void SteadyStateGA::nextGeneration()
{
double cross_random;
double mutate_random;
// sort the population bases on the parameters
pop->sortPopulation();
vector<Genome *> *new_genomes = new vector<Genome *>;
// we only replace replace_percentage of the population.
// decide which percentage to replace...the lowest scores... (-1) fitnesses
// reverse interate through and delete the genomes.
// copy the top (pop_size * replacement_percentage) genomes into the new genome vector
for(int i = 0; i < (int)(pop->getPopSize() * replace_percentage); ++i)
{
new_genomes->push_back(pop->getGenome(i)->clone());
}
// now i need to add 1 - pop->getPopSize() * replace_percentage more genomes
for(int i = 0; i < (int)(pop->getPopSize() - (pop->getPopSize() * replace_percentage)); ++i)
{
// need to select a dad and mom .. select...
Genome *dad = pop->select();
Genome *mom = pop->select();
// then create child for crossover.
Genome *child;
cross_random = objRand->randomPercentage();
// if crossover does not happen make the children = to the parents
if(cross_random <= crossover_percentage)
{
// perform crossover
child = dad->crossover(*mom);
}
else
{
// make child and dad equal to each other
child = dad->clone();
}
mutate_random = objRand->randomPercentage();
// chance of mutation to each of the children
if(mutate_random <= mutation_percentage)
{
// perform crossover
child->mutate();
}
if((int)new_genomes->size() < pop->getPopSize())
new_genomes->push_back(child);
}
++current_generation;
pop->setPopGenomes(new_genomes);
init(); // init the new population
}
