private PathChosen AsyncGACall(int numIterations, int populationSize, int crossoverPercentage, double mutationPercentage)
{
PathChosen bestPath = null;
int shortestLength = int.MaxValue;
int totalLength = 0;
DateTime timeStart = DateTime.Now;
for (int i = 0; i < numIterations; i++)
{
//Dispatcher.BeginInvoke(new Action(() =>
//{
// gaResultTB.Text = "Calculating #" + (i + 1) + "...";
//}),
//DispatcherPriority.Background);
// do this long-running process in a non-UI thread
int elitism = 10;
int tourneySize = 2;
var path = new GeneticAlgorithmEngine(populationSize, crossoverPercentage,
mutationPercentage, elitism, tourneySize).FindOptimalPath(Cities);
int pathLength = path.GetTotalLength();
totalLength += pathLength;
if (pathLength < shortestLength)
{
shortestLength = pathLength;
bestPath = new PathChosen(path); // copy constructor
}
}
var elapsed = DateTime.Now.Subtract(timeStart).TotalSeconds;
Debug.WriteLine(numIterations + " iterations took " + elapsed + " seconds ");
bestPath.AvLength = (int)(totalLength / numIterations);
//// tell the UI thread to update the results, and display the best one
//Dispatcher.BeginInvoke(new Action(() =>
//{
// if (gaDrawnPath != null)
// canvas.Children.Remove(gaDrawnPath);
// gaDrawnPath = bestPath.Draw(canvas, Brushes.Green);
// gaResultTB.Text = "Avg. path length: " + (int)(totalLength / numIterations) + " \nBest path length: " + shortestLength;
//}),
// DispatcherPriority.Background);
return(bestPath);
}
public void algorithm_loop()
{
this.output.Focus();
this.output.AppendText("START\n");
do
{
var eval_list = population.Select(x => evalSelector.Evaluate(x));
var avg = eval_list.Average();
var max = eval_list.Max();
series.Points.AddXY(it, avg);
series2.Points.AddXY(it, max);
this.output.AppendText(String.Format(" iteracja nr {0} avg:{1:0.000} max:{2:0.000}\n", it++, avg, max));
//wydzielenie podzbiorow
var planyGrup = new Dictionary<Grupa, IList<Timetable>>();
foreach (var osobnik in population)
{
var lista_grup = osobnik.Gens.Select(x => x.Grupa).Distinct();
foreach (var grupa in lista_grup)
{
Application.DoEvents();
var planGrupy = new Timetable(mutation);
IList<TimetableLocus> loci =
(from locus in osobnik.Loci
where osobnik[locus].Grupa.Equals(grupa)
select locus).ToList();
foreach (var locus in loci)
planGrupy[locus] = osobnik[locus];
IList<Timetable> lista;
if (!planyGrup.TryGetValue(grupa, out lista))
{
lista = new List<Timetable>();
planyGrup.Add(grupa, lista);
}
lista.Add(planGrupy);
}
}
var listaGrup = planyGrup.Keys.ToList();
for (int i = 0; i < listaGrup.Count; i++)
{
Application.DoEvents();
IGeneticAlgorithmEngine<Timetable> ga_grupy;
ga_grupy = new GeneticAlgorithmEngine<Timetable, TimetableLocus, Zajecia>(planyGrup[listaGrup[i]], selector, reproducer, 20);
ga_grupy.Iterate(() =>
{
//planGrupy.Value = ga_grupy.Population;
return ga_grupy.Population.Select(x => evalSelector.Evaluate(x)).Average() > 0.95;
});
planyGrup[listaGrup[i]] = ga_grupy.Population;
}
//polaczenie podzbiorow
var nowaPopulacja = new List<Timetable>();
for (int i = 0; i < population.Count; i++)
{
Application.DoEvents();
var osobnik = new Timetable(mutation);
foreach (var planGrupy in planyGrup)
{
osobnik.Concat(planGrupy.Value[i]);
}
}
} while ((population.Select(x => evalSelector.Evaluate(x)).Average() < 0.95) && !stop_button);
if (!stop_button)
{
//wypisanie wyniku
this.output.AppendText("ZAKOŃCZONO\n");
stop_Click(null, null);
stop_button = false;
}
else
{
try
{
output.AppendText("STOP\n");
}
catch (Exception){}
}
}
public void SolveWithGenetics()
{
var genetics = new GeneticAlgorithmEngine <ItemList, Knapsack>(100, 10, 10, new SolutionGenerator(), Randomizer);
ResultContents = genetics.Run(this);
}
