|
import java.util.ArrayList; |
|
import java.util.Random; |
|
|
|
public class GA { |
|
public static ArrayList<ArrayList<Integer>> crossover(Individual a, Individual b, int[][] imgArray){ |
|
ArrayList<ArrayList<Integer>> newRoots = new ArrayList<>(); |
|
|
|
for (int i = 0; i < a.getSegments().size(); i++) { |
|
|
|
ArrayList<Integer> currRootCoords = new ArrayList<>(); |
|
|
|
|
|
int newX = (a.getSegments().get(i).getRootNode().getX() + b.getSegments().get(i).getRootNode().getX() ) / 2; |
|
int newY = (a.getSegments().get(i).getRootNode().getY() + b.getSegments().get(i).getRootNode().getY() ) / 2; |
|
|
|
currRootCoords.add(newX); |
|
currRootCoords.add(newY); |
|
|
|
newRoots.add(currRootCoords); |
|
|
|
} |
|
|
|
return newRoots; |
|
} |
|
|
|
public static Individual tournamentSelection(Individual a, Individual b){ |
|
if (a.getRank() == b.getRank()){ |
|
return a.getCrowdingDistance() > b.getCrowdingDistance() ? a : b; |
|
} else { |
|
return a.getRank() > b.getRank() ? a: b; |
|
} |
|
|
|
} |
|
|
|
public static ArrayList<ArrayList<ArrayList<Integer>>> mutate(ArrayList<ArrayList<ArrayList<Integer>>> children, int[][] imgArray){ |
|
Random r = new Random(); |
|
|
|
for (int i = 0; i < children.size(); i++) { |
|
for (int j = 0; j < children.get(i).size(); j++) { |
|
if(r.nextDouble() < 0.8){ |
|
int newX = r.nextInt(imgArray[0].length); |
|
int newY = r.nextInt(imgArray.length); |
|
|
|
children.get(i).get(j).set(0, newY); |
|
children.get(i).get(j).set(1, newX); |
|
|
|
} |
|
} |
|
} |
|
|
|
return children; |
|
} |
|
|
|
|
|
public static ArrayList<ArrayList<ArrayList<Integer>>> doGA(int[][] imgArray, Population parentPopulation, int numIndividuals){ |
|
ArrayList<Individual> parents = parentPopulation.getIndividuals(); |
|
ArrayList<ArrayList<ArrayList<Integer>>> children = new ArrayList<>(); |
|
|
|
for (int i = 0; i < numIndividuals * 3; i++) { |
|
Random r = new Random(); |
|
|
|
|
|
Individual crossover_individual_a = GA.tournamentSelection(parents.get(r.nextInt(parents.size())), parents.get(r.nextInt(parents.size()))); |
|
Individual crossover_individual_b = GA.tournamentSelection(parents.get(r.nextInt(parents.size())), parents.get(r.nextInt(parents.size()))); |
|
|
|
children.add((GA.crossover(crossover_individual_a, crossover_individual_b, imgArray))); |
|
} |
|
|
|
children = mutate(children, imgArray); |
|
|
|
return children; |
|
} |
|
} |
|
|