终于写完了遗传算法（基于WEKA平台）不过还没调试，第一次写这么多行程序，暴露了自己很多问题，也锻炼了自己的编程能力。 嗯嗯，以后要多写程序，全方位提高自己 ^_^.
package weka.Mygenetic;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.util.Random;
import weka.core.Instance;
import weka.core.Instances;
/**
* @author ZhiChao Wang
*
*/
public class Genetic {
// private int speciesNum=50;//种群数量
private double[] finness;//适应度
private static int generationNum=200;//进化代数
private double crossPro=0.25;//交叉概率
private double mutationPro=0.02;//变异概率
Random random = new Random();
instanceOpe instanceope = new instanceOpe();
int attSelectNum;
Instances instances;
public double fitness(Instance instance,Instances instances){//求当前样例和其他样例的距离
double sum=0;
double result=0;
for(int i=0;i<instances.numInstances();i++){
for(int j=0;j<instance.numAttributes();j++){
sum+=Math.pow(instance.value(j)-instances.instance(i).value(j), 2);
}
result+=sum;
}
return result/instances.numInstances();
}
public void Genetic(){
}
/**
* Select Method
*/
public void select(Instance[] instance){
double sumFitness=0;
double[] instanceFitness = new double[instances.numInstances()];
double[] instancePro = new double[instances.numInstances()];
for(int i=0;i<instances.numInstances();i++){//计算适应度
instanceFitness[i]=fitness(instance[i],instances);
sumFitness+=instanceFitness[i];//计算所有使用度总和
}
for(int i=0;i<instances.numInstances();i++){//计算适用度概率
instancePro[i]+=instanceFitness[i]/sumFitness;
}
double wheelBorder[] = new double[instances.numInstances()];
for(int i=0;i<instances.numInstances();i++){//计算轮盘赌边界
for(int j=0;j<i;j++){
wheelBorder[i]+=instancePro[i];
}
}
Random random = new Random();
Instance[] tmpInstance = new Instance[instances.numInstances()];
for(int i=0;i<instances.numInstances();i++){
double pointer = random.nextDouble();
for(int j=0;j<wheelBorder.length;j++){
if(pointer>wheelBorder[j]&&pointer<wheelBorder[j+1]){
tmpInstance[i] = instance[j];//临时保存在tmpInstance
}
}
}
instance=tmpInstance; //把选出来的样本重新赋给instance
}
/**
* Cross Method
*/
public void cross(Instance[] instance){
for (int i = 0; i < instances.numInstances(); i++) {
if (crossPro>random.nextDouble()) {
int inSelectNumOne = random.nextInt(instances
.numInstances());//随机选择第一个样例索引
int inSelectNumTwo = random.nextInt(instances
.numInstances());
Instance instanceSelectOne = instance[inSelectNumOne];//选择第一个样例
Instance instanceSelectTwo = instance[inSelectNumOne];
String strSelectOne = encode(instanceSelectOne);//把第一个样例随机属性进行编码
int attSelectNumOne = attSelectNum;//临时保存一下随机属性索引
String strSelectTwo = encode(instanceSelectTwo);//把第一个样例随机属性进行编码
int attSelectNumTwo = attSelectNum;
int crossPosition = random.nextInt(strSelectOne.length());//随机选择交叉位置
//交叉操作(多点交叉)
String newStrOne = strSelectOne.substring(0,
crossPosition – 1)
+ strSelectTwo.substring(crossPosition);
String newStrTwo = strSelectTwo.substring(0,
crossPosition – 1)
+ strSelectOne.substring(crossPosition);
int intDecodeOne = decode(newStrOne);//解码
instance[inSelectNumOne].setValue(attSelectNumOne,
intDecodeOne);//解码之后赋值
int intDecodeTwo = decode(newStrTwo);
instance[inSelectNumOne].setValue(attSelectNumTwo,
intDecodeTwo);//解码之后赋值
}
}
}
/**
* Mutation Method
*/
public void mutation(Instance[] instance){
for (int i = 0; i < instances.numInstances(); i++) {
if (mutationPro>random.nextDouble()) {
int inSelectNum = random.nextInt(instances.numInstances());//随机样例序列
Instance instanceSelect = instance[inSelectNum];//随机选择一个样例
String attString = encode(instanceSelect);
int mutPosition = random.nextInt(attString.length());//随机选择一个交叉位置
char ch = attString.charAt(mutPosition);//选出一个字符来0 或 1
if (ch == ‘0’) {
ch = ‘1’;
} else {
ch = ‘0’;
}
attString.replace(attString.charAt(mutPosition), ch);//用新的字符串代替原来的
instance[i].setValue(attSelectNum, decode(attString));//给样例重新赋属性值
}
}
}
/**
* Encode Method
*/
public String encode(Instance instance){//选出一个样例的属性进行编码
attSelectNum = random.nextInt(instance.numAttributes());//随即选择0~属性长度的位置
double getAttValue = instance.value(attSelectNum);//根据随即属性索引得到属性值
String str = Integer.toBinaryString((int)getAttValue); //把该属性转换为二进制
return str;
}
/**
* Decode Method
*/
public int decode(String str){//把String-里面包含二进制 转换成十进制
return Integer.parseInt(str, 2);
}
public static void main(String args[]) {
Genetic genetic=new Genetic();
try {
File file = new File(“文件路径”);
FileReader fr = new FileReader(file);
BufferedReader br = new BufferedReader(fr);
Instances instances = new Instances(br);
Instance[] instance = new Instance[instances.numInstances()];
for(int i=0;i<instances.numInstances();i++){
instance[i] = instances.instance(i);
}
for(int i=0;i<generationNum;i++){
genetic.select(instance);//选择
genetic.cross(instance);//交叉
genetic.mutation(instance);//变异
//1 J48 j48 = new J48();使用分类器进行分类,每一代得到一个模型
//2 利用这些模型在测试机上进行测试，得到一个正确率数组
//3 选出使正确率最高的那个模型对应的Instance[]
}
} catch (Exception e) {
e.printStackTrace();
}
}
}
