/*
* RapidMiner
*
* Copyright (C) 2001-2008 by Rapid-I and the contributors
*
* Complete list of developers available at our web site:
*
* http://rapid-i.com
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see http://www.gnu.org/licenses/.
*/
package com.rapidminer.operator.features.aggregation;
import java.io.Serializable;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;
import java.util.Iterator;
import java.util.List;
import com.rapidminer.operator.performance.PerformanceVector;
/**
* Performs the non-dominated sorting selection from NSGA-II.
*
* @author Ingo Mierswa
* @version $Id: AggregationNonDominatedSortingSelection.java,v 1.5 2006/04/14
* 07:47:16 ingomierswa Exp $
*/
public class AggregationNonDominatedSortingSelection implements AggregationSelection {
/**
* The comparator for aggregation individuals using the fitness values of
* the m-th criterion.
*/
private static class CriteriaComparator implements Comparator<AggregationIndividual>, Serializable {
private static final long serialVersionUID = 4188525398056332510L;
private int m;
public CriteriaComparator(int m) {
this.m = m;
}
public int compare(AggregationIndividual i1, AggregationIndividual i2) {
PerformanceVector pv1 = i1.getPerformance();
PerformanceVector pv2 = i2.getPerformance();
return (-1) * Double.compare(pv1.getCriterion(m).getFitness(), pv2.getCriterion(m).getFitness());
}
}
/** The comparator for aggregation individuals using the crowding distance. */
private static class CrowdingComparator implements Comparator<AggregationIndividual>, Serializable {
private static final long serialVersionUID = 4712425942757940295L;
public int compare(AggregationIndividual i1, AggregationIndividual i2) {
double cd1 = i1.getCrowdingDistance();
double cd2 = i2.getCrowdingDistance();
return (-1) * Double.compare(cd1, cd2);
}
}
/** The desired population size. */
private int popSize;
/** Creates a new NSGA-II selection. */
public AggregationNonDominatedSortingSelection(int popSize) {
this.popSize = popSize;
}
/** Performs non-dominated sorting selection from NSGA-II. */
public void performSelection(List<AggregationIndividual> population) {
List<List<AggregationIndividual>> ranks = new ArrayList<List<AggregationIndividual>>();
while (population.size() > 0) {
List<AggregationIndividual> rank = getNextRank(population);
ranks.add(rank);
Iterator<AggregationIndividual> i = rank.iterator();
while (i.hasNext()) {
population.remove(i.next());
}
}
population.clear();
int index = 0;
while ((index < ranks.size()) && ((population.size() + ranks.get(index).size()) <= popSize)) {
population.addAll(ranks.get(index));
index++;
}
if (population.size() < popSize) {
List<AggregationIndividual> rank = ranks.get(index);
sortByCrowdingDistance(rank);
while (population.size() < popSize) {
population.add(rank.remove(0));
}
}
}
/**
* Sorts the given rank by crowding distance (i.e. the
* "neighborhoodship" in the fitness space).
*/
private void sortByCrowdingDistance(List<AggregationIndividual> rank) {
Iterator<AggregationIndividual> f = rank.iterator();
int numberOfCriteria = 0;
while (f.hasNext()) {
AggregationIndividual current = f.next();
current.setCrowdingDistance(0.0d);
numberOfCriteria = Math.max(numberOfCriteria, current.getPerformance().getSize());
}
for (int m = 0; m < numberOfCriteria; m++) {
Comparator<AggregationIndividual> comparator = new CriteriaComparator(m);
Collections.sort(rank, comparator);
rank.get(0).setCrowdingDistance(Double.POSITIVE_INFINITY);
rank.get(rank.size() - 1).setCrowdingDistance(Double.POSITIVE_INFINITY);
for (int i = 1; i < (rank.size() - 1); i++) {
AggregationIndividual current = rank.get(i);
double currentCrowdingDistance = current.getCrowdingDistance();
AggregationIndividual afterI = rank.get(i + 1);
AggregationIndividual beforeI = rank.get(i - 1);
double afterPerformance = afterI.getPerformance().getCriterion(m).getFitness();
double beforePerformance = beforeI.getPerformance().getCriterion(m).getFitness();
current.setCrowdingDistance(currentCrowdingDistance + Math.abs(afterPerformance - beforePerformance));
}
}
Collections.sort(rank, new CrowdingComparator());
}
/** Returns a list of non-dominated individuals. */
private List<AggregationIndividual> getNextRank(List<AggregationIndividual> population) {
List<AggregationIndividual> rank = new ArrayList<AggregationIndividual>();
for (int i = 0; i < population.size(); i++) {
AggregationIndividual current = population.get(i);
rank.add(current);
boolean delete = false;
for (int j = rank.size() - 2; j >= 0; j--) {
AggregationIndividual ranked = rank.get(j);
if (isDominated(ranked, current))
rank.remove(ranked);
if (isDominated(current, ranked)) {
delete = true;
// break;
}
}
if (delete)
rank.remove(current);
}
return rank;
}
/**
* Returns true if the second performance vector is better in all fitness
* criteria than the first one (remember: the criteria should be maximized).
*/
public static boolean isDominated(AggregationIndividual i1, AggregationIndividual i2) {
PerformanceVector pv1 = i1.getPerformance();
PerformanceVector pv2 = i2.getPerformance();
double[][] performances = new double[pv1.getSize()][2];
for (int p = 0; p < performances.length; p++) {
performances[p][0] = pv1.getCriterion(p).getFitness();
performances[p][1] = pv2.getCriterion(p).getFitness();
}
boolean dominated = true;
for (int p = 0; p < performances.length; p++) {
dominated &= (performances[p][1] >= performances[p][0]);
}
boolean oneActuallyBetter = false;
for (int p = 0; p < performances.length; p++) {
oneActuallyBetter |= (performances[p][1] > performances[p][0]);
}
dominated &= oneActuallyBetter;
return dominated;
}
}