public double Evaluate(IChromosome chromosome)
{//Play n games vs a random (to be Neural Net), % win is Fitness
FloatingPointChromosome myChromosome = (FloatingPointChromosome)chromosome;
double[] genes = myChromosome.ToFloatingPoints();
double fitness = 0;
double wonCount = 0;
object wonCountLock = new object();
for (int index = 0; index < TEST_COUNT; index++)
{
//Parallel.For(0, TEST_COUNT, new ParallelOptions() { MaxDegreeOfParallelism = 2},
// (index) => {
BoardStates player = (index % 2 == 0) ? BoardStates.black : BoardStates.white;
OthelloGame othelloGame = new OthelloGame();
IEvaluationAgent heurAgent = new HeuristicAgent(genes);
while (!othelloGame.GameComplete)
{
if (othelloGame.WhosTurn == player)
{
othelloGame.MakeMove(heurAgent.MakeMove(othelloGame, player));
}
else
{
othelloGame.MakeMove(opposingAgent.MakeMove(othelloGame, ~player));
}
}
if (othelloGame.GameComplete)//just gotta check
{
if (othelloGame.FinalWinner == player)
{
lock (wonCountLock)
{
wonCount++;
}
}
else if (othelloGame.FinalWinner == BoardStates.empty)
{
lock (wonCountLock)
{
wonCount += .5;
}
}
}
else
{
throw new Exception("EvaluationFitness didn't complete a game");
}
// });
}
fitness = (double)wonCount / TEST_COUNT;
//System.Diagnostics.Debug.WriteLine("Fitness: " + fitness);
return(fitness);
}
public void ToFloatingPoints_NoArgs_Double()
{
RandomizationProvider.Current = Substitute.For <IRandomization>();
RandomizationProvider.Current.GetDouble(0, 10).Returns(1);
RandomizationProvider.Current.GetDouble(1, 11).Returns(2);
RandomizationProvider.Current.GetDouble(2, 12).Returns(3);
var target = new FloatingPointChromosome(new double[] { 0, 1, 2 }, new double[] { 10, 11, 12 }, new int[] { 8, 8, 8 }, new int[] { 0, 0, 0 });
var actual = target.ToFloatingPoints();
Assert.AreEqual(3, actual.Length);
Assert.AreEqual(1, actual[0]);
Assert.AreEqual(2, actual[1]);
Assert.AreEqual(3, actual[2]);
}
public void ToFloatingPoints_NoArgs_Double()
{
RandomizationProvider.Current = MockRepository.GenerateMock <IRandomization>();
RandomizationProvider.Current.Expect(r => r.GetDouble(0, 10)).Return(1);
RandomizationProvider.Current.Expect(r => r.GetDouble(1, 11)).Return(2);
RandomizationProvider.Current.Expect(r => r.GetDouble(2, 12)).Return(3);
var target = new FloatingPointChromosome(new double[] { 0, 1, 2 }, new double[] { 10, 11, 12 }, new int[] { 8, 8, 8 }, new int[] { 0, 0, 0 });
var actual = target.ToFloatingPoints();
Assert.AreEqual(3, actual.Length);
Assert.AreEqual(1, actual[0]);
Assert.AreEqual(2, actual[1]);
Assert.AreEqual(3, actual[2]);
}
public static ProductMatchConfig GetMatchConfig(this FloatingPointChromosome fpc)
{
var values = fpc.ToFloatingPoints();
return(new DefaultProductMatchConfig
{
NameMetricConfig =
{
ABCompoundPositionalWeightRatio = values[0],
APositionalWeightingCoefficientPower = values[1],
BPositionalWeightingCoefficientPower = values[2]
}
});
}
public override void ConfigGA(GeneticSharp.Domain.GeneticAlgorithm ga)
{
var latestFitness = 0.0;
Context.GA.GenerationRan += (object sender, EventArgs e) =>
{
m_bestChromosome = Context.GA.BestChromosome as FloatingPointChromosome;
if (m_bestChromosome != null)
{
lock (m_positions)
{
var fitness = m_bestChromosome.Fitness.Value;
var points = m_bestChromosome.ToFloatingPoints();
m_positions.Add(new KeyValuePair <double, double[]>(fitness, points));
if (fitness != latestFitness)
{
latestFitness = fitness;
var phenotype = m_bestChromosome.ToFloatingPoints();
Console.WriteLine(
"Generation {0,2}: ({1},{2}),({3},{4}) = {5}",
ga.GenerationsNumber,
phenotype[0],
phenotype[1],
phenotype[2],
phenotype[3],
fitness
);
}
}
}
};
}
private static void SaveChromosome(FloatingPointChromosome chromosome, string chromosomeLabel, string path)
{
string[] genes = chromosome.ToFloatingPoints().Select((gene) => gene.ToString()).ToArray();
System.IO.File.WriteAllLines(path + @chromosomeLabel, genes);
}
public static wynikGA GeneticExt
(int counter,
Item[] dane,
double back_volume,
int populationSize = 500,
ICrossover crossover = null,
double crossoverChance = 0.20f,
double mutationChance = 0.05f,
int tournamentSize = 50)
{
double[] minValue = new double[counter];
double[] maxValue = new double[counter];
int[] totalBits = new int[counter];
int[] fractionDigits = new int[counter];
for (int i = 0; i < counter; i++)
{
minValue[i] = 0;
maxValue[i] = 1;
totalBits[i] = 16;
fractionDigits[i] = 4;
}
ISelection selection = new TournamentSelection(tournamentSize > populationSize ? populationSize : tournamentSize, true);
if (crossover == null)
{
crossover = new TwoPointCrossover();
}
IMutation mutation = new FlipBitMutation();
var chromosome = new FloatingPointChromosome(minValue, maxValue, totalBits, fractionDigits);
// var fitobj = new MyProblemFitness(dane, chromosome, back_volume);
//var fitness = fitobj.Evaluate;
IPopulation population = new Population(populationSize, populationSize, chromosome);
IFitness fitness = new FuncFitness((c) =>
{
FloatingPointChromosome fc = c as FloatingPointChromosome;
//var _items = dane;
//var _bvol = back_volume;
/* var values = fc.ToFloatingPoints();
* var x1 = values[0];
* var y1 = values[1];
* var x2 = values[2];
* var y2 = values[3];
* return Math.Sqrt(Math.Pow(x2 - x1, 2) + Math.Pow(y2 - y1, 2));
*
*/
//chromosome.ReplaceGenes(0, chromosome.GetGenes());
double[] proc_wag = fc.ToFloatingPoints();
double suma_spakowanych = 0;
double wartosc_spakowanych = 0;
for (int i = 0; i < proc_wag.Length; i++)
{
suma_spakowanych += dane[i].item_volume * proc_wag[i];
wartosc_spakowanych += dane[i].item_value * proc_wag[i];
}
if (suma_spakowanych <= back_volume)
{
double temp = (wartosc_spakowanych * (suma_spakowanych / back_volume));
if (temp < 0)
{
if (Experiments.doOutput)
{
System.Console.WriteLine("LOL ");
}
}
}
return((suma_spakowanych <= back_volume) ? (wartosc_spakowanych * (suma_spakowanych / back_volume)) : (-suma_spakowanych));
});
GeneticAlgorithm ga = new GeneticAlgorithm(population, fitness, selection, crossover, mutation);
ga.Termination = new GenerationNumberTermination(100);
if (Experiments.doOutput)
{
System.Console.WriteLine("GA running...");
}
ga.MutationProbability = 0.05f;
ga.CrossoverProbability = 0.20f;
//ga.Selection.
ga.Start();
// FITNESS if (Experiments.doOutput) System.Console.WriteLine("Best solution found has fitness = " + ga.BestChromosome.Fitness);
wynikGA w = new wynikGA();
w.ga = ga;
w.ch = chromosome;
w.ch.ReplaceGenes(0, ga.BestChromosome.GetGenes());
return(w);
}