/// <summary>
/// Compares the performance between
/// </summary>
/// <param name="testSize">How many tests to execute.</param>
/// <param name="parameters">The parameters used to run the test.</param>
/// <returns>A tuple consisting of average relative number of evaluations and average relative success rate.</returns>
static Tuple <double, double> ComparePerformance(int testSize, double evaluationRate, ExecutionParameters parameters)
{
int totalEvaluations_optimised = 0;
int totalEvaluations_default = 0;
int successfulExecutions_optimised = 0;
int successfulExecutions_default = 0;
for (int i = 0; i < testSize; i++)
{
ExecutionParameters notOptimisedParameters = parameters;
Tuple <bool, int> results_optimised = ExecuteInstance(evaluationRate, parameters);
Tuple <bool, int> results_default = ExecuteInstance(1, notOptimisedParameters);
if (results_optimised.Item1 && results_default.Item1)
{
totalEvaluations_optimised += results_optimised.Item2;
totalEvaluations_default += results_default.Item2;
}
if (results_optimised.Item1)
{
successfulExecutions_optimised++;
}
if (results_default.Item1)
{
successfulExecutions_default++;
}
}
double relativeEvaluations = (double)totalEvaluations_optimised / (double)totalEvaluations_default;
double relativeSuccessRate = (double)successfulExecutions_optimised / (double)successfulExecutions_default;
return(new Tuple <double, double>(relativeEvaluations, relativeSuccessRate));
}
static void Main(string[] args)
{
var stopwatch = Stopwatch.StartNew();
ExecutionParameters parameters = new ExecutionParameters();
parameters.blueprintChromosome = new RealValuedChromosome(10, -4, 4, 0.5);
parameters.fitnessFunction = new GriewankFitnessFunction();
parameters.elitismRate = 0.01;
parameters.populationSize = 500;
parameters.mutationWeigth = 0.01;
parameters.nonEvaluationWeigth = 0.001;
parameters.maxEvaluations = 50000;
parameters.crossoverRate = 0.9;
parameters.mutationRate = 0.05;
parameters.selectionTechnique = new TournamentSelection(0.1, true, true);
parameters.target = 0;
parameters.targetError = 1e-3;
int testSize = 100;
for (double evaluationRate = 0.1; evaluationRate < 1; evaluationRate += 0.1)
{
Tuple <double, double> results = ComparePerformance(testSize, evaluationRate, parameters);
Trace.WriteLine("Relative evaluations: " + results.Item1 + "; Relative success rate: " + results.Item2);
}
stopwatch.Stop();
Trace.WriteLine("Elapsed time:" + stopwatch.Elapsed);
}
/// <summary>
/// Executes an instance of a genetic algorithm with the specified parameters.
/// </summary>
/// <param name="parameters"></param>
/// <returns></returns>
static Tuple <bool, int> ExecuteInstance(double evaluationRate, ExecutionParameters parameters)
{
Generation firstGeneration = new Generation(parameters.populationSize, parameters.blueprintChromosome);
GeneticAlgorithm geneticAlgorithm = new GeneticAlgorithm(evaluationRate, firstGeneration, parameters);
int nGenerations = 0;
bool success = true;
while (Math.Abs(geneticAlgorithm.CurrentGeneration.MinFitness - parameters.target) > parameters.targetError)
{
geneticAlgorithm.RunEpoch();
nGenerations++;
if (geneticAlgorithm.totalEvaluations == parameters.maxEvaluations)
{
success = false;
break;
}
}
return(new Tuple <bool, int>(success, geneticAlgorithm.totalEvaluations));
}
public GeneticAlgorithm(double evaluationRate, Generation firstGeneration, ExecutionParameters parameters) :
this(evaluationRate, firstGeneration, parameters.maxEvaluations, parameters.crossoverRate, parameters.mutationRate, parameters.selectionTechnique, parameters.fitnessFunction, parameters.elitismRate, parameters.mutationWeigth, parameters.nonEvaluationWeigth)
{
}