def build_and_run(early_conv_avoid, mutation_rate, crossover_rate,
population_size, elitism_count, ga_type, chromo_type):
ga = ga_type(elitism_count, mutation_rate, crossover_rate, population_size)
population = Population()
population.init_population(population_size, chromo_type)
print(ga)
print(early_conv_avoid)
return run(ga, population, early_conv_avoid)
def crossover(self, parents: List[Tuple[Chromosome]], pop_size: int) -> Population:
"""
get list of parents to mate, return the new population
NOTE: when population (size - elitism amount) doesn't divide in (offspring amount) then the remainder
offsprings won't enter the new population.
"""
new_population = Population()
for chromosomes in parents:
# do crossover with probability self.crossover_rate
if random.random() < self.crossover_rate:
offsprings = self.pair_chromosomes(chromosomes)
else:
offsprings = chromosomes # [ch.__copy__() for ch in chromosomes] # todo: maybe redundant
for ch in offsprings:
new_population.add_chromosome(ch)
# if there are too many offsprings then stop adding them to population
if new_population.get_size() == pop_size:
return new_population
def selection(self, population: Population) -> List[Tuple[Chromosome]]:
"""
select parents for crossover
implements basic selection method: 1) create a pool of chromosomes, 2) select the amount needed for mating.
"""
parents = []
pairings = math.ceil(population.get_size() / self.get_offsprings_amount())
self.set_selection_pool([ch for ch in population])
for _ in range(pairings):
parents.append(tuple([self.select() for _ in range(self.get_parents_amount())]))
return parents
def mutation(self, population: Population) -> Population:
""" """
new_pop = Population()
for ch in population:
new_pop.add_chromosome(self.mutate(ch))
return new_pop
def build_and_run(mutation_rate, crossover_rate, population_size,
elitism_count, ga_type, chromo_type):
ga = ga_type(elitism_count, mutation_rate, crossover_rate, population_size)
population = Population()
population.init_population(population_size, chromo_type)
return run(ga, population)
def check_for_ec(self, gen: int, ga: GeneticAlgorithm,
population: Population):
w = population.get_least_fit().get_fitness()
b = population.get_fittest().get_fitness()
return b - w < self.dist
def check_for_ec(self, gen: int, ga: GeneticAlgorithm,
population: Population):
f = population.get_fittest().get_fitness()
m = mean(ch.get_fitness() for ch in population)
return f - m < self.dist_from_avg