private void FitnessSort(GENE[] a, int st, int ed)//适应度排序
{
if (st >= ed)
{
return;
}
int i = st, j = ed;
GENE tmp = a[i];
while (i < j)
{
//while (i < j && CompareSmall(a[j].Fitness,tmp.Fitness))
while (i < j && a[j].Fitness >= tmp.Fitness)
{
j--;
}
if (i < j)
{
a[i] = a[j];
}
//while (i < j && CompareSmall(tmp.Fitness,a[i].Fitness))
while (i < j && a[i].Fitness < tmp.Fitness)
{
i++;
}
if (i < j)
{
a[j] = a[i];
}
}
a[i] = tmp;
FitnessSort(a, st, i - 1);
FitnessSort(a, i + 1, ed);
}
private void Replicate()
{
GENE[] NextPopulation = new GENE[PopulationSize]; //声明下一代种群
for (int i = 0; i < PopulationSize; i++) //对种群中每个个体进行声明
{
NextPopulation[i] = new GENE(GENE.GeneLength);
}
double[] pi = new double[PopulationSize];
double FitnessSum = 0;
//复制
FitnessSum = 0;
FitnessSort(Population, 0, PopulationSize - 1);
//采取“保留最佳精英策略”，将最优个体直接替代最差个体;
//但最差个体仍有机会参与复制
NextPopulation[0] = Population[PopulationSize - 1];
//构造轮盘赌
for (int i = 0; i < PopulationSize; i++)
{
FitnessSum += Population[i].Fitness;
}
pi[0] = Population[0].Fitness / FitnessSum;
for (int i = 1; i < PopulationSize; i++)
{
pi[i] = Population[i].Fitness / FitnessSum + pi[i - 1];
}
Random rd = new Random();
for (int i = 1; i < PopulationSize; i++)
{
double tmp = rd.NextDouble();
int copy = 0;
for (int j = 0; j < PopulationSize; j++)
{
if (tmp <= pi[j])
{
copy = j; break;
}
}
NextPopulation[i] = Population[copy];
}
for (int i = 0; i < PopulationSize; i++)
{
Population[i] = NextPopulation[i];
}
//调试
for (int i = 0; i < PopulationSize; i++)
{
Population[i].Fitness = FitnessFunc(Population[i]);
}
}
private double FitnessFunc(GENE x) //适应度函数，index为染色体编号
{
var result = network.Compute(x.GeneSerial); //计算对应函数值
double tmp = 0;
for (int i = 0; i < result.Length; i++)
{
tmp += result[i];
}
return(tmp);
}
public GA_12(string saveNetFile, int inputDimension, int outputDimension)//默认构造函数
{
SaveNetFile = saveNetFile;
InputDimension = inputDimension;
OutputDimension = outputDimension;
Good = 0.8;
PopulationSize = 50;
MutationProb = 0.001;
GA_Iter_Max = 1000;
Population = new GENE[PopulationSize];
BestPerson = new GENE[1000];
for (int i = 0; i < PopulationSize; i++)
{
Population[i] = new GENE(GENE.GeneLength);
BestPerson[i] = new GENE(GENE.GeneLength);
}
}
public GA_12(int populationSize, double mutationProb, int ga_iter_Max, string saveNetFile,
int inputDimension, int outputDimension, double good)//构造函数重载
{
SaveNetFile = saveNetFile;
InputDimension = inputDimension;
OutputDimension = outputDimension;
Good = good;
PopulationSize = populationSize;
MutationProb = mutationProb;
GA_Iter_Max = ga_iter_Max;
//CrossProb = crossProb;
Population = new GENE[PopulationSize];
BestPerson = new GENE[1000];
for (int i = 0; i < PopulationSize; i++)
{
Population[i] = new GENE(GENE.GeneLength);
BestPerson[i] = new GENE(GENE.GeneLength);
}
}
private double FitnessFunc(GENE x)//适应度函数
{
double tmp = 0;
for (int i = 0; i < GENE.GeneLength; i++)
{
tmp += x.GeneSerial[i];
}
if (tmp > C)
{
return(0);
}
double tmp1 = 0;
double alpha = 3.0;
for (int i = 0; i < GENE.GeneLength; i++)
{
tmp1 += x.GeneSerial[i] * StockData[InveseIndex[i]][3];
}
double tmp2 = C;
double beta = 3.0;
for (int i = 0; i < GENE.GeneLength; i++)
{
tmp2 -= x.GeneSerial[i];
}
double tmp3 = 0;
double gama = 1.0;
for (int i = 0; i < GENE.GeneLength; i++)
{
tmp3 += x.GeneSerial[i] * abs(StockData[InveseIndex[i]][4] - EquityRatio) / EquityRatio;
}
return(alpha * tmp1 + beta * tmp2 - gama * tmp3);
}
public GA_3(double[][] stockData, int[] inveseIndex)
{
InveseIndex = inveseIndex;
C = 100;
StockData = stockData;
PopulationSize = 50;
MutationProb = 0.001;
GA_Iter_Max = 100;
EquityRatio = 50.0;
BestAnswer = new GENE[1000];
Population = new GENE[PopulationSize];
for (int i = 0; i < PopulationSize; i++)
{
Population[i] = new GENE(GENE.GeneLength);
}
for (int i = 0; i < 1000; i++)
{
BestAnswer[i] = new GENE(GENE.GeneLength);
}
}
public GA_3(double[][] stockData, int[] inveseIndex, double c, double equityRatio,
int populationSize, double mutationPro, int ga_iter_max)//构造函数重载
{
InveseIndex = inveseIndex;
C = c;
StockData = stockData;
PopulationSize = populationSize;
MutationProb = mutationPro;
GA_Iter_Max = ga_iter_max;
EquityRatio = equityRatio;
Population = new GENE[PopulationSize];
BestAnswer = new GENE[1000];
for (int i = 0; i < PopulationSize; i++)
{
Population[i] = new GENE(GENE.GeneLength);
}
for (int i = 0; i < 1000; i++)
{
BestAnswer[i] = new GENE(GENE.GeneLength);
}
}
