private void DoPostSolveActivities(GeneticAlgorithmState state)
{
Console.ForegroundColor = ConsoleColor.Green;
Console.WriteLine($"Best solution found: {state.BestChromosomeOptimizationResult.HighestLambdaId}");
Console.ForegroundColor = ConsoleColor.DarkCyan;
Console.WriteLine($"Genetic Algorithm exited after {(double)state.ElapsedTime.ElapsedMilliseconds / 1000} seconds. " +
$"Generations cultured: {state.NumberOfGenerations}.");
_outputWriter.SaveOutputToTheFile(state.BestChromosomeOptimizationResult, state.BestChromosomeNetworkSolution.PathAllocations, _currentFileName);
}
private void PrintBestAlgorithmInGeneration(GeneticAlgorithmState state, bool bestResult)
{
if (bestResult)
{
Console.ForegroundColor = ConsoleColor.Green;
}
Console.WriteLine($"Best solution in generation { state.NumberOfGenerations }: " + state.BestChromosomeOptimizationResult.HighestLambdaId);
if (bestResult)
{
Console.ForegroundColor = ConsoleColor.DarkCyan;
}
}
public void Solve()
{
var state = new GeneticAlgorithmState
{
ElapsedTime = Stopwatch.StartNew(),
NumberOfGenerations = 0,
NumberOfGenerationsWithoutImprovement = 0,
NumberOfMutations = 0,
BestChromosomeOptimizationResult = null,
BestChromosomeFitness = 0
};
var random = new Random(_parameters.RandomSeed);
var population = GenerateInitialPopulation(random);
EvaluateFitness(population.Chromosomes);
var bestChromosome = population.Chromosomes.OrderByDescending(x => x.Fitness).First();
state.BestChromosomeNetworkSolution = ((NetworkSolution)bestChromosome).CloneNS();
state.BestChromosomeOptimizationResult = CalculateNetworkSolutionOptimizationResult((NetworkSolution)bestChromosome);
if (state.BestChromosomeOptimizationResult.HighestLambdaId != 0)
{
state.BestChromosomeFitness = (1000000 / state.BestChromosomeOptimizationResult.HighestLambdaId);
}
else
{
state.BestChromosomeFitness = int.MaxValue;
}
PrintBestAlgorithmInGeneration(state, true);
while (!EvaluateStoppingCriteria(state))
{
EvaluateFitness(population.Chromosomes);
var bestChromosomeInGeneration = population.Chromosomes.OrderByDescending(x => x.Fitness).First();
if (bestChromosomeInGeneration.Fitness > state.BestChromosomeFitness)
{
state.BestChromosomeNetworkSolution = ((NetworkSolution)bestChromosomeInGeneration).CloneNS();
state.BestChromosomeOptimizationResult = CalculateNetworkSolutionOptimizationResult((NetworkSolution)bestChromosomeInGeneration);
state.BestChromosomeFitness = (1000000 / state.BestChromosomeOptimizationResult.HighestLambdaId);
state.NumberOfGenerationsWithoutImprovement = 0;
PrintBestAlgorithmInGeneration(state, true);
}
else
{
state.NumberOfGenerationsWithoutImprovement++;
PrintBestAlgorithmInGeneration(state, false);
}
var eliteOffsprings = SelectEliteOffsprings(population);
var crossoveredOffsprings = CrossoverOffsprings(population, eliteOffsprings.Count, random);
var mutatedOffsprings = MutateOffsprings(crossoveredOffsprings, state, random);
EvaluateFitness(mutatedOffsprings);
population = SelectSurvivors(population, eliteOffsprings, mutatedOffsprings);
state.NumberOfGenerations++;
}
state.ElapsedTime.Stop();
DoPostSolveActivities(state);
}
private bool EvaluateStoppingCriteria(GeneticAlgorithmState state) =>
_parameters.StoppingCriteria switch
{
private List <Chromosome> MutateOffsprings(List <Chromosome> crossoveredOffsprings, GeneticAlgorithmState state, Random random)
{
List <Chromosome> mutatedOffsprings = new List <Chromosome>();
foreach (var offspring in crossoveredOffsprings)
{
var rand = random.NextDouble();
if (rand >= _parameters.MutationProbability)
{
state.NumberOfMutations++;
mutatedOffsprings.Add(offspring.Mutate(random));
}
else
{
mutatedOffsprings.Add(offspring);
}
}
return(mutatedOffsprings);
}