Genetic Algorithms Example
public class MaxOnesExperiment {
public static void main(String[] args) {
// Build a exemplar individual
Individual ex = new Individual();
for (int i = 0; i < 100; i++) {
ex.addGene(new BitGene()); // true or false gene
}
// Create genetic operators, selection mechanism and fitness function.
MutationOperator mu = new FixedRateMutation(0.001);
CrossoverOperator co = new OnePointCrossover();
ParentSelector ps = new RouletteWheel();
FitnessFunction ff = new MaxOnesFitnessFunction();
// Create a population of 500 individuals.
Population pop = new FixedSizePopulation(500, ex, ps, mu, co, ff, null);
// Print the average population fitness every 5 generations.
pop.addStatTracker(new AveFitnessTracker(5));
// Run the simulation for 500 generations.
for (int gen = 0; gen < 500; gen++) {
pop.nextGeneration();
}
}
private class MaxOnesFitnessFunction implements IndividualFitnessFunction {
// Fitness is given by the number of true genes.
public double evaluate(Individual ind) {
double fitness = 0;
for (Gene g : ind) {
if (((BitGene) g).getValue()) {
fitness += 1;
}
}
return fitness;
}
}
}
The example below illustrates how to create a genetic algorithm in dLife. This genetic algorithm evolves individuals with 100 1-bit genes. An individual’s fitness is given by the number of bit genes with the value true.