/// <summary>
/// Evalutate the fitness of the population.
/// </summary>
/// <param name="ad_fitness"></param>
/// <param name="ad_bestfitness"></param>
/// <param name="ao_pop"></param>
/// <param name="ao_notablechroms"></param>
public static void EvaluateFitness(ref double ad_fitness, ref double ad_bestfitness, Chromosome <T>[] ao_pop, NotableChromosomes <T> ao_notablechroms = null)
{
ad_bestfitness = 0;
double ld_tfitness = 0;
Globals <T> .MODIFYBONUS = 0;
for (int i = 0; i < Globals <T> .POOLSIZE; i++)
{
ld_tfitness += ao_pop[i].fitness;
if (ao_pop[i].fitness < ad_bestfitness || ad_bestfitness == 0)
{
ad_bestfitness = ao_pop[i].fitness;
}
if (ao_notablechroms != null)
{
ao_notablechroms.UpdateFinalBest(ao_pop[i]);
}
}
ad_fitness = (ld_tfitness / ao_pop.Length);
GeneticAPI.Fitness.Elitism <T> .MarkElite(ao_pop);
}
/// <summary>
/// Initialize the population by calling the initialize chromosome factory until the population is full.
/// </summary>
/// <param name="ad_fitness"></param>
/// <param name="ad_bestfitness"></param>
/// <param name="ao_pop"></param>
/// <param name="ao_notablechroms"></param>
public static void Initialize(ref double ad_fitness, ref double ad_bestfitness, Chromosome <T>[] ao_pop, NotableChromosomes <T> ao_notablechroms = null)
{
for (int i = 0; i < Globals <T> .POOLSIZE; i++)
{
Chromosome <T> temp = InitialChromosomeFactory <T> .GenerateChromosome();
ao_pop[i] = temp;
if (ao_notablechroms != null)
{
ao_notablechroms.UpdateInitialBest(temp);
}
}
//Evaluate the fitness of the population.
EvaluateFitness(ref ad_fitness, ref ad_bestfitness, ao_pop);
}
/// <summary>
/// Starts the Genetic Algorithm.
/// </summary>
/// <param name="ao_data">List of IData to be converted into Genes and stored in Chromosome.</param>
/// <param name="ai_poolsize">Size of population.</param>
/// <param name="ai_generations">Number of generations to iterate algorithm.</param>
/// <param name="ad_modifyprob">Probability of modification/mutation.</param>
/// <param name="ad_recomprob">Probability of recombination/crossover.</param>
/// <param name="aen_selector">Selector Operator.</param>
/// <param name="aen_recomb">Recombination Operator.</param>
/// <param name="aen_random">Random basic vs adv</param>
/// <param name="ai_elites">Number of Elites.</param>
/// <param name="ai_ts_contestants">Number of Contestants for Tournament Selector.</param>
/// <param name="ab_adaptivemut">Adaptive Mutation Enabler.</param>
/// <param name="ab_rog">ROG Enabler.</param>
/// <param name="ab_lrog">LROG Enabler.</param>
public void Execute
(
List <T> ao_data,
int ai_poolsize,
int ai_generations,
double ad_modifyprob,
double ad_recomprob,
Selectors aen_selector,
Recombinators aen_recomb,
Randoms aen_random,
int ai_elites = 1,
int ai_ts_contestants = 2,
bool ab_adaptivemut = true,
bool ab_rog = false,
bool ab_lrog = true
)
{
//Initialize global variables.
Globals <T> .DATA = DataEncoder <T> .EncodeListFromData(ao_data);
Globals <T> .GENERATIONS = ai_generations;
Globals <T> .POOLSIZE = ai_poolsize;
Globals <T> .MODIFYPROB = ad_modifyprob;
Globals <T> .MODIFYBONUS = 0;
Globals <T> .RECOMPROB = ad_recomprob;
Globals <T> .ELITENUM = ai_elites;
Globals <T> .ADAPTMUT = ab_adaptivemut;
Globals <T> .ROG = ab_rog;
Globals <T> .LROG = ab_lrog;
if (aen_random == Randoms.Basic)
{
Globals <T> .RAND = new BasicRandom();
}
else
{
Globals <T> .RAND = new AdvRandom();
}
//Initialize local variables.
NotableChromosomes <T> lo_noteablechroms = new NotableChromosomes <T>();
Chromosome <T>[] lo_pop = new Chromosome <T> [Globals <T> .POOLSIZE];
double ld_fitness = 0;
double ld_popbestfitness = 0;
double ld_inifitness = 0;
int li_generation = 0;
//Start LROG thread if it's active.
//if (Globals<T>.LROG && io_thread == null)
//{
// Globals<T>.CPQ = new ConcurrentPriorityQueue<Chromosome<T>>(100);
// LessROG<T> lo_seeder = new LessROG<T>(ref Globals<T>.CPQ);
// io_thread = new Thread(new ThreadStart(lo_seeder.Run));
// io_thread.Start();
//}
if (Globals <T> .LROG && Globals <T> .SROGTHREAD == null)
{
Globals <T> .CPQ = new ConcurrentPriorityQueue <Chromosome <T> >(100);
LessROG <T> lo_seeder = new LessROG <T>(ref Globals <T> .CPQ);
Globals <T> .SROGTHREAD = new Thread(new ThreadStart(lo_seeder.Run));
Globals <T> .SROGTHREAD.Start();
}
//Initialize population.
ExecutionFunctions <T> .Initialize(ref ld_inifitness, ref ld_popbestfitness, lo_pop, lo_noteablechroms);
Chromosome <T>[] lo_newpop = new Chromosome <T> [Globals <T> .POOLSIZE];
//Iterate Genetic Algorithm.
while (ContinueGA(ref li_generation))
{
ExecutionFunctions <T> .EvaluateElite(lo_pop);
ExecutionFunctions <T> .Select(lo_pop, lo_newpop, aen_selector, ai_ts_contestants);
ExecutionFunctions <T> .Recombination(lo_newpop, aen_recomb);
ExecutionFunctions <T> .Modification(lo_newpop);
ExecutionFunctions <T> .EvaluateFitness(ref ld_fitness, ref ld_popbestfitness, lo_newpop, lo_noteablechroms);
lo_pop = lo_newpop;
//Send statistics to UI.
OnChanged(new APIEventArgs("", false, ld_fitness, ld_popbestfitness, lo_noteablechroms.GetFinalBest().fitness, lo_noteablechroms.GetFinalBest().ToString()));
}
//Send final statistics to UI.
OnChanged(new APIEventArgs("Initial avg fitness: ", false, ld_inifitness, ld_popbestfitness, lo_noteablechroms.GetFinalBest().fitness, lo_noteablechroms.GetFinalBest().ToString()));
OnChanged(new APIEventArgs("Final avg fitness: ", false, ld_fitness, ld_popbestfitness, lo_noteablechroms.GetFinalBest().fitness, lo_noteablechroms.GetFinalBest().ToString()));
OnChanged(new APIEventArgs("Initial best fitness: ", false, ld_fitness, ld_popbestfitness, lo_noteablechroms.GetFinalBest().fitness, lo_noteablechroms.GetFinalBest().ToString()));
OnChanged(new APIEventArgs("Overall best fitness: ", false, ld_fitness, ld_popbestfitness, lo_noteablechroms.GetFinalBest().fitness, lo_noteablechroms.GetFinalBest().ToString(), true));
}
