public void ShouldComposeNewProcessInformationForNewGeneration()
{
Random rng = new Random();
for (Int32 i = 0; i < TestConstants.TestRepeats; i++)
{
Int32 populationSize = rng.Next(1, 1000);
Int32 geneSize = rng.Next(1, 1000);
Population <TestIndividual <Int32>, Int32> GetPopulation()
{
var population = new Population <TestIndividual <Int32>, Int32>();
IList <TestIndividual <Int32> > individuals = new List <TestIndividual <Int32> >();
for (Int32 j = 0; j < populationSize; j++)
{
var newIndividual = new TestIndividual <Int32>();
Chromosome <Int32> newChromosome =
ChromosomeInitializers.GetRandomChromosome <Chromosome <Int32>, Int32>(geneSize
, rng_ => ChromosomeInitializers.GetRandomInt32(rng_, 0, 10));
newIndividual.SetChromosome(newChromosome);
individuals.Add(newIndividual);
}
population.SetPopulation(individuals);
population.OrderByFitness(new PopulationTests.TestFitnessFunctionInt32());
return(population);
}
Population <TestIndividual <Int32>, Int32> population1 = GetPopulation();
Population <TestIndividual <Int32>, Int32> population2 = GetPopulation();
var composer =
new ProcessInformationComposer <Population <TestIndividual <Int32>, Int32>, TestIndividual <Int32>,
Int32>();
ProcessInformation firstProcessInformation = composer.ComposeProcessInformation(null, population1);
ProcessInformation secondProcessInformation =
composer.ComposeProcessInformation(firstProcessInformation, population2);
Assert.Equal(firstProcessInformation.Generation, secondProcessInformation.Generation - 1);
Assert.Equal(firstProcessInformation.TotalFitnessDelta, firstProcessInformation.TotalFitness);
Assert.Equal(firstProcessInformation.FittestFitnessDelta, firstProcessInformation.FittestFitness);
Assert.Equal(firstProcessInformation.FittestNFitnessDelta, firstProcessInformation.FittestNFitness);
Assert.Equal(secondProcessInformation.TotalFitnessDelta
, secondProcessInformation.TotalFitness - firstProcessInformation.TotalFitness);
Assert.Equal(secondProcessInformation.FittestFitnessDelta
, secondProcessInformation.FittestFitness - firstProcessInformation.FittestFitness);
Assert.Equal(secondProcessInformation.FittestNFitnessDelta
, secondProcessInformation.FittestNFitness - firstProcessInformation.FittestNFitness);
}
}
public void ShouldTerminateBasedOnGeneration()
{
Random rng = new Random();
Population <TestIndividual <Int32>, Int32> population = TestUtils.GetPopulation(10
, () => TestUtils.GetRandomTestIndividual(10, () => ChromosomeInitializers.GetRandomInt32(rng, -5, 5)));
population.OrderByFitness(new PopulationTests.TestFitnessFunctionInt32());
for (Int32 i = 0; i < TestConstants.TestRepeats; i++)
{
Int32 maxRun = 1000;
Int32 terminationPoint = rng.Next(1, maxRun);
var composer =
new ProcessInformationComposer <Population <TestIndividual <Int32>, Int32>, TestIndividual <Int32>,
Int32>();
var terminator =
new GenerationsCountTerminator <Population <TestIndividual <Int32>, Int32>, TestIndividual <Int32>,
Int32>((UInt64)terminationPoint);
ProcessInformation previousProcessInformation = null;
for (Int32 j = 1; j < maxRun; j++)
{
ProcessInformation newProcessInformation =
composer.ComposeProcessInformation(previousProcessInformation, population);
previousProcessInformation = newProcessInformation;
Boolean shouldTerminate = terminator.ShouldTerminate(population, newProcessInformation);
if (j >= terminationPoint)
{
Assert.True(shouldTerminate);
}
else
{
Assert.False(shouldTerminate);
}
}
}
}
public void ShouldComposeNewFirstProcessInformation()
{
Random rng = new Random();
Population <TestIndividual <Int32>, Int32> population = TestUtils.GetPopulation(10
, () => TestUtils.GetRandomTestIndividual(10, () => ChromosomeInitializers.GetRandomInt32(rng, 0, 10)));
population.OrderByFitness(new PopulationTests.TestFitnessFunctionInt32());
var composer =
new ProcessInformationComposer <Population <TestIndividual <Int32>, Int32>, TestIndividual <Int32>, Int32
>();
ProcessInformation newProcessInformation = composer.ComposeProcessInformation(null, population);
Assert.Equal((UInt64)1, newProcessInformation.Generation);
Double expectedTotalFitness = population
.GetPopulationByFitness().Select(fitness => fitness.GetFitness()).Sum();
Assert.Equal(expectedTotalFitness, newProcessInformation.TotalFitness);
}
