static bool HandleCommandLineArgs(Game game, IEnumerable<string> args)
{
string fitnessType = "", crossoverType = "";
bool status = true;
options = new OptionSet
{
{"m|mutation=", "The mutation rate (0-1)", (double v) => mutationRate = v},
{"s|crossover=", "The crossover rate (0-1)", (double v) => crossoverRate = v},
{"e|elitism=", "The elitism rate (0-1)", (double v) => elitismRate = v},
{"c|crcount=", "The number of chromosomes per population (>1)", (int v) => chromosomeCount = v},
{"fitness=", "The fitness calculator [sum | levenshtein | hamming]", v => fitnessType = v},
{"ctype=", "The crossover type [one | two ]", v => crossoverType = v},
{"t|truncate=", "The rate of the chromosomes to keep from a population before advancing (0 < t <= 1)", (double v) => truncationRate = v},
{"?|h|help", "Show help", v => { status = false; }},
//{"<>", v => target = v} // this can be used for a seed
};
try
{
options.Parse(args);
}
catch (OptionException ex)
{
status = false;
}
fitness = (FitnessCalculator)Activator.CreateInstance(GetOperationType(fitnessTypes, fitnessType), new object[] { game });
crossover = (ICrossover)Activator.CreateInstance(GetOperationType(crossoverTypes, crossoverType), new object[] { fitness });
mutation = new SinglePointMutation(new ReplayCharacterSet(), fitness);
selection = new RouletteWheelSelection();
if (mutationRate > 1 || crossoverRate > 1 || elitismRate > 1 || chromosomeCount <= 1 || truncationRate <= 0 || truncationRate > 1)
{
status = false;
}
return status;
}
public HighestScoreFitnessCalculator(Game game)
{
this.game = game;
}
static void Main(string[] args)
{
const int replayLength = 750;
//IReplayGenerator replayGenerator = new FromStringReplay("AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA A");
IReplayGenerator replayGenerator = new RandomReplayGenerator(new ReplayCharacterSet(), replayLength);
var game = new Game();
if (!HandleCommandLineArgs(game, args))
{
ShowHelp(options);
return;
}
Console.WriteLine("Mutation Rate: {0}%", mutationRate * 100);
Console.WriteLine("Crossover Rate: {0}%", crossoverRate * 100);
Console.WriteLine("Truncation Rate: {0}%", truncationRate * 100);
Console.WriteLine("Chromosomes / population: {0}", chromosomeCount);
Console.WriteLine("Elitism / population: {0} ({1}%)", elitismRate * chromosomeCount, elitismRate);
Console.WriteLine("Fitness Calculator: {0}", fitness);
Console.WriteLine("Crossover Type: {0}", crossover);
Console.WriteLine();
//game.Start(new Replay("AAAA A00A AAAA AAAA AAA2 8000 AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA"));
var fitnessCalculator = new HighestScoreFitnessCalculator(game);
var ga = new GeneticAlgorithm(0.6, 0.25, 0.1, 0.8, chromosomeCount,
new RouletteWheelSelection(),
new OnePointCrossover(fitnessCalculator),
new SinglePointMutation(new ReplayCharacterSet(),fitnessCalculator),
fitnessCalculator);
Console.WriteLine(DateTime.Now);
Population population = new Population(GenerateRandomReplays(replayGenerator, chromosomeCount), fitnessCalculator);
int generation = 1;
while(true)
{
var bestOne = population.ElementAt(0);
Console.WriteLine("Generation {0}", generation++);
Console.WriteLine("Top fitness: {0}", bestOne.Fitness);
Console.WriteLine("Replay: {0}\n", bestOne.Replay);
population = ga.AdvancePopulation(population);
}
}
