public override Chromosome PerformMutation(Chromosome chromosome)
{
Chromosome mutated = new Chromosome(shouldInitGenes: false);
mutated.Genes = new List <int>(chromosome.Genes);
if (Random.NextDouble() < Settings.MutationRate)
{
// Generate a random gene
//mutated.GetRandomGenes();
Shuffle(mutated.Genes);
mutated.MakeProperGenes();
//Console.WriteLine("{0}", string.Join(",", chromosome.GetReadableGenes()));
//Console.WriteLine("{0}\n", string.Join(",", mutated.GetReadableGenes()));
mutated.CalculateFitness();
}
return(mutated);
}
public void CalculateAllFitness()
{
// initialize best individual to be the first element of the Population
Chromosome best = Population[0];
best.CalculateFitness();
for (int i = 0; i < Population.Count; i++)
{
Population[i].CalculateFitness();
if (Population[i].Fitness <= best.Fitness)
{
best = Population[i];
}
}
BestFitness = best.Fitness;
BestGenes = best.Genes;
BestChromosome = best;
}
// Crossover function
public override Chromosome PerformCrossover(Chromosome firstParent, Chromosome secondParent)
{
Chromosome child = new Chromosome(shouldInitGenes: false);
if (firstParent.Genes.SequenceEqual(secondParent.Genes))
{
child.Genes = firstParent.Genes;
}
else
{
bool isFeasible = false;
int counter = 0;
while (!isFeasible)
{
// create new child Chromosome with same gene array size as parent; improve performance by setting shouldInitGenes: false
Chromosome firstChild = new Chromosome(shouldInitGenes: false);
Chromosome secondChild = new Chromosome(shouldInitGenes: false);
counter++;
if (counter > 20)
{
//Console.WriteLine(counter);
child.Genes = child.GetRandomGenes();
child.CalculateFitness();
break;
}
//Console.WriteLine("firstChild: {0}", firstChild.GetReadableGenes());
//Console.WriteLine("secondChild: {0}", secondChild.GetReadableGenes());
List <int> positionProfile = new List <int>(new int[firstParent.Genes.Count]);
double prob;
for (int i = 0; i < firstParent.Genes.Count; i++)
{
prob = Random.NextDouble();
if (prob < secondParent.Fitness / (firstParent.Fitness + secondParent.Fitness))
{
positionProfile[i] = 1;
}
}
for (int i = 0; i < positionProfile.Count; i++)
{
if (positionProfile[i] == 1)
{
firstChild.Genes[i] = firstParent.Genes[i];
}
else
{
secondChild.Genes[i] = secondParent.Genes[i];
}
}
//Console.WriteLine("positionProfile: {0}", string.Join(" ", positionProfile));
//Console.WriteLine("firstParent: {0}", firstParent.GetReadableGenes());
//Console.WriteLine("secondParent: {0}", secondParent.GetReadableGenes());
//Console.WriteLine("firstChild: {0}", firstChild.GetReadableGenes());
//Console.WriteLine("secondChild: {0}", secondChild.GetReadableGenes());
for (int i = 0; i < firstChild.Genes.Count; i++)
{
if (firstChild.Genes[i] == -1)
{
firstChild.Genes[i] = NextUnrepeatedElem(firstChild.Genes, secondParent.Genes);
}
else if (secondChild.Genes[i] == -1)
{
secondChild.Genes[i] = NextUnrepeatedElem(secondChild.Genes, firstParent.Genes);
}
}
//Console.WriteLine("firstChild: {0}", firstChild.GetReadableGenes());
//Console.WriteLine("secondChild: {0}\n", secondChild.GetReadableGenes());
firstChild.CalculateFitness();
secondChild.CalculateFitness();
if (firstChild.Fitness > secondChild.Fitness)
{
child = secondChild;
}
else
{
child = firstChild;
}
isFeasible = child.Schedule.IsOverallFeasible() && child.Fitness <= (firstParent.Fitness + secondParent.Fitness) / 2;
//if (isFeasible)
//{
// Console.WriteLine("positionProfile: {0}", string.Join(" ", positionProfile));
// Console.WriteLine("firstParent: {0}", firstParent.GetReadableGenes());
// Console.WriteLine("secondParent: {0}", secondParent.GetReadableGenes());
// Console.WriteLine("firstChild: {0}", firstChild.GetReadableGenes());
// Console.WriteLine("secondChild: {0}", secondChild.GetReadableGenes());
// Console.WriteLine("child.Fitness: {0}\n", child.Fitness);
// Console.WriteLine("if child Feasible {0}\n", isFeasible);
// Console.WriteLine(counter);
//}
}
//Console.WriteLine(counter);
}
return(child);
}
