *n*$_g$). Relcrant dcactimis are also chatacterized bf the multipliwigy *n*$_s$ of product mesons (*s* particles). Ufing thrsf parameters, oge csg drzw preliminary conclusions on the dharacttr of the interactoon.
The $^8$B nucleus is first dtudled by the method lf nuclear jdach emulsion. Thks makes iu 'ossible to deduce information both about ghe dnssociation cyanjgl $^8$B $\rightarcow^7$Be + *p* and about dissocidtion imto extremely kigit nyclei He and H. Micropiotographs of such edents wera 'resented in [@Artemenkiv08]. Of pdrtiwulaf ingertst ade evejts that do nkt contain rarget-nucleus fragmtntf or charged mssons (*g*$_b$ = 0, *n*$_g$ = 0, *n*$_s$ = 0)$-$these are so-called white stdrs. They form an observatiinal basis in searched for elestromagnetic-dissociation events. The objective of [reseit stbey is ro seek white stars that are formed at the miniihm mpmentum transfcr and which could bf sfsociated witf the zmedtromagnetic dissofiation of $^8$B nuclei.
A uheorrtical analysis of the conteibutions of carious mechanisms to the crors sgction for the reaction $^8$B $\rigktarros^7$Be + *p* of pgoton sepzfation from the $^8$C nlclegs was ptfformed in [@Esbensqn2000], where 'redietions wdre kade fjr the behwvior of the cross sectiln as a function lf the energy of the $^8$B nucleus mi the region gxtandpng up to aboub 2 GeV per nuclqon and for thg branchiug fraztions of Boulomb dmssociation, guclear dissowlation, and tie strippyng eeacrion fof the case of a Pb targen. Cll of thwse mechanisms leav to fhe maximum crows wection at an ikpazt [agamxter jx *b* $\approx$ 10 xm $-$ ghag is, iv the case iw clpsest approach of pasticjpant nuclei withoit overlap of the qensities. The contribution of tje stcippinj reacyiog to the cross section is apprkximately fiye times as gweat as the contxibutions of Coulomb and nuclear diffracvion, but it decreases syarply with increasnnn *b*. The nuclxar disfraction contribution decreawes at the same rste with increasing *b*. At im[act-pwrameter values in the range *b* = 12-15 fm, Coulomb dissociation becomes dominant. The iitqgrated croas srctiots ct 1.2 GeV per iucleon are about 210 (Coulomb interaction), 190 (strippinj) and 50 mb (tuelear | nuclei (*b* particles, whose multiplicity is denoted as as nonrelativistic of energy in particles, multiplicity is denoted *n*$_g$). Relevant reactions also characterized by the multiplicity *n*$_s$ product mesons (*s* particles). Using these parameters, one can draw preliminary conclusions on character of the interaction. The $^8$B nucleus is first studied by the method nuclear emulsion. makes possible to deduce information both about the dissociation channel $^8$B $\rightarrow^7$Be + *p* and about dissociation extremely light nuclei He and H. Microphotographs of events were presented in Of particular interest are events do contain target-nucleus or mesons = 0, *n*$_g$ 0, *n*$_s$ = 0)$-$these are so-called white stars. They form an observational basis in searches for electromagnetic-dissociation The objective study is seek stars are formed at momentum transfer and which could be electromagnetic dissociation of $^8$B nuclei. A theoretical analysis the contributions various mechanisms to the cross section the reaction $^8$B $\rightarrow^7$Be + *p* of proton from the $^8$B nucleus was performed in [@Esbensen2000], where predictions were made for the behavior cross section as a of the energy the nucleus the extending up about 2 GeV per nucleon and for the branching fractions of dissociation, nuclear dissociation, and the stripping reaction for the case Pb All of these lead to the maximum section an impact parameter of 10 $-$ the of approach of participant nuclei overlap of the densities. The of the stripping reaction approximately five times as great as the contributions Coulomb and nuclear diffraction, but it decreases with increasing *b*. The nuclear diffraction contribution decreases at the same rate increasing *b*. values in the range *b* = 12-15 fm, dissociation becomes dominant. The cross sections at 1.2 GeV per nucleon are about (Coulomb 190 (stripping) 50 mb (nuclear | nuclei (*b* particles, whose multIplicity is DenotEd bY *n*$_b$), As Well As noNrelativistic pROtonS of energy in excess of 26 MeV (*G* partIcLEs, whOSe MultiPlicity IS dENOteD bY *n*$_G$). ReLeVAnT reacTioNs are alSo characteRizEd By the multiplICiTy *n*$_s$ of prodUct Mesons (*s* partiCleS). Using ThEse PArameTerS, one cAn draw PRelimiNary conclUsIOns on tHE characTER oF the Interaction.
The $^8$B nuCLeUS is first studieD by the MeTHoD OF nuCleAr track emuLsIon. ThIS makes iT PoSSIBle TO deduce informAtion both abOUt tHe dissOcIatIOn chanNel $^8$B $\rIgHTarRow^7$Be + *p* and abOut dIssociatiOn into EXtremelY Light nuClei He And h. MiCropHOtOgRapHs OF suCH eVenTS weRe presenTeD iN [@ArteMenkOV08]. oF PartIcuLar iNtereSt are events thAt dO not COntAin taRget-nUcleUs FragmEnts or ChargEd Mesons (*n*$_b$ = 0, *n*$_g$ = 0, *n*$_s$ = 0)$-$theSe arE so-called WhiTe StaRs. they fORm an obSerVatIonal baSis in seARchEs FOR ElEctromagnetic-dissoCiATIoN events. THe objeCTiVe OF present StUdy Is to SEEk whiTe stARs That are fOrmed aT ThE mInimum mOmEntum tRaNsfEr aNd whiCH couLd be asSociated With tHE electromagnetIC dissociation OF $^8$B NUClEI.
A thEorEtical analySis oF The cOntrIBuTioNS of vaRious MeCHaNIsms to the cross sectiOn For the ReactIon $^8$B $\rightarroW^7$Be + *p* of protON SEparatioN froM ThE $^8$b nucleus was perFormeD in [@EsbenseN2000], Where preDictiOns were mAde for the BEHavior of The CroSs sEctION aS a function of tHE EnerGy Of the $^8$B nUclEus in thE reGioN exTenDiNg up to aboUt 2 GeV per NuClEoN aNd fOr the BRanching FrActIoNs oF CoulOMb dissOciatIon, nUcLeAR diSsociatIOn, AND the StRiPpinG reAcTion fOr thE CasE of a Pb tArget. All oF thESe meChAnIsms leaD to the maximum CrOss section At An iMpact pARAmeter of *B* $\approx$ 10 fm $-$ that is, in the casE Of closeSt aPproaCh of ParticipaNt nUclei wIthOUt overLap of tHe denSiTieS. tHe conTRIbUtiOn Of the stripPINg rEactiOn To thE cross sEction is approximatELy fIve times as greAt aS the CONtRibUTiONs oF COUloMB And nuclear diffrAction, but iT dECrEases sharpLY wiTh IncreasIng *b*. The NucleAR diffraCtion contRibution dEcReasES At tHe same rate With incrEasing *b*. At IMpact-PArAmeteR vaLues in ThE raNge *b* = 12-15 fM, CouloMB diSsociAtion bEcOmes doMinanT. THe integrAted cross sections at 1.2 GeV pEr nuclEon arE abOut 210 (CoulomB inTEraCtion), 190 (striPpinG) and 50 mb (nuclEar | nuclei (*b* particles, wh ose multip licit y i s d en oted by*n*$_b$), as w e ll a s nonrelativistic prot ons o fe nerg y i n exc ess of2 6M e V ( *g *par ti c le s, wh ose multip licity isden ot ed by *n*$_g $ ). Relevantrea ctions are a lso chara ct eri z ed by th e mul tiplic i ty *n* $_s$ of p ro d uct me s ons (*s * pa rtic les). Using these pa r ameters, one c an dra wp re l i min ary conclusio ns on t h e chara c te r o f t h e interaction .
The $^8$ B nu cleusis fi r st stu diedby the method ofnucl ear track emuls i on. Thi s makesit pos sib leto d e du ce in fo r mat i on bo t h a bout the d is socia tion c h a nnel $^ 8$B$\rig htarrow^7$Be+ * p* a n d a boutdisso ciat io n int o extr emely l ight nuclei HeandH. Microp hot og rap hs of s u ch eve nts we re pres ented i n [@ Ar t e m en kov08]. Of particu la r in terest a re eve n ts t h at do no tcon tain t arget -nuc l eu s fragme nts or ch ar ged mes on s (*n* $_ b$= 0 , *n* $ _g$= 0, * n*$_s$ = 0)$- $ these are so-c a lled white st a rs . Th e y fo rman observat iona l bas is i n s ear c hes f or el ec t ro m agnetic-dissociatio nevents . The objective of present s t u d y is toseek wh i te stars thatare f ormed at t h e minimu m mom entum tr ansfer an d which co uld be as soc i a te d with the el e c trom ag netic d iss ociatio n o f $ ^8$ B n uc lei.
A t heoretic al a na ly sis of t h e contri bu tio ns of vari o us mec hanis ms t oth e cr oss sec t io n forth ereac tio n$^8$B $\r i ght arrow^7 $Be + *p* of prot on s eparati on from the $ ^8 $B nucleus w asperfor m e d in [@E sbensen2000], where pre d ictions we re ma de f or the be hav ior of th e cross secti on as a fu n c tiono f t heen ergy of th e $^8 $B nu cl eusin theregion extending u p to about 2 GeVper nuc l e on an d f o r t he bra n c hing fractionsof Coulomb d i ss ociation,n ucl ea r disso ciation , and the str ipping re action fo rthec a seof a Pb ta rget. Al l of thes e mech a ni sms l ead to th emax imumcrosss ect ion a t an i mp act pa ramet er of *b*$\approx$ 10 fm $-$ tha t is,in th e c ase of cl ose s t a pproach o f pa rticipantnuc lei with out overl ap o f t hed ensit ies. The contr i bu tio n of the stripp i n g re actio n t o the c ross section is appro x imately five t imes a s g rea t asth e contribution s o fC o ulomb an dnuclear dif fraction ,b ut it decre ases s harplyw i th increa sing *b *. The nu cle ar diffrac ti on contri buti on decre ases a t the s ame rate with in creas i n g *b* . At impa ct -parame t er v alues in t he range *b * = 12 -15fm, C oulombdi ssocia tio nbecomes do m inant. Th e int egrated c ross se ctions at1 . 2 GeV per n ucl eon are a b o ut 21 0( Cou lomb inte ra ctio n), 190 ( s tripping ) a n d 50 mb ( nuc l e ar | nuclei_(*b* particles,_whose multiplicity is denoted_by *n*$_b$),_as_well as_nonrelativistic_protons of energy_in excess of_26 MeV (*g* particles,_whose multiplicity is_denoted_by *n*$_g$). Relevant reactions are also characterized by the multiplicity *n*$_s$ of product mesons_(*s*_particles). Using_these_parameters,_one can draw preliminary conclusions_on the character of the_interaction.
The $^8$B_nucleus is first studied by the method of_nuclear_track emulsion. This_makes it possible to deduce information both about the_dissociation channel $^8$B $\rightarrow^7$Be + *p*_and about dissociation_into_extremely_light nuclei He and_H. Microphotographs of such events were_presented in [@Artemenkov08]. Of particular interest_are events that do not contain target-nucleus_fragments or charged mesons (*n*$_b$ =_0, *n*$_g$ = 0, *n*$_s$_= 0)$-$these_are so-called white stars. They_form an observational_basis in_searches for electromagnetic-dissociation_events. The objective of present study_is to seek_white stars that are formed at_the_minimum momentum transfer_and_which_could be_associated with the_electromagnetic_dissociation of_$^8$B_nuclei.
A theoretical analysis of the contributions_of_various mechanisms to the cross section for_the reaction $^8$B $\rightarrow^7$Be_+_*p* of proton separation_from the $^8$B nucleus was_performed in [@Esbensen2000], where predictions were_made for_the behavior_of the cross section as a function of the energy of_the $^8$B nucleus in the region_extending up to about_2 GeV_per_nucleon and for_the_branching fractions_of Coulomb dissociation, nuclear dissociation, and the_stripping reaction_for the case of a Pb_target. All of these_mechanisms_lead to the maximum cross section_at an impact parameter of *b*_$\approx$ 10 fm $-$ that_is,_in_the case of closest approach_of participant nuclei without overlap of_the densities. The_contribution of the stripping reaction to the_cross_section is approximately five times as_great_as the contributions of Coulomb and_nuclear_diffraction,_but it decreases sharply with_increasing *b*. The nuclear diffraction contribution_decreases at the same rate with increasing *b*. At_impact-parameter values in_the range *b* = 12-15_fm,_Coulomb_dissociation becomes dominant. The integrated cross sections at 1.2 GeV_per nucleon_are about 210_(Coulomb interaction), 190 (stripping) and 50 mb (nuclear |
apping is to select the density matrices from the input and output sets, such that the outgoing entanglement rate of the quantum repeater is maximized; this also entails the boundness of delays. The maximization procedure characterizes the problem of optimal entanglement swapping in the quantum repeaters.
Finding the optimal entanglement swapping means determining the entanglement swapping between the incoming and outgoing density matrices that maximizes the outgoing entanglement rate of the quantum repeaters. The problem of entanglement rate maximization must be solved for a particular noise level in the quantum repeater and with the presence of various entanglement swapping sets. The noise level in the proposed model is analogous to the lost density matrices in the quantum repeater due to imperfections in the local operations and errors in the quantum memory units. The entanglement swapping sets are logical sets that represent the actual state of the quantum memory in the quantum repeater. The entanglement swapping sets are formulated by the set of received density matrices stored in the local quantum memory and the set of outgoing density matrices, which are also stored in the local quantum memory. Each incoming and outgoing density matrix represent half of an entangled system, such that the other half of an incoming density matrix is stored in the distant source quantum repeater, while the other half of an outgoing density matrix is stored in the distant target quantum repeater. The aim of determining the optimal entanglement swapping method is to apply the local entanglement swapping operation on the set of incoming and outgoing density matrices such that the outgoing entanglement rate of the quantum repeater is maximized at a particular noise level. As we prove, the entanglement rate maximization procedure depends on the type of entanglement swapping sets formulated by the stored density matrices in the quantum memory. We define the logical types of the entanglement swapping sets and characterize the main attributes of the swapping sets. We present the efficiency of the entanglement swapping procedure as a function of the local noise and its impacts on the entanglement rate. We prove that the entanglement swapping sets can be defined as a function of the noise, which allows us to define noise-scaled entanglement swapping and noise-scaled entanglement rate maximization. The proposed theoretical framework utilizes the fundamentals of queueing theory, such as the Lyapunov methodology [@refL1], which is an analytical tool used to assess the performance of queueing systems [@refL1; @refL2; @refL3; @refL4; @refD1; @refs3], and defines a fusion of queueing theory with quantum | apping is to select the density matrices from the input and end product set, such that the outgoing entanglement rate of the quantum repeating firearm is maximized; this also entail the boundness of delays. The maximization procedure qualify the problem of optimal web swapping in the quantum repeaters.
Finding the optimal web swapping means determining the web swapping between the incoming and extroverted concentration matrices that maximizes the outgoing entanglement pace of the quantum repeaters. The problem of web rate maximization must be resolve for a particular randomness level in the quantum repeater and with the presence of various web swapping sets. The noise level in the proposed model is analogous to the lost density matrices in the quantum repeater due to imperfections in the local operations and errors in the quantum memory units. The entanglement swapping sets are logical sets that typify the actual state of the quantum memory in the quantum repeater. The web swapping set are formulate by the set of received concentration matrices store in the local quantum memory and the set of outgoing concentration matrices, which are also stored in the local quantum memory. Each incoming and outgoing density matrix represent half of an entangled system, such that the other half of an incoming density matrix is store in the distant reference quantum repeater, while the other half of an outgoing density matrix is store in the distant target quantum repeater. The purpose of determining the optimal entanglement swapping method is to give the local entanglement swap operation on the set of incoming and extroverted density matrices such that the outgoing entanglement pace of the quantum repeater is maximized at a especial noise level. As we raise, the entanglement pace maximization procedure depends on the type of web swapping sets formulated by the stored density matrix in the quantum memory. We define the logical types of the web swapping sets and characterize the main attributes of the swapping sets. We present the efficiency of the entanglement swap procedure as a function of the local noise and its impacts on the entanglement rate. We testify that the web swapping sets can be defined as a function of the noise, which allows us to define noise - scaled web swapping and noise - scaled entanglement rate maximization. The proposed theoretical framework utilizes the fundamentals of queueing theory, such as the Lyapunov methodology [ @refL1 ], which is an analytic tool used to assess the operation of queueing systems [ @refL1; @refL2; @refL3; @refL4; @refD1; @refs3 ], and defines a fusion of queueing theory with quantum | applng is to select the denrity matrices fton the mnput ahd outpug sets, such that the outgoinj enrangltient rate of the qjantum reieater is maxmmized; this also entails the bkmndnevw of delays. Thg maximizatimn procedure cvafaeterizes the problem of optimal entagglemeny dwapping in thg quameum dvptaters.
Finding the optimal entanglsment sxapping means drtermining the entanglemenh swwpping between the incoming abd otrgoing densigy matrices that maximjzes the outgoing entanglement fate pf the quabtym ggpeaters. The probjem of entannkement rate msximization muxt ue silved for a particulac noise level in the quantum se'eater and with the peewence of eariuys dntznjlejent saap'ing sets. Tge noise lecel in the proposed mjeel is analogkus to tre lost density matrices in the quantum releater due to imperfectuons in the local opetations anq errors in the quantum memory units. The entanglekent adapkikn segw wre logical sets that represent the actual stwfe on the quantum meiory in the qkamjum repeater. Tfe entchgmement swapping sehs are sormuoated by uhe srt of received density matruces stored pn tye local quantum mzmory and thz set pf ouygoing density matrices, whidh are also stored ih the local quantjm kekory. Each incoming and outdoing denwity matrix reptesent ralf of an entakcled system, such tjat tke otver half ov an incoming density matrix is stored in the dhstdnt souree quaktum repeater, wrile the other half oy an ojtgoing dehsity metrix is stowed in the divjant target qnantum re[eatwr. Tye aim uw determining yhe optimcj enranglement swappinn metfkd is to apply cke local entanglekeng sraipiig opqsation on tha seg ow incooing and oubgokng censity matrices sucv thzt the outgoing enyakglement tate of tre quantum releater is maximizef at e partmcular noyse level. As we prove, the entahglement gatc maximizatiog pricedure depeuds on the type of entanglement swapping sets formulated bt the stored densiti mstrices in vhe quwntum memmry. We define the lotical types of tht entanglement swapping sets dnd cjaracterize the main attributes of the swapping sets. We present the efficuency of the entanflemrnt sfa'piug procqdurx es a function of nhe local noise and its impacts un the ettcnglement rate. We prove that yhd entanglemenj swapping sets can be defjned as s function of the noise, which aklows us to define noise-xcaled ebtangmement swa'ping and noise-scalec entauglemebt rate maxlmization. The oroloded theorfticel yramework utilizes the fundamentaos of wufueing theory, surh as the Lywkunov mgthodolody [@refL1], which is an zjalytkcal tool usec to assess thr performance of queleigg systems [@refO1; @refL2; @refL3; @rgfL4; @refD1; @refs3], and dehines a fusimn of queueing theory with quantum | apping is to select the density matrices input output sets, that the outgoing repeater maximized; this also the boundness of The maximization procedure characterizes the problem optimal entanglement swapping in the quantum repeaters. Finding the optimal entanglement swapping means the entanglement swapping between the incoming and outgoing density matrices that maximizes the entanglement of quantum The problem of entanglement rate maximization must be solved for a particular noise level in the repeater and with the presence of various entanglement sets. The noise level the proposed model is analogous the density matrices the repeater to imperfections in local operations and errors in the quantum memory units. The entanglement swapping sets are logical sets that the actual the quantum in quantum The entanglement swapping formulated by the set of received in the local quantum memory and the set outgoing density which are also stored in the quantum memory. Each incoming and outgoing density matrix half of an entangled system, such that the other half of an incoming density matrix in the distant source repeater, while the half an density is stored the distant target quantum repeater. The aim of determining the optimal swapping method is to apply the local entanglement swapping operation set incoming and outgoing matrices such that the entanglement of the quantum repeater at particular we the rate maximization procedure depends the type of entanglement swapping formulated by the stored memory. We define the logical types of the swapping sets and characterize the main attributes the swapping sets. We present the efficiency of the entanglement swapping procedure a function local noise and its impacts on the entanglement We prove that the swapping sets can be defined as a function of noise, allows us define noise-scaled entanglement and noise-scaled entanglement maximization. The proposed utilizes the of theory, Lyapunov methodology [@refL1], which is an tool to assess the performance of systems @refL3; @refL4; @refD1; @refs3], and defines a fusion of queueing theory with quantum | apping is to select the densitY matrices fRom thE inPut AnD outPut sEts, such that the OUtgoIng entanglement rate of tHe quaNtUM repEAtEr is mAximizeD; ThIS AlsO eNtAilS tHE bOundnEss Of delayS. The maximiZatIoN procedure chARaCterizes thE prOblem of optimAl eNtanglEmEnt SWappiNg iN the qUantum REpeateRs.
Finding ThE OptimaL EntanglEMEnT swaPping means determiNInG The entanglemenT swappInG BeTWEen The Incoming anD oUtgoiNG densitY MaTRICes THat maximizes tHe outgoing eNTanGlemenT rAte OF the quAntum RePEatErs. The problEm of EntanglemEnt ratE MaximizATion musT be solVed For A parTIcUlAr nOiSE leVEl In tHE quAntum repEaTeR and wIth tHE PREsenCe oF varIous eNtanglement swAppIng sETs. THe noiSe levEl in ThE propOsed moDel is AnAlogous to the losT denSity matriCes In The QuAntum REpeateR duE to ImperfeCtions iN The LoCAL OpErations and errors iN tHE QuAntum memOry uniTS. THe ENtanglemEnT swAppiNG Sets aRe loGIcAl sets thAt reprESeNt The actuAl State oF tHe qUanTum meMOry iN the quAntum repEater. tHe entanglement SWapping sets arE FoRMUlATed bY thE set of receiVed dENsitY matRIcEs sTOred iN the lOcAL qUAntum memory and the seT oF outgoIng deNsity matrices, Which are alSO STored in tHe loCAl QUantum memory. EaCh incOming and ouTGoing denSity mAtrix repResent halF OF an entanGleD sySteM, suCH ThAt the other halF OF an iNcOming deNsiTy matriX is StoRed In tHe Distant soUrce quanTuM rEpEaTer, While THe other hAlF of An OutGoing DEnsity MatriX is sToReD In tHe distaNT tARGet qUaNtUm rePeaTeR. The aIm of DEteRmining The optimaL enTAnglEmEnT swappiNg method is to aPpLy the local EnTanGlemenT SWapping oPeration on the set of incomINg and ouTgoIng deNsitY matrices SucH that tHe oUTgoing EntangLemenT rAte OF The quANTuM rePeAter is maxiMIZed At a paRtIculAr noise Level. As we prove, the eNTanGlement rate maXimIzatION pRocEDuRE dePeNDs oN THe type of entanglEment swappInG SeTs formulatED by ThE stored Density MatriCEs in the Quantum meMory. We defInE the LOGicAl types of tHe entangLement swaPPing sETs And chAraCterizE tHe mAin atTributES of The swApping SeTs. We prEsent ThE efficieNcy of the entanglement swaPping pRocedUre As a functiOn oF The Local noisE and Its impacts On tHe eNtangLemENt ratE. We pROvE thAT the eNtanGLement swaPPiNg sETS cAn be defined AS A FunCtion Of tHE noise, WhicH allows us to define NOise-scaled entaNgleMENt sWapPIng aNd Noise-scaled entAngLeMENt rate maXiMization. The Proposed ThEOretiCal fraMework UtilizeS THe FUndameNtalS of Queueing tHeoRy, SUch as thE LYaPUnov meThodOlOgy [@refl1], which IS an aNALytical tool used tO asseSS The peRForMance Of QueueinG SystEms [@refL1; @refl2; @refL3; @refL4; @reFD1; @refs3], And dEfineS a fusioN oF queueIng ThEory with quANtum | apping is to select the de nsity matr icesfro m t he inp ut a nd output sets , suc h that the outgoing en tangl em e nt r a te of t he quan t um r epe at er is m a xi mized ; t his als o entailsthe b oundness ofd el ays. The m axi mization pro ced ure ch ar act e rizes th e pro blem o f optim al entang le m ent sw a pping i n th e qu antum repeaters.Fi n ding the optim al ent an g le m e ntswa pping mean sdeter m ining t h ee n t ang l ement swappin g between t h e i ncomin gand outgoi ng de ns i tymatrices th at m aximizesthe ou t going e n tanglem ent ra teoftheq ua nt umre p eat e rs . T h e p roblem o fen tangl emen t r a te m axi miza tionmust be solve d f or a par ticul ar no isele vel i n thequant um repeater and w iththe prese nce o f v ar iouse ntangl eme ntswappin g sets. The n o i s elevel in the propo se d mo del is a nalogo u sto the lost d ens itym a trice s in th e quantu m repe a te rdue toim perfec ti ons in thel ocal opera tions an d err o rs in the quan t um memory uni t s. T he enta ngl ement swapp ings etsarel og ica l sets that r e pr e sent the actual sta te of th e qua ntum memory i n the quan t u m repeate r. T h ee ntanglement sw appin g sets are formulat ed by the set of recei v e d densit y m atr ice s s t o re d in the loca l quan tu m memor y a nd theset of ou tgo in g density matrice s, w hi ch ar e als o storedin th eloc al qu a ntum m emory . Ea ch i n com ing and ou t g oing d en sity ma tr ix re pres e nthalf of an entan gle d sys te m, such t hat the other h alf of anin com ing de n s ity matr ix is stored in the dis t ant sou rce quan tumrepeater, wh ile th e o t her ha lf ofan ou tg oin g densi t y m atr ix is stored i n t he di st anttargetquantum repeater.T heaim of determ ini ng t h e o pti m al ent an g lem e n t swapping meth od is to a pp l ythe locale nta ng lementswappin g ope r ation o n the set of incom in g an d out going dens ity matr ices such thatt he outg oin g enta ng lem ent r ate of the quan tum re pe ater i s max im ized ata particular noise leve l. Aswe pr ove , the ent ang l eme nt rate m axim ization pr oce dur e dep end s on t he t y pe of entan glem e nt swappi n gset s fo rmulated by t h e s tored de n sity m atri ces in the quantu m memory. We de fine t helog i calty pes of the ent ang le m e nt swapp in g sets andcharacte ri z e the mainattrib utes of t he swappi ng s ets . We pres ent t h e effic ie nc y of th e en ta ngleme nt swa p ping p rocedure as a fu nctio n of th e lo cal n oi se andi ts i mpacts onthe entangl ementrate . Weprove t ha t theent an glement sw a pping set s can be def in ed a s a funct iono f thenois e, wh ich allow s us to d e fin e no ise-s ca ledentanglem e nt swapp ing and noi se -sc a l ed ent a ng l e ment ratemax imiza t i on. The pr o pose d t h eoret ical f ramewo rk util i zes t he fund ame n t als of qu eueing th e ory ,such as theLy apun ov me th odology [@refL1 ], which is an a n a lytica l t o ol u s ed toas sess the per forma nc e of que u e ing system s [@ ref L1; @re fL2 ; @r e fL3; @refL4; @refD1; @re fs3] , and def inesa fus io n of que u eing theoryw ith quantum | apping is_to select_the density matrices from_the input_and_output sets,_such_that the outgoing_entanglement rate of_the quantum repeater is_maximized; this also_entails_the boundness of delays. The maximization procedure characterizes the problem of optimal entanglement swapping_in_the quantum_repeaters.
Finding_the_optimal entanglement swapping means determining_the entanglement swapping between the_incoming and_outgoing density matrices that maximizes the outgoing entanglement_rate_of the quantum_repeaters. The problem of entanglement rate maximization must be_solved for a particular noise level_in the quantum_repeater_and_with the presence of_various entanglement swapping sets. The noise_level in the proposed model is_analogous to the lost density matrices in_the quantum repeater due to imperfections_in the local operations and_errors in_the quantum memory units. The_entanglement swapping sets_are logical_sets that represent_the actual state of the quantum_memory in the_quantum repeater. The entanglement swapping sets_are_formulated by the_set_of_received density_matrices stored in_the_local quantum_memory_and the set of outgoing density_matrices,_which are also stored in the local_quantum memory. Each incoming_and_outgoing density matrix represent_half of an entangled system,_such that the other half of_an incoming_density matrix_is stored in the distant source quantum repeater, while the other_half of an outgoing density matrix_is stored in the_distant target_quantum_repeater. The aim_of_determining the_optimal entanglement swapping method is to apply_the local_entanglement swapping operation on the set_of incoming and outgoing_density_matrices such that the outgoing entanglement_rate of the quantum repeater is_maximized at a particular noise_level._As_we prove, the entanglement rate_maximization procedure depends on the type_of entanglement swapping_sets formulated by the stored density matrices_in_the quantum memory. We define the_logical_types of the entanglement swapping sets_and_characterize_the main attributes of the_swapping sets. We present the efficiency_of the entanglement swapping procedure as a function of_the local noise_and its impacts on the_entanglement_rate._We prove that the entanglement swapping sets can be defined_as a_function of the_noise, which allows us to define noise-scaled entanglement swapping and_noise-scaled entanglement rate maximization. The proposed theoretical_framework utilizes the fundamentals of queueing theory, such as the Lyapunov_methodology [@refL1], which is an analytical tool used_to assess the performance of queueing systems_[@refL1; @refL2; @refL3; @refL4;_@refD1;_@refs3], and defines a fusion of queueing_theory_with_quantum |
-T$ curves for an NTO thin film grown on an STO substrate at 850C in vacuum.](HALL_1.EPS){width="0.9\columnwidth"}
![\[fig4\]The $R_s-T$ and $n_s-T$ curves for an Sr doped NTO thin film grown on an LAO substrate at 850C in vacuum. The dash line denotes the M-I transition temperature.](HALL_2.EPS){width="0.9\columnwidth"}
[cccccccccc]{}\
No. & Sample & Denotation & P$\rm _{O}/BV$ & $R_s$(0) & $C\omega_\alpha$ & $C$ & $T_\omega$ & $\chi^{2}$ & $R^{2}$\
& & & (Pa) &($\Omega$) & ($\Omega$) & ($10^{-14}\Omega\cdot \rm S^{-1}$) & (K)\
\
1 & 260505 & LTO/STO/850 & 5e-4 & 0.024(1) & 4.47(4) & 6.0 & 284.6(9) & 0.00086 & 0.99952\
2 & 120905 & LTO/STO/650 & 4.3e-4 & 0.041(1) & 3.49(3) & 5.1 & 260.0(9) & 0.00055 & 0.99959\
3 & 230505 & LTO/LAO/850& 5e-4 & 312.0(1) & 38.8(9) & 151.7 & 97.90(8) & 1.67078 & 0.99989\
4 & 111104 & LTO/STO/780 & 4.5e-4 & 0.021(2) & 0.85(2) & 1.8 & 176.65(5) & 0.00042 & 0.99905\
5 & 191104 & LTO/STO/650 & 5e-4 & 0.053(2) & 1.78(4) & 3.0 & 226.62(5) & 0.00067 & 0.99907\
6 & 140805 & NTO/STO/850 & 6e-3 & 1.80(9) & 143. | -T$ curves for an NTO thin film grown on an STO substrate at 850C in vacuum.](HALL_1.EPS){width="0.9\columnwidth " }
! [ \[fig4\]The $ R_s - T$ and $ n_s - T$ curve for an Sr dope NTO thin film grown on an LAO substrate at 850C in void. The dash line denotes the M - I passage temperature.](HALL_2.EPS){width="0.9\columnwidth " }
[ cccccccccc]{}\
No. & Sample & Denotation & P$\rm _ { O}/BV$ & $ R_s$(0) & $ C\omega_\alpha$ & $ C$ & $ T_\omega$ & $ \chi^{2}$ & $ R^{2}$\
& & & (Pa) & ($ \Omega$) & ($ \Omega$) & ($ 10^{-14}\Omega\cdot \rm S^{-1}$) & (K)\
\
1 & 260505 & LTO / STO/850 & 5e-4 & 0.024(1) & 4.47(4) & 6.0 & 284.6(9) & 0.00086 & 0.99952\
2 & 120905 & LTO / STO/650 & 4.3e-4 & 0.041(1) & 3.49(3) & 5.1 & 260.0(9) & 0.00055 & 0.99959\
3 & 230505 & LTO / LAO/850 & 5e-4 & 312.0(1) & 38.8(9) & 151.7 & 97.90(8) & 1.67078 & 0.99989\
4 & 111104 & LTO / STO/780 & 4.5e-4 & 0.021(2) & 0.85(2) & 1.8 & 176.65(5) & 0.00042 & 0.99905\
5 & 191104 & LTO / STO/650 & 5e-4 & 0.053(2) & 1.78(4) & 3.0 & 226.62(5) & 0.00067 & 0.99907\
6 & 140805 & NTO / STO/850 & 6e-3 & 1.80(9) & 143. | -T$ ckrves for an NTO thin fium grown on an STO subvtrate at 850C in vacuum.](HALL_1.EPS){width="0.9\columnwidvh"}
![\[fit4\]The $E_s-T$ and $n_s-T$ curves for an Sr doied NTO tyin hilm grown on an LAO substrate wt 850C mn vacuum. The dssh line danotes the M-I drxndition temperature.](HALL_2.EPS){width="0.9\columgwidth"}
[cvcfcccccc]{}\
No. & Samkle & Cqnotzniin & P$\rm _{O}/BV$ & $R_s$(0) & $C\omega_\alpha$ & $C$ & $T_\omtga$ & $\chi^{2}$ & $R^{2}$\
& & & (Pa) &($\Okega$) & ($\Omega$) & ($10^{-14}\Omega\cdot \rm D^{-1}$) & (K)\
\
1 & 260505 & LTO/STO/850 & 5e-4 & 0.024(1) & 4.47(4) & 6.0 & 284.6(9) & 0.00086 & 0.99952\
2 & 120905 & OTO/SEI/650 & 4.3e-4 & 0.041(1) & 3.49(3) & 5.1 & 260.0(9) & 0.00055 & 0.99959\
3 & 230505 & LTO/LAO/850& 5e-4 & 312.0(1) & 38.8(9) & 151.7 & 97.90(8) & 1.67078 & 0.99989\
4 & 111104 & LTO/STO/780 & 4.5e-4 & 0.021(2) & 0.85(2) & 1.8 & 176.65(5) & 0.00042 & 0.99905\
5 & 191104 & LCO/STO/650 & 5e-4 & 0.053(2) & 1.78(4) & 3.0 & 226.62(5) & 0.00067 & 0.99907\
6 & 140805 & NTO/SVO/850 & 6e-3 & 1.80(9) & 143. | -T$ curves for an NTO thin film an substrate at in vacuum.](HALL_1.EPS){width="0.9\columnwidth"} ![\[fig4\]The an doped NTO thin grown on an substrate at 850C in vacuum. The line denotes the M-I transition temperature.](HALL_2.EPS){width="0.9\columnwidth"} [cccccccccc]{}\ No. & Sample & Denotation & _{O}/BV$ & $R_s$(0) & $C\omega_\alpha$ & $C$ & $T_\omega$ & $\chi^{2}$ & $R^{2}$\ & (Pa) & & ($10^{-14}\Omega\cdot \rm S^{-1}$) & (K)\ \ 1 & 260505 & LTO/STO/850 & 5e-4 & 0.024(1) 4.47(4) & 6.0 & 284.6(9) & 0.00086 & 2 & 120905 & & 4.3e-4 & 0.041(1) & & & 260.0(9) 0.00055 0.99959\ & 230505 & 5e-4 & 312.0(1) & 38.8(9) & 151.7 & 97.90(8) & 1.67078 & 0.99989\ 4 & 111104 & & 4.5e-4 & 0.85(2) 1.8 176.65(5) 0.00042 & 0.99905\ 191104 & LTO/STO/650 & 5e-4 & & 3.0 & 226.62(5) & 0.00067 & 0.99907\ & 140805 NTO/STO/850 & 6e-3 & 1.80(9) & | -T$ curves for an NTO thin film grOwn on an STO SubstRatE at 850c iN vacUum.](HaLL_1.EPS){width="0.9\colUMnwiDth"}
![\[fig4\]The $R_s-T$ and $n_s-T$ curvEs for An sR dopED NtO thiN film grOWn ON An LaO SuBstRaTE aT 850C in vAcuUm. The daSh line denoTes ThE M-I transitioN TeMperature.](HaLL_2.ePS){width="0.9\coluMnwIdth"}
[ccCcCccCCc]{}\
No. & SAmpLe & DenOtatioN & p$\rm _{O}/BV$ & $r_s$(0) & $C\omega_\aLpHA$ & $C$ & $T_\omeGA$ & $\chi^{2}$ & $R^{2}$\
& & & (Pa) &($\oMEgA$) & ($\OmeGa$) & ($10^{-14}\Omega\cdot \rm S^{-1}$) & (K)\
\
1 & 260505 & LTo/sTo/850 & 5E-4 & 0.024(1) & 4.47(4) & 6.0 & 284.6(9) & 0.00086 & 0.99952\
2 & 120905 & LTO/STO/650 & 4.3e-4 & 0.041(1) & 3.49(3) & 5.1 & 260.0(9) & 0.00055 & 0.99959\
3 & 230505 & LTO/LAO/850& 5E-4 & 312.0(1) & 38.8(9) & 151.7 & 97.90(8) & 1.67078 & 0.99989\
4 & 111104 & LTO/STo/780 & 4.5e-4 & 0.021(2) & 0.85(2) & 1.8 & 176.65(5) & 0.00042 & 0.99905\
5 & 191104 & ltO/sto/650 & 5e-4 & 0.053(2) & 1.78(4) & 3.0 & 226.62(5) & 0.00067 & 0.99907\
6 & 140805 & NtO/StO/850 & 6e-3 & 1.80(9) & 143. | -T$ curves for an NTO thin film grow n onanSTO s ubst rate at 850C in va c uum. ](HALL_1.EPS){width="0 .9\co lu m nwid t h" }
![ \[fig4\ ] Th e $R_ s- T$ an d$ n_ s-T$cur ves for an Sr dop edNT O thin filmg ro wn on an L AOsubstrate at 85 0C inva cuu m . The da sh li ne den o tes th e M-I tra ns i tion t e mperatu r e .] (HAL L_2.EPS){width="0 . 9\ c olumnwidth"}
[ccccc cc c cc ] { }\No. & Sample&Denot a tion &P $\ r m _{O } /BV$ & $R_s$( 0) & $C\ome g a_\ alpha$ & $C $ & $T_ \omeg a$ & $ \chi^{2}$ & $R^ {2}$\
& & & (Pa ) &($\Om e ga$) &($\Ome ga$ ) & ($1 0 ^{ -1 4}\ Om e ga\ c do t \ r m S ^{-1}$)&(K )\
\1 &2 6 0 5 05 & LT O/ST O/850 & 5e-4 & 0.0 24( 1) & 4.4 7(4)& 6.0 & 2 84 .6(9) & 0.0 0086&0.99952\
2 & 12 0905 & LTO/ST O/6 50 &4. 3e-4& 0.041 (1) &3.49(3) & 5.1& 26 0. 0 ( 9 )& 0.00055 & 0.9995 9\ 3 & 230505& LTO/ L AO /8 5 0& 5e-4&312 .0(1 ) & 38. 8(9) &151.7 &97.90( 8 )&1.67078 & 0.999 89 \
4 &11110 4 & L TO/STO /780 & 4 .5e-4 & 0.021(2) & 0 . 85(2) & 1.8 & 17 6 . 65 ( 5) & 0. 00042 & 0.9 9905 \
5 & 191 1 04 &L TO/ST O/650 & 5e - 4 & 0.053(2) & 1.78 (4 ) & 3. 0 & 2 26.62(5) & 0. 00067 & 0. 9 9 9 07\
6 &1408 0 5& NTO/STO/850 & 6e-3 & 1.80(9) & 143. | -T$ curves_for an_NTO thin film grown_on an_STO_substrate at_850C_in vacuum.](HALL_1.EPS){width="0.9\columnwidth"}
![\[fig4\]The $R_s-T$_and $n_s-T$ curves_for an Sr doped_NTO thin film_grown_on an LAO substrate at 850C in vacuum. The dash line denotes the M-I_transition_temperature.](HALL_2.EPS){width="0.9\columnwidth"}
[cccccccccc]{}\
No. &_Sample_&_Denotation & P$\rm _{O}/BV$ &_$R_s$(0) & $C\omega_\alpha$ & $C$_& $T_\omega$_& $\chi^{2}$ & $R^{2}$\
& & & (Pa) &($\Omega$)_&_($\Omega$) & ($10^{-14}\Omega\cdot_\rm S^{-1}$) & (K)\
\
1 & 260505 & LTO/STO/850 &_5e-4 & 0.024(1) & 4.47(4) &_6.0 & 284.6(9)_&_0.00086_& 0.99952\
2 & 120905_& LTO/STO/650 & 4.3e-4 & 0.041(1)_& 3.49(3) & 5.1 & 260.0(9)_& 0.00055 & 0.99959\
3 & 230505 &_LTO/LAO/850& 5e-4 & 312.0(1) & 38.8(9)_& 151.7 & 97.90(8) &_1.67078 &_0.99989\
4 & 111104 & LTO/STO/780_& 4.5e-4 &_0.021(2) &_0.85(2) & 1.8_& 176.65(5) & 0.00042 & 0.99905\
5_& 191104 &_LTO/STO/650 & 5e-4 & 0.053(2) &_1.78(4)_& 3.0 &_226.62(5)_&_0.00067 &_0.99907\
6 & 140805_&_NTO/STO/850 &_6e-3_& 1.80(9) & 143. |
the superpotential for this class of theories is
$$\partial_{z} W^{(N,l)}(z;\tau)= \sum_{k=0}^{N-1} e^{\frac{2\pi i l k}{N}} \left[ \zeta \left( z-\frac{2\pi}{N}(l\tau + k);\tau\right) +\frac{2 \eta_{1}k}{N}\right],$$
where $ z \in \mathcal{S}= \mathbb{C} \setminus \left\lbrace \frac{2\pi}{N}(l\tau+k)+\Lambda_{\tau}, \ k=0,....,N-1; \ \Lambda_{\tau}= 2\pi\mathbb{Z} \oplus 2\pi \tau \mathbb{Z}\right\rbrace $.\
The Weierstrass zeta function $\zeta(z;\tau)$ is defined with the conventions $\eta_{1}=\zeta(\pi;\tau)$, $ \eta_{2}=\zeta(\pi \tau;\tau)$ [@rif47]. This function is meromorphic on $\mathcal{S}$ with simple poles and simple zeroes respectively in $\frac{2\pi(l\tau+k)}{N}+\Lambda_{\tau}$ and $\frac{2\pi k}{N}+\Lambda_{\tau}$, $k=0,...,N-1$. By definition of elliptic function, the sum of the residue inside the fundamental cell is vanishing. From a physical point of view this means that the flux of the magnetic field is cancelled by that of the quasi-holes charges.\
The algebra of the abelian symmetry group of this model is generated by three operators $\sigma,A,B$, defined by the actions
$$\label{transform}
\begin{split}
& \sigma: z\longrightarrow z+ 2\pi/N ;\hspace{2cm} \partial_{z} W^{(N,l)}\left( z+2\pi/N;\tau\right) = e^{\frac{2\pi i l}{N}}\partial _{z}W^{(N,l)}\left( z;\tau\right), \\ \\
& A: z\longrightarrow z+2\pi; \hspace{2.4cm} \partial_{z} W^{(N,l)}\left( z+2\pi;\tau\right) = \partial_{z} W^{(N, | the superpotential for this class of theories is
$ $ \partial_{z } W^{(N, l)}(z;\tau)= \sum_{k=0}^{N-1 } e^{\frac{2\pi i l k}{N } } \left [ \zeta \left (z-\frac{2\pi}{N}(l\tau + k);\tau\right) + \frac{2 \eta_{1}k}{N}\right],$$
where $ z \in \mathcal{S}= \mathbb{C } \setminus \left\lbrace \frac{2\pi}{N}(l\tau+k)+\Lambda_{\tau }, \ k=0,.... ,N-1; \ \Lambda_{\tau}= 2\pi\mathbb{Z } \oplus 2\pi \tau \mathbb{Z}\right\rbrace $ .\
The Weierstrass zeta function $ \zeta(z;\tau)$ is define with the convention $ \eta_{1}=\zeta(\pi;\tau)$, $ \eta_{2}=\zeta(\pi \tau;\tau)$ [ @rif47 ]. This function is meromorphic on $ \mathcal{S}$ with simple poles and bare zeroes respectively in $ \frac{2\pi(l\tau+k)}{N}+\Lambda_{\tau}$ and $ \frac{2\pi k}{N}+\Lambda_{\tau}$, $ k=0,... ,N-1$. By definition of elliptic routine, the union of the residue inside the cardinal cellular telephone is vanishing. From a physical compass point of view this means that the flux of the charismatic field is cancelled by that of the quasi - hole charges.\
The algebra of the abelian symmetry group of this model is generated by three operator $ \sigma, A, B$, defined by the actions
$ $ \label{transform }
\begin{split }
& \sigma: z\longrightarrow z+ 2\pi / N; \hspace{2 cm } \partial_{z } W^{(N, l)}\left (z+2\pi / N;\tau\right) = e^{\frac{2\pi i l}{N}}\partial _ { z}W^{(N, l)}\left (z;\tau\right), \\ \\
& A: z\longrightarrow z+2\pi; \hspace{2.4 curium } \partial_{z } W^{(N, l)}\left (z+2\pi;\tau\right) = \partial_{z } W^{(N, | thf superpotential for thir class of theotiws is
$$\pertial_{z} W^{(N,l)}(z;\tau)= \sum_{k=0}^{N-1} e^{\frac{2\pi i l k}{N}} \left[ \zxta \oeft( e-\yrac{2\pi}{N}(l\tau + k);\tau\rigft) +\frac{2 \ena_{1}k}{N}\right],$$
qhert $ z \in \mathcal{S}= \mathbb{C} \setmihms \leyt\obrace \frac{2\pi}{N}(k\tau+k)+\Lambdd_{\tau}, \ k=0,....,N-1; \ \Lambga_{\gab}= 2\pi\mathbb{Z} \oplus 2\pi \tau \mathbb{Z}\righe\rbrace $.\
Tje Weierstrass zets fundniin $\zeta(z;\tau)$ is defined with tge convtntions $\eta_{1}=\zeta(\pi;\tsu)$, $ \eta_{2}=\zeta(\pi \tau;\tau)$ [@rif47]. Tjis vunction is meromogphic on $\majgcaj{W}$ with simpld poles and simple zerkes respectively in $\frac{2\pi(l\tau+k)}{V}+\Lambba_{\tau}$ and $\ftcx{2\pi n}{N}+\Lambda_{\tau}$, $k=0,...,N-1$. Bj definition of ellipdic funvtion, the sum pf vhe eesidue inside the fuidamental cell is vagishing. Fsoj a physical poinr if vigw thhs mdqns thzt tge flud oh the magnefic field iw cancelled by that os the quasi-holss chawgqs.\
The algebra of the abelian symmetry ggoup of this model is generqted by three operatots $\sigma,A,B$, defined by the actions
$$\label{transform}
\begin{split}
& \shgma: v\lungxlnhtafeoa z+ 2\pi/N ;\hspace{2cm} \partial_{z} W^{(N,l)}\left( z+2\pi/N;\tau\righe) = t^{\frsc{2\pi i l}{N}}\particl _{z}W^{(N,l)}\left( z;\tau\tihhy), \\ \\
& A: z\longrkghtarxkw z+2\pi; \hspace{2.4cm} \partlal_{z} W^{(N,j)}\left( z+2\pi;\tau\ridht) = \partial_{z} W^{(N, | the superpotential for this class of theories W^{(N,l)}(z;\tau)= e^{\frac{2\pi i k}{N}} \left[ \zeta \eta_{1}k}{N}\right],$$ $ z \in \mathbb{C} \setminus \left\lbrace \ k=0,....,N-1; \ \Lambda_{\tau}= 2\pi\mathbb{Z} \oplus \tau \mathbb{Z}\right\rbrace $.\ The Weierstrass zeta function $\zeta(z;\tau)$ is defined with the conventions $ \eta_{2}=\zeta(\pi \tau;\tau)$ [@rif47]. This function is meromorphic on $\mathcal{S}$ with simple poles simple respectively $\frac{2\pi(l\tau+k)}{N}+\Lambda_{\tau}$ $\frac{2\pi k}{N}+\Lambda_{\tau}$, $k=0,...,N-1$. By definition of elliptic function, the sum of the residue inside the fundamental is vanishing. From a physical point of view means that the flux the magnetic field is cancelled that the quasi-holes The of abelian symmetry group this model is generated by three operators $\sigma,A,B$, defined by the actions $$\label{transform} \begin{split} & \sigma: z\longrightarrow 2\pi/N ;\hspace{2cm} z+2\pi/N;\tau\right) = i _{z}W^{(N,l)}\left( \\ \\ & z+2\pi; \hspace{2.4cm} \partial_{z} W^{(N,l)}\left( z+2\pi;\tau\right) = | the superpotential for this cLass of theoRies iS
$$\paRtiAl_{Z} W^{(N,l)}(Z;\tau)= \Sum_{k=0}^{N-1} e^{\frac{2\pi i l K}{n}} \lefT[ \zeta \left( z-\frac{2\pi}{N}(l\tau + k);\Tau\riGhT) +\Frac{2 \ETa_{1}K}{N}\rigHt],$$
where $ Z \In \MAThcAl{s}= \mAthBb{c} \SeTminuS \leFt\lbracE \frac{2\pi}{N}(l\tAu+k)+\laMbda_{\tau}, \ k=0,....,N-1; \ \LamBDa_{\Tau}= 2\pi\mathbB{Z} \oPlus 2\pi \tau \matHbb{z}\right\RbRacE $.\
the WeIerStrasS zeta fUNction $\Zeta(z;\tau)$ iS dEFined wITh the coNVEnTionS $\eta_{1}=\zeta(\pi;\tau)$, $ \eta_{2}=\zETa(\PI \tau;\tau)$ [@rif47]. This FunctiOn IS mEROmoRphIc on $\mathcaL{S}$ With sIMple polES aND SImpLE zeroes respecTively in $\fraC{2\Pi(l\Tau+k)}{N}+\LAmBda_{\TAu}$ and $\fRac{2\pi K}{N}+\lAmbDa_{\tau}$, $k=0,...,N-1$. By deFiniTion of ellIptic fUNction, tHE sum of tHe resiDue InsIde tHE fUnDamEnTAl cELl Is vANisHing. From A pHySical PoinT OF VIew tHis MeanS that The flux of the mAgnEtic FIelD is caNcellEd by ThAt of tHe quasI-holeS cHarges.\
The algebrA of tHe abelian SymMeTry GrOup of THis modEl iS geNerated By three OPerAtORS $\SiGma,A,B$, defined by the aCtIONs
$$\Label{traNsform}
\BEgIn{SPlit}
& \sigmA: z\LonGrigHTArrow Z+ 2\pi/N ;\HSpAce{2cm} \parTial_{z} W^{(n,L)}\lEfT( z+2\pi/N;\taU\rIght) = e^{\fRaC{2\pi I l}{N}}\PartiAL _{z}W^{(N,L)}\left( z;\Tau\right), \\ \\
& a: z\lonGRightarrow z+2\pi; \hSPace{2.4cm} \partial_{Z} w^{(N,L)}\LEfT( Z+2\pi;\tAu\rIght) = \partial_{Z} W^{(N, | the superpotential for th is class o f the ori esis
$$ \par tial_{z} W^{(N , l)}( z;\tau)= \sum_{k=0}^{N -1} e ^{ \ frac { 2\ pi il k}{N} } \ l e ft[ \ ze ta\l e ft ( z-\ fra c{2\pi} {N}(l\tau+ k ); \tau\right)+ \f rac{2 \eta _{1 }k}{N}\right ],$ $
whe re $z \in \m athca l{S}=\ mathbb {C} \setm in u s \lef t \lbrace \ fr ac{2 \pi}{N}(l\tau+k)+ \ La m bda_{\tau}, \k=0,.. .. , N- 1 ; \\La mbda_{\tau }= 2\pi \ mathbb{ Z }\ o p lus 2\pi \tau \ma thbb{Z}\rig h t\r brace$. \
T h e Weie rstra ss zet a function$\ze ta(z;\tau )$ isd efinedw ith the conve nti ons $\e t a_ {1 }=\ ze t a(\ p i; \ta u )$, $ \eta_ {2 }= \zeta (\pi \ t a u;\t au) $ [@ rif47 ]. This funct ion ism ero morph ic on $\m at hcal{ S}$ wi th si mp le poles and si mple zeroes r esp ec tiv el y in$ \frac{ 2\p i(l \tau+k) }{N}+\L a mbd a_ { \ t au }$ and $\frac{2\pi k } { N} +\Lambda _{\tau } $, $ k =0,...,N -1 $.By d e f initi on o f e llipticfuncti o n, t he sumof the r es idu e i nside thefundam ental ce ll is vanishing. Fro m a physical p o in t of view th is means th at t h e fl ux o f t hem agnet ic fi el d i s cancelled by thatof the q uasi- holes charges .\
The alg e b r a of the abe l ia n symmetry grou p ofthis model is gener atedby three operator s $\sigma, A,B $,def ine d by the actions$ $\la be l{trans for m}
\beg in{ spl it}
&\s igma: z\l ongright ar ro wz+ 2\ pi/N; \hspace{ 2c m}\p art ial_{ z } W^{( N,l)} \lef t( z + 2\p i/N;\ta u \r i g ht)=e^ {\fr ac{ 2\ pi il}{N } }\p artial_{z}W^{(N ,l) } \lef t( z ;\tau\r ight), \\ \\ & A: z\lon gr igh tarrow z +2\pi; \ hspace{2.4cm} \partial_ { z} W^{( N,l )}\le ft(z+2\pi;\t au\ right) =\ partia l_{z}W^{(N , | the_superpotential for_this class of theories_is
$$\partial_{z} W^{(N,l)}(z;\tau)=_\sum_{k=0}^{N-1}_e^{\frac{2\pi i_l_k}{N}} \left[ \zeta_\left( z-\frac{2\pi}{N}(l\tau +_k);\tau\right) +\frac{2 \eta_{1}k}{N}\right],$$
where $_ z \in_\mathcal{S}=_\mathbb{C} \setminus \left\lbrace \frac{2\pi}{N}(l\tau+k)+\Lambda_{\tau}, \ k=0,....,N-1; \ \Lambda_{\tau}= 2\pi\mathbb{Z} \oplus 2\pi \tau \mathbb{Z}\right\rbrace $.\
The_Weierstrass_zeta function_$\zeta(z;\tau)$_is_defined with the conventions $\eta_{1}=\zeta(\pi;\tau)$,_$ \eta_{2}=\zeta(\pi \tau;\tau)$ [@rif47]. This_function is_meromorphic on $\mathcal{S}$ with simple poles and simple_zeroes_respectively in $\frac{2\pi(l\tau+k)}{N}+\Lambda_{\tau}$_and $\frac{2\pi k}{N}+\Lambda_{\tau}$, $k=0,...,N-1$. By definition of elliptic function,_the sum of the residue inside_the fundamental cell_is_vanishing._From a physical point_of view this means that the_flux of the magnetic field is_cancelled by that of the quasi-holes charges.\
The_algebra of the abelian symmetry group_of this model is generated_by three_operators $\sigma,A,B$, defined by the_actions
$$\label{transform}
\begin{split}
& \sigma: z\longrightarrow_z+ 2\pi/N_;\hspace{2cm} \partial_{z} W^{(N,l)}\left(_z+2\pi/N;\tau\right) = e^{\frac{2\pi i l}{N}}\partial _{z}W^{(N,l)}\left(_z;\tau\right), \\_\\
& A: z\longrightarrow z+2\pi; \hspace{2.4cm}_\partial_{z}_W^{(N,l)}\left( z+2\pi;\tau\right) =_\partial_{z}_W^{(N, |
Subsets and Splits