public void NewGeneration()
{
int finalCount = Settings.PopulationSize + Settings.NumNewDNA;
List <Chromosome> newPopulation = new List <Chromosome>();
if (finalCount <= 0)
{
return;
}
if (Population.Count > 0)
{
CalculateAllFitness();
Population = Population.OrderBy(chrmsm => chrmsm.Fitness).ToList();
}
for (int i = 0; i < finalCount; i++)
{
// Keep only top individuals of the previous generation
if (i < Elitism && i < Settings.PopulationSize)
{
newPopulation.Add(Population[i]);
//Console.WriteLine("Generation: " + Generation);
//Console.WriteLine("/n");
//Console.WriteLine("genes: [{0}]", string.Join(",\t", Population[i].genes));
//Console.WriteLine(Population[i].fitness);
}
else if (i < Settings.PopulationSize)
{
Chromosome parent1 = Selection.PerformSelection(Population);
Chromosome parent2 = Selection.PerformSelection(Population);
Chromosome child = Crossover.PerformCrossover(parent1, parent2);
child = Mutation.PerformMutation(child);
newPopulation.Add(child);
}
else if (Settings.EnableCrossoverNewDNA)
{
newPopulation.Add(new Chromosome(shouldInitGenes: true));
}
}
Population = newPopulation;
Generation++;
}
public virtual void PerformIteration()
{
foreach (var chromosome in Population)
{
chromosome.FittnessValue = FittnessFunction.Calculate(chromosome);
}
var selection = Selection.Select(Population);
var crossover = Crossover.Perform(selection);
foreach (var chromosome in crossover)
{
Mutation.Mutate(chromosome);
}
Population.Remove(Population.FirstOrDefault(
x => x.FittnessValue == Population.Min(y => y.FittnessValue)));
Population.Remove(Population.FirstOrDefault(
x => x.FittnessValue == Population.Min(y => y.FittnessValue)));
Population.AddRange(crossover);
}
public GeneticAlgorithm(Selection <T> selectionMethod, Crossover <T> crossoverMethod, Mutation <T> mutationMethod, FitnessFunction <T> fitnessFunction, CreateNewSolution <T> createNewSolution, AcceptResult <T> acceptResult, int population, double mutationProbability, int selectionCount, int eliteCount = 0)
{
if (population < 2)
{
throw new Exception("Population must be greater or equal 2.");
}
if (eliteCount >= population)
{
throw new Exception("EliteCount must be less than population.");
}
selection = selectionMethod;
crossover = crossoverMethod;
mutation = mutationMethod;
fitness = fitnessFunction;
createSolution = createNewSolution;
accept = acceptResult;
maxPopulation = population;
this.mutationProbability = mutationProbability;
this.selectionCount = selectionCount;
this.eliteCount = eliteCount;
}
private void start_Click(object sender, EventArgs e)
{
ReadFile.Read();
Program.itr = 1000;
Program.pM = Double.Parse(mutationRate.Text, culture);
Program.pC = Double.Parse(crossoverRate.Text, culture);
int tryCount = Convert.ToInt32(repeatTime.Text);
Program.size = Program.weight.Count;
Program.n = Convert.ToInt32(populationSize.Text);
for (int k = 0; k < tryCount; k++)
{
double[] fit = new double[Program.n]; // (fx/overall) or fx values
double[] norm = new double[Program.n]; // fitness values get norm
int[] choosenId = new int[Program.n]; // choose for new generation
int bestId = 0;
int worstId = 0;
int bestFx = 0;
int max;
int index;
Stopwatch stopw = new Stopwatch();
stopw.Start();
// Generate Population
Population population = new Population();
bool gotcha = true;
Population newGeneration = new Population(gotcha);
for (int i = 0; i < Program.itr; i++)
{
// Normalize
norm = Population.Normalize(population.curPopulation);
// Select from population
switch (selectionMethod.SelectedItem.ToString())
{
case "Tournament Selection":
choosenId = Selection.TournamentSelection(norm);
break;
case "Roulet Selection":
choosenId = Selection.RouletChark(norm);
break;
}
// Keep best
bestId = Population.GetBest(norm);
newGeneration.curPopulation = Population.CopyToNew(choosenId, population.curPopulation, newGeneration.curPopulation);
switch (crossoverMethod.SelectedItem.ToString())
{
case "One Point Crossover":
newGeneration.curPopulation = CrossingOver.Crossover(newGeneration.curPopulation, 1);
break;
case "Two Point Crossover":
newGeneration.curPopulation = CrossingOver.Crossover(newGeneration.curPopulation, 2);
break;
}
// Mutation
newGeneration.curPopulation = Mutation.Mutate(newGeneration.curPopulation);
norm = Population.Normalize(newGeneration.curPopulation);
// Get worst
worstId = Population.GetWorst(norm);
// Copy the best element to new gen
Array.Copy(population.curPopulation[bestId].genes, newGeneration.curPopulation[worstId].genes, newGeneration.curPopulation[worstId].genes.Length);
newGeneration.curPopulation[worstId].calcFitness();
Chromosome.Fitness(newGeneration.curPopulation);
// Copy new generation
max = -1;
for (int p = 0; p < population.curPopulation.Count; p++)
{
Array.Copy(newGeneration.curPopulation[p].genes, population.curPopulation[p].genes, population.curPopulation[p].genes.Length);
population.curPopulation[p].calcFitness();
if (population.curPopulation[p].value > max)
{
max = population.curPopulation[p].value;
index = p;
}
}
if (max > bestFx)
{
bestFx = max;
}
}
Console.WriteLine(bestFx);
stopw.Stop();
TimeSpan ts = stopw.Elapsed;
string elapsedTime = String.Format("{0:00}:{1:00}:{2:00}.{3:00}", ts.Hours, ts.Minutes, ts.Seconds, ts.Milliseconds);
Console.WriteLine("RunTime " + elapsedTime);
Console.WriteLine("---------------------");
}
Console.WriteLine("Process Completed...");
}
private void RunAll_Click(object sender, EventArgs e)
{
string[] txts = new string[] { "test1.txt", "test2.txt", "test3.txt", "test4.txt", "test5.txt", "test6.txt", "test7.txt", "test8.txt", "test9.txt", "test10.txt" };
foreach (string k in txts)
{
ReadFile.path = "../../../../Datasets/" + k;
Console.WriteLine(ReadFile.path);
ReadFile.Read();
List <int> bests = new List <int>();
List <double> times = new List <double>();
Program.itr = 1000;
Program.pM = Double.Parse(mutationRate.Text, culture);
Program.pC = Double.Parse(crossoverRate.Text, culture);
int tryCount = Convert.ToInt32(repeatTime.Text);
Program.size = Program.weight.Count;
Program.n = Convert.ToInt32(populationSize.Text);
for (int t = 0; t < tryCount; t++)
{
double[] fit = new double[Program.n]; // (fx/overall) or fx values
double[] norm = new double[Program.n]; // fitness values get norm
int[] choosenId = new int[Program.n]; // choose for new generation
int bestId = 0;
int worstId = 0;
int bestFx = 0;
int max;
int index;
Stopwatch stopw = new Stopwatch();
stopw.Start();
// Create Population
Population population = new Population();
bool gotcha = true;
Population newGeneration = new Population(gotcha);
for (int i = 0; i < Program.itr; i++)
{
// Normalize
norm = Population.Normalize(population.curPopulation);
switch (selectionMethod.SelectedItem.ToString())
{
case "Tournament Selection":
choosenId = Selection.TournamentSelection(norm);
break;
case "Roulet Selection":
choosenId = Selection.RouletChark(norm);
break;
}
// Keep best of population
bestId = Population.GetBest(norm);
newGeneration.curPopulation = Population.CopyToNew(choosenId, population.curPopulation, newGeneration.curPopulation);
switch (crossoverMethod.SelectedItem.ToString())
{
case "One Point Crossover":
newGeneration.curPopulation = CrossingOver.Crossover(newGeneration.curPopulation, 1);
break;
case "Two Point Crossover":
newGeneration.curPopulation = CrossingOver.Crossover(newGeneration.curPopulation, 2);
break;
}
// Mutation
newGeneration.curPopulation = Mutation.Mutate(newGeneration.curPopulation);
norm = Population.Normalize(newGeneration.curPopulation);
worstId = Population.GetWorst(norm);
// Copy the best of prev generation to new generation
Array.Copy(population.curPopulation[bestId].genes, newGeneration.curPopulation[worstId].genes, newGeneration.curPopulation[worstId].genes.Length);
newGeneration.curPopulation[worstId].calcFitness();
Chromosome.Fitness(newGeneration.curPopulation);
// Copy the new generation to old generation
max = -1;
for (int p = 0; p < population.curPopulation.Count; p++)
{
Array.Copy(newGeneration.curPopulation[p].genes, population.curPopulation[p].genes, population.curPopulation[p].genes.Length);
population.curPopulation[p].calcFitness();
if (population.curPopulation[p].value > max)
{
max = population.curPopulation[p].value;
index = p;
}
}
if (max > bestFx)
{
bestFx = max;
}
}
stopw.Stop();
Console.WriteLine(k.Split('.')[0] + " Try:" + (t + 1) + " " + bestFx + " " + stopw.ElapsedMilliseconds + " ms");
bests.Add(bestFx);
times.Add(stopw.ElapsedMilliseconds);
}
//Optional write file
//ReadFile.Write(bests,times,k);
}
Console.WriteLine("Process Completed...");
}