public override int GeneticAlgorithm(MainGraph graph, ProblemData problemData)
{
algorithmInfo = new AlgorithmInfo();
adjacencyList = AdjacencyList.GenerateList(graph);
cost = Cost.GanerateCostArray(graph, problemData); problem = problemData;
var crossover = GeneticMethod.ChosenCrossoverMethod(crossoverMethod, CrossoverProbability, dotCrossover);
var mutation = GeneticMethod.ChosenMutationMethod(mutationMethod, MutationProbability, dotMutation);
var selection = GeneticMethod.ChosenSelectionMethod(selectionMethod, CountTour, CountSelected);
ReductionForClassicGA reduction = new ReductionForClassicGA();
Chromosome bestChromosome = null;
Population startPopulation = new Population(PopulationSize, cost);
var population = startPopulation;
int stepGA = 0;
double MediumFitness = Solution.MediumFitnessPopulation(population);
RandomSelection randomSelection = new RandomSelection();
FitnessCalculation(population.populationList);
stopwatch = new Stopwatch();
stopwatch.Start();
while (stepGA < IterateSize)
{
List <Chromosome> midPopulation = selection.Selection(population);
List <Chromosome> parentPopulation = randomSelection.Selection(midPopulation, selection.indexSelectChrom);
List <Chromosome> childList = crossover.Crossover(parentPopulation);
if (childList.Count == 0)
{
continue;
}
mutation.Mutation(childList);
FitnessCalculation(childList);
reduction.Reduction(childList, randomSelection.indexTwoParant, population.populationList);
double tempMediumFitness = Solution.MediumFitnessPopulation(population);
double absFitness = Math.Abs(tempMediumFitness - MediumFitness);
MediumFitness = tempMediumFitness;
if (absFitness >= 0 && absFitness <= 1)
{
bestChromosome = population.BestChromosome();
var worstChromosome = population.WorstChromosome();
if (bestChromosome.fitness - worstChromosome.fitness <= 1 && bestChromosome.fitness - worstChromosome.fitness >= 0)
{
if (Solution.isAnswerTrue(bestChromosome, cost, problemData))
{
stopwatch.Stop();
algorithmInfo.Time = stopwatch.Elapsed;
algorithmInfo.BestFx = bestChromosome.fitness;
algorithmInfo.Steps = stepGA;
break;
}
}
}
stepGA++;
}
stopwatch.Stop();
if (stepGA == IterateSize)
{
bool answer = false;
bestChromosome = population.BestChromosome();
while (population.populationList.Count != 0)
{
if (Solution.isAnswerTrue(bestChromosome, cost, problemData))
{
algorithmInfo.Time = stopwatch.Elapsed;
algorithmInfo.BestFx = bestChromosome.fitness;
algorithmInfo.Steps = stepGA;
answer = true;
break;
}
else
{
population.populationList.Remove(bestChromosome);
if (population.populationList.Count == 0)
{
break;
}
bestChromosome = population.BestChromosome();
}
}
if (!answer)
{
bestChromosome = null;
}
}
return(Solution.Answer(cost, bestChromosome, problemData, graph));
}
/// <summary>
/// Реализация генетического алгоритма "Genitor".
/// </summary>
/// <param name="graph">Граф.</param>
/// <param name="problemData">Параметры задачи</param>
public override int GeneticAlgorithm(MainGraph graph, ProblemData problemData)
{
this.algorithmInfo = new AlgorithmInfo();
// Инициализация основных структур.
adjacencyList = AdjacencyList.GenerateList(graph);
cost = Cost.GanerateCostArray(graph, problemData); ProblemData problem = problemData;
var crossover = GeneticMethod.ChosenCrossoverMethod(crossoverMethod, 100, dotCrossover);
var mutation = GeneticMethod.ChosenMutationMethod(mutationMethod, MutationProbability, dotMutation);
var selection = GeneticMethod.ChosenSelectionMethod(selectionMethod, CountTour, CountSelected);
Chromosome bestChromosome = null;
double MediumFitness = 0;
Population startPopulation = new Population(PopulationSize, cost);
var population = startPopulation;
int stepGA = 0;
// вычесление пригодности хромосом.
for (int i = 0; i < PopulationSize; i++)
{
population.populationList[i].fitness = Fitness.FunctionTrue(cost, problemData, population.populationList[i]);
}
RandomSelection randomSelection = new RandomSelection();
MediumFitness = Solution.MediumFitnessPopulation(population);
stopwatch = new Stopwatch();
stopwatch.Start();
while (stepGA < IterateSize)
{
List <Chromosome> midPopulation = selection.Selection(population);
List <Chromosome> parentPopulation = randomSelection.Selection(midPopulation);
Chromosome child = crossover.Crossover(parentPopulation[0], parentPopulation[1]);
if (mutation != null)
{
mutation.Mutation(child);
}
// вычесление ранга хромосомы.
//population.Sort();
Ranking2(population);
// пригодность потомка
child.fitness = Fitness.FunctionTrue(cost, problemData, child);
// поиск самой худщей хромосомы
Chromosome bedChrom = null;
bedChrom = population.populationList.First();
// замена худшей хромосомы на потомка
int index = population.populationList.IndexOf(bedChrom);
population.populationList.Insert(index, child);
population.populationList.RemoveAt(index + 1);
double tempMediumFitness = Solution.MediumFitnessPopulation(population);
double absFitness = Math.Abs(tempMediumFitness - MediumFitness);
MediumFitness = tempMediumFitness;
if (absFitness >= 0 && absFitness <= 1)
{
bestChromosome = population.BestChromosome();
var worstChromosome = population.WorstChromosome();
if (bestChromosome.fitness - worstChromosome.fitness <= 1 && bestChromosome.fitness - worstChromosome.fitness >= 0)
{
if (Solution.isAnswerTrue(bestChromosome, cost, problemData))
{
stopwatch.Stop();
algorithmInfo.Time = stopwatch.Elapsed;
algorithmInfo.BestFx = bestChromosome.fitness;
algorithmInfo.Steps = stepGA;
break;
}
}
}
stepGA++;
}
stopwatch.Stop();
if (stepGA == IterateSize)
{
bool answer = false;
bestChromosome = population.BestChromosome();
while (population.populationList.Count != 0)
{
if (Solution.isAnswerTrue(bestChromosome, cost, problemData))
{
algorithmInfo.Time = stopwatch.Elapsed;
algorithmInfo.BestFx = bestChromosome.fitness;
algorithmInfo.Steps = stepGA;
answer = true;
break;
}
else
{
population.populationList.Remove(bestChromosome);
if (population.populationList.Count == 0)
{
break;
}
bestChromosome = population.BestChromosome();
}
}
if (!answer)
{
bestChromosome = null;
}
}
return(Solution.Answer(cost, bestChromosome, problemData, graph));
}