protected Population crossoverHelper(Population parents, int random_gens1, int random_gens2, int[] permutation)
{
List <Individual> offsprings1 = new List <Individual>();
List <Individual> offsprings2 = new List <Individual>();
List <int> used = new List <int>();
if (random_gens2 < random_gens1)
{
int temp = random_gens1;
random_gens1 = random_gens2;
random_gens2 = temp;
}
for (int i = 0; i < parents.Count / 2; i++)
{
Tuple <Individual, Individual> children = crossover(parents.getIndividual(permutation[i]), parents.getIndividual(permutation[parents.Count - 1 - i]), random_gens1, random_gens2);
offsprings1.Add(children.Item1);
offsprings2.Add(children.Item2);
}
offsprings2.Reverse(0, offsprings2.Count);
offsprings1.AddRange(offsprings2);
Population offspring = new Population(offsprings1);
return(offspring);
}
public Individual RunGA_GT()
{
population = new Population();
population.RandomPopulation(cheaterRate, chromosomeLength, populationSize);
maxFitness = knapsackList.MaxFitness(); //maxFitness only calculated once (sum of the values of the objects)
maxPayoff = gameModel.maxPayoff;
population.Evaluation();
for (int epoch = 0; epoch < numberOfEpochs; epoch++)
{
Population parents = population.TournamentSelection();
Population offspring = population.UniformCrossover(parents);//population.TwoPointsCrossover(parents);
offspring.Mutation();
parents.Sort();
offspring.Sort();
// Another way of choosing the new generation:
//population.Clear();
//population.AddRange(parents);
//population.AddRange(offspring);
//population.Sort();
//population.RemoveRange(populationSize, population.Count - populationSize);
//// TODO: number of cheaters should be less than a cheater rate? or some other value?
//population.Clear();
//for (int i = 0; i < populationSize; i++)
//{
// //population.Add(parents.getIndividual(i));
// population.Add(offspring.getIndividual(i));
//}
population.Clear();
int nonFeasibleAllowed = (int)(populationSize * (1.0 - feasibleRate));
int nonFeasibleInPopulation = 0;
for (int i = 0; i < offspring.Count; i++)
{
if (!offspring.getIndividual(i).isFeasible)
{
if (nonFeasibleInPopulation == nonFeasibleAllowed)
{
continue;
}
else
{
population.Add(offspring.getIndividual(i));
nonFeasibleInPopulation++;
}
}
else
{
population.Add(offspring.getIndividual(i));
}
}
if (population.Count < populationSize)
{
for (int i = 0; i < parents.Count; i++)
{
if (population.Count == populationSize)
{
break;
}
if (parents.getIndividual(i).isFeasible)
{
population.Add(parents.getIndividual(i));
}
}
}
population.GetBestFeasible().ShowValue();
population.Evaluation();
}
population.Sort();
population.Show();
Console.ReadLine();
return(population.GetBestFeasible());
}
public Population UniformCrossover(Population parents)
{
Population offspring = new Population();
List <int> unused = new List <int>();
for (int i = 0; i < parents.Count; i++)
{
unused.Add(i);
}
int chromosomeSize = getChromosomeSize();
for (int i = 0; i < parents.Count; i++)
{
if (!unused.Contains(i))
{
continue;
}
unused.Remove(i);
int randomIndex = unused.ElementAt(GA_GT.random.Next() % unused.Count);
unused.Remove(randomIndex);
if (GA_GT.random.NextDouble() >= GA_GT.crossoverRate)
{
offspring.Add(parents.getIndividual(i));
offspring.Add(parents.getIndividual(randomIndex));
}
else
{
Chromosome ch1 = new Chromosome(chromosomeSize);
Chromosome ch2 = new Chromosome(chromosomeSize);
for (int j = 0; j < chromosomeSize; j++)
{
if (GA_GT.random.NextDouble() < 0.5)
{
ch1[j] = parents.getIndividual(i)[j];
ch2[j] = parents.getIndividual(randomIndex)[j];
}
else
{
ch1[j] = parents.getIndividual(randomIndex)[j];
ch2[j] = parents.getIndividual(i)[j];
}
}
Individual ind1 = new Individual(ch1);
ind1.strategy = parents.getIndividual(i).strategy;
ind1.Update(GA_GT.knapsackList);
Individual ind2 = new Individual(ch2);
ind2.strategy = parents.getIndividual(randomIndex).strategy;
ind2.Update(GA_GT.knapsackList);
offspring.Add(ind1);
offspring.Add(ind2);
}
}
return(offspring);
}
protected Population crossoverHelper(Population parents, int random_gens1, int random_gens2, int[] permutation)
{
List<Individual> offsprings1 = new List<Individual>();
List<Individual> offsprings2 = new List<Individual>();
List<int> used = new List<int>();
if (random_gens2 < random_gens1)
{
int temp = random_gens1;
random_gens1 = random_gens2;
random_gens2 = temp;
}
for (int i = 0; i < parents.Count / 2; i++)
{
Tuple<Individual, Individual> children = crossover(parents.getIndividual(permutation[i]), parents.getIndividual(permutation[parents.Count - 1 - i]), random_gens1, random_gens2);
offsprings1.Add(children.Item1);
offsprings2.Add(children.Item2);
}
offsprings2.Reverse(0, offsprings2.Count);
offsprings1.AddRange(offsprings2);
Population offspring = new Population(offsprings1);
return offspring;
}
public Population UniformCrossover(Population parents)
{
Population offspring = new Population();
List<int> unused = new List<int>();
for (int i = 0; i < parents.Count; i++)
{
unused.Add(i);
}
int chromosomeSize = getChromosomeSize();
for (int i = 0; i < parents.Count; i++)
{
if (!unused.Contains(i))
continue;
unused.Remove(i);
int randomIndex = unused.ElementAt(GA_GT.random.Next() % unused.Count);
unused.Remove(randomIndex);
if (GA_GT.random.NextDouble() >= GA_GT.crossoverRate)
{
offspring.Add(parents.getIndividual(i));
offspring.Add(parents.getIndividual(randomIndex));
}
else
{
Chromosome ch1 = new Chromosome(chromosomeSize);
Chromosome ch2 = new Chromosome(chromosomeSize);
for (int j = 0; j < chromosomeSize; j++)
{
if (GA_GT.random.NextDouble() < 0.5)
{
ch1[j] = parents.getIndividual(i)[j];
ch2[j] = parents.getIndividual(randomIndex)[j];
}
else
{
ch1[j] = parents.getIndividual(randomIndex)[j];
ch2[j] = parents.getIndividual(i)[j];
}
}
Individual ind1 = new Individual(ch1);
ind1.strategy = parents.getIndividual(i).strategy;
ind1.Update(GA_GT.knapsackList);
Individual ind2 = new Individual(ch2);
ind2.strategy = parents.getIndividual(randomIndex).strategy;
ind2.Update(GA_GT.knapsackList);
offspring.Add(ind1);
offspring.Add(ind2);
}
}
return offspring;
}