gajava.txt

这是一个应用遗传算法解决的函数优化问题

    private static Random random = new Random();
    
    public Randomizer(int lower, int upper){
        if(upper <= lower){
            throw new IllegalStateException("Upper is smaller than lower!");
        }
        this.lower = lower;
        this.upper = upper;
    }
    public Double nextDouble(){
        return Double.valueOf(lower + (upper - lower) * random.nextDouble());
    }
    
    public Integer nextInteger(){
        return Integer.valueOf(lower +random.nextInt(upper - lower));
    }
    
    public char[] nextBitArray(int length){
        if(length <= 0){
            throw new IllegalStateException("Length is less than ZERO!");
        }
        char[] temp = new char[length];
        for(int i = 0; i < length ; i++){
            temp[i] = random.nextBoolean() ? '1' : '0';
        }
        return temp;
    }
}

染色体：
package edu.zsu.zouang.inheritence;

import java.util.Arrays;

import edu.zsu.zouang.util.Randomizer;

public class Chromosome implements Cloneable{
    private double fitness = -1; //代表未计算适应函数
    
    private double select = -1; // 选择概率
    
    private char[] chromo; //染色体串
    
    private Randomizer random;
    
    private int lower;
    
    private int upper;
    
    public Chromosome(int lower, int upper, int length){
        this.lower = lower;
        this.upper = upper;
        random = new Randomizer(lower, upper);
        chromo = random.nextBitArray(length);
    }

    /** *//**
     * 克隆一个染色体
     */
    public Chromosome clone(){
        Chromosome c = new Chromosome(lower,upper,chromo.length);
        char[] temp = new char[c.chromo.length];
        System.arraycopy(chromo, 0, temp, 0, chromo.length);
        c.setChromo(temp);
        return c;
    }
    public char[] getChromo() {
        return chromo;
    }

    public void setChromo(char[] chromo) {
        this.chromo = chromo;
    }

    public double getFitness() {
        return fitness;
    }

    public void setFitness(double fitness) {
        this.fitness = fitness;
    }

    public double getSelect() {
        return select;
    }

    public void setSelect(double select) {
        this.select = select;
    }
    
    
}

适应函数接口：
package edu.zsu.zouang.inheritence;

public interface FitnessCalculate {
    public double calculate(char[] chromosome);
}

本函数的适应函数，既然是求最大值，干脆用求解的函数来做适应函数
package edu.zsu.zouang.inheritence;

/** *//**
 * 计算xyz*sin(xyz)最大值的遗传函数适应值
 * 2007-4-25
 * @author Zou Ang
 * Contact <a href ="mailto:[e-mail]">Zou Ang</a>
 */
public class FunctionFitness implements FitnessCalculate {

    
    public double calculate(char[] chromosome) {
        /**//*
         * x、y、z都用8位来编码，即x,y,z都属于[0,255]
         */
        double x = 0;
        double y = 0;
        double z = 0;
        for(int i = 0; i < 8; i++){
            int j = i + 8;
            int k = i + 16;
            x = x + Math.pow(2, 7 - i) * (Integer.valueOf(chromosome[i]) - 48);
            y = y + Math.pow(2, 7 - i) * (Integer.valueOf(chromosome[j]) - 48);
            z = z + Math.pow(2, 7 - i) * (Integer.valueOf(chromosome[k]) - 48);
        }
        return x * y * z * Math.sin(x*y*z);
    }

}
种群详细信息类：
package edu.zsu.zouang.inheritence;

public class GenerationDetail {
    private double averageFitness = 0.0;
    
    private double maxFitness = 0.0;
    
    private double minFitness = 0.0;

    public double getAverageFitness() {
        return averageFitness;
    }

    public void setAverageFitness(double averageFitness) {
        this.averageFitness = averageFitness;
    }

    public double getMaxFitness() {
        return maxFitness;
    }

    public void setMaxFitness(double maxFitness) {
        this.maxFitness = maxFitness;
    }

    public double getMinFitness() {
        return minFitness;
    }

    public void setMinFitness(double minFitness) {
        this.minFitness = minFitness;
    }
    
    
}

最后是主类：
package edu.zsu.zouang.inheritence;

import java.util.ArrayList;
import java.util.List;

import edu.zsu.zouang.util.Randomizer;

public class GeneticAlgorithm {
    
    private int generation; //进化代数
    
    private int population; //种群数量
    
    private double crossoverPossibility; //繁殖概率
    
    private double mutationPossibility; //变异概率
    
    private FitnessCalculate calculator = new FunctionFitness(); //适应函数计算
    
    private List<Chromosome> clist = new ArrayList<Chromosome>();
    
    private Randomizer random1; //随机数生成器1,用于生成变异位和交配位
    
    private Randomizer random2 = new Randomizer(0,1); //随机数生成器2，用于生成0-1之间的概率
    
    private GenerationDetail detail = new GenerationDetail();
    
    public GeneticAlgorithm(int population, double sp, double cp, double mp,int length){
        this.population = population;
        this.crossoverPossibility = cp;
        this.mutationPossibility = mp;
        random1 = new Randomizer(0,length - 1);
        generatePopulation(0,255,length); //用24位表示一组x,y,z的值
    }
    
    /** *//**
     * 生成初始种群
     * @param lower
     * @param upper
     * @param length
     */
    private void generatePopulation(int lower, int upper, int length){
        //随机生成染色体
        for(int i = 0; i < population; i ++){
            clist.add(new Chromosome(lower,upper,length));
        }
        //计算染色体的适应值
        evaluate();
    }
    
    /** *//**
     * 计算群体的适应值
     */
    private void evaluate(){
        double sum = 0.0;
        double min = Double.MAX_VALUE;
        double max = Double.MIN_VALUE;
        for(Chromosome c : clist){
            double fitness = calculator.calculate(c.getChromo());
            if(fitness > max){
                max = fitness;
            }
            if(fitness < min){
                min = fitness;
            }
            c.setFitness(fitness);
            sum += fitness;
        }
        detail.setMaxFitness(max);
        detail.setMinFitness(min);
        detail.setAverageFitness(sum/population);
        for(Chromosome c : clist){
            c.setSelect((c.getFitness())/sum);    //设置选择概率
        }
    }
    
    /** *//**
     * 在后代中选择新种群
     */
    private void selectPopulation(){
        List<Chromosome> tempList = new ArrayList<Chromosome>();
        for(Chromosome c : clist){
            long expectation = Math.round(c.getSelect() * population);
            for(int i = 0; i < expectation; i ++){
                tempList.add(c.clone());
            }
        }
        //如果选择种群数量大于种群规定数量，则淘汰适应值最小的染色体
        while(tempList.size() > population){
            int location = 0;
            double min = tempList.get(0).getFitness();
            for(int i = 0 ; i < tempList.size(); i ++){
                if(tempList.get(i).getFitness() < min){
                    location = i;
                }
            }
            tempList.remove(location);
        }
        //如果选择种群数量小于种群规定数量，则加入适应值最大的染色体
        while(tempList.size() < population){
            int location = 0;
            double max = tempList.get(0).getFitness();
            for(int i = 0; i < tempList.size(); i++){
                if(tempList.get(i).getFitness() > max){
                    location = i;
                }
            }
            tempList.add(tempList.get(location).clone());
        }
        clist = tempList;
    }
    
    /** *//**
     * 交配两个染色体
     * @param c1
     * @param c2
     * @param location
     */
    private void crossover(Chromosome c1, Chromosome c2){
        if(c1.getChromo().length != c2.getChromo().length){
            throw new IllegalStateException("染色体长度不同!");
        }
        //交换染色体上的基因
        //随机确定交配位
        int location = random1.nextInteger();
        for(int i = location ; i < c1.getChromo().length; i ++){
            char temp;
            temp = c1.getChromo()[i];
            c1.getChromo()[i] = c2.getChromo()[i];
            c2.getChromo()[i] = temp;
        }
    }
    
    /** *//**
     * 交配整个种群,完成一次进化
     *
     */
    public void crossoverPopulation(){
        for(int j = 0; j < population; j ++){
            for(int i = 0; i < population - 1; i++){
                double temp = random2.nextDouble();
                if(temp < crossoverPossibility){//在交配概率之内
                    crossover(clist.get(i), clist.get(i + 1));
                }
            }
            double mutation = random2.nextDouble();
            if(mutation < mutationPossibility){//在变异概率之内
                mutation(clist.get(j));
            }
        }
        //重新计算群体适应值
        evaluate();
        //重新选择种群
        selectPopulation();
        generation ++;
    }
    
    //随机变异一位
    private void mutation(Chromosome ch){
        int location = random1.nextInteger();
        char c = ch.getChromo()[location];
        if(c == '1'){
            ch.getChromo()[location] = '0';
        }else{
            ch.getChromo()[location] = '1';
        }
    }
    
    public void printDetail(){
        System.out.println("/*****************************************");
        System.out.println("*           -----遗传算法报告-----                 ");
        System.out.println("*  当前是第" + generation + "代");
        System.out.println("*  当前种群平均适应值为: " + detail.getAverageFitness());
        System.out.println("*  其中最大适应值为:" + detail.getMaxFitness());
        System.out.println("*  其中最小适应值为:" + detail.getMinFitness());
        System.out.println("*******************************************/");
    }
    
    public static void main(String[] args){
        //种群数量：30
        //选择概率：0.0
        //交配概率：0.9
        //变异概率: 0.01
        GeneticAlgorithm ga = new GeneticAlgorithm(30,0.0,0.9,0.01,24);
        for(int i = 0; i < 5; i ++){
            ga.crossoverPopulation();
        }
        ga.printDetail();
        for(int j = 0; j < 25; j++){
            ga.crossoverPopulation();
        }
        ga.printDetail();
        for(int k = 0; k < 1000; k++){
            ga.crossoverPopulation();
        }
        ga.printDetail();
    }
}

输出的结果是
/** *//*****************************************
*           -----遗传算法报告-----                 
*  当前是第5代
*  当前种群平均适应值为: 7592451.0488077225
*  其中最大适应值为:9031331.437029611
* 对应最大适应值的染色体为:111011101101010010111000
*  其中最小适应值为:-1.2521097155031694E7
* 其中对应最小适应值的染色体为111011101101010011111011
*******************************************/
/** *//*****************************************
*           -----遗传算法报告-----                 
*  当前是第30代
*  当前种群平均适应值为: 8785113.746617064
*  其中最大适应值为:9836674.727850026
* 对应最大适应值的染色体为:111111101101010010111000
*  其中最小适应值为:-1.1676031572464054E7
* 其中对应最小适应值的染色体为111011101101010011111000
*******************************************/
/** *//*****************************************
*           -----遗传算法报告-----                 
*  当前是第1030代
*  当前种群平均适应值为: 8763072.724911185
*  其中最大适应值为:9836674.727850026
* 对应最大适应值的染色体为:111111101101010010111000
*  其中最小适应值为:-9580464.117842415
* 其中对应最小适应值的染色体为111101101101010010111000
*******************************************/



遗传算法的JAVA实现4 
2008-06-19 01:20:33 本文已公布到博客频道校园·教育分类 
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
public class Population {
private ArrayList m_chromosomes = new ArrayList();
private Chromosome m_fittestChromosome=null;
private