def create_genotype(parents: List[ByteGenotype]) -> ByteGenotype:
"""
Randomly select bits of the parent Genotypes to create a new child Genotype
:param parents: A list of all the possible parents for the child Genotype
"""
child_byte_array = bytearray()
for i in range(len(parents[0])):
parent = choice(parents)
child_byte_array.append(parent[i])
child = ByteGenotype(child_byte_array)
child.mutate(MUTATION_RATE)
return child
def crossover_and_mutate(parent1: ByteGenotype,
parent2: ByteGenotype) -> ByteGenotype:
child_array = []
for i in range(len(parent1)):
parent_choice = randint(0, 1)
parent: ByteGenotype
if parent_choice == 0:
parent = parent1
else:
parent = parent2
child_array.append(parent[i])
child = ByteGenotype(bytearray(child_array))
child.mutate(MUTATION_RATE)
return child
def _choose_genotypes_from_sorted_unique_list(
self, genotypes: List[ByteGenotype]) -> List[ByteGenotype]:
"""
Takes in all the unique Genotypes from the population and selects some to keep
:param genotypes: A sorted list of unique Genotypes
"""
splitting_point = math.floor(self._M / 2)
good = [genotypes[i] for i in range(self._M - splitting_point)]
# bad = [genotypes[i] for i in range(-1, -splitting_point-1, -1)]
bad = ByteGenotype.generate_random_population(
splitting_point, len(genotypes[0])) #insert garbage Genotypes
return good + bad
def main():
# string_to_find = "hello world"
# number_of_primes = 50
# size_of_primes_genotype = 16
black_jack_genotype_length = PlayerParameters.number_of_parameters(
) * 4 # 4 bytes per float
fitness_calculator = make_FitnessCalculator(Application.BLACK_JACK,
[1000, 20])
population = ByteGenotype.generate_random_population(
20, black_jack_genotype_length)
simple_manager = mitosis.SimpleManager(population, fitness_calculator,
"simple_mitosis_manager")
truncation_manager = mitosis.TruncationManager(population,
fitness_calculator,
"truncation_manager")
brute_force_manager = mitosis.BruteForce(population, fitness_calculator,
"brute_force_manager")
simple_meiosis_manager = meiosis.SimpleManager(population,
fitness_calculator,
"simple_meiosis_manager")
multiple_children_manager = meiosis.MultipleChildrenManager(
population, fitness_calculator, "multiple_children_manager")
diversity_manager = meiosis.DiversityManager(population,
fitness_calculator,
"diversity_manager")
tournament_manager = meiosis.TournamentManager(population,
fitness_calculator,
"tournament_manager")
managers = [
simple_meiosis_manager, multiple_children_manager, diversity_manager,
brute_force_manager
]
optimizer = EvolutionaryOptimizer(managers, True)
name_to_reports = optimizer.run_many_lifecycles(5000)
rep.generate_many_reports(LifecycleReport.header(), name_to_reports, {}, 1)
solutions: List[Tuple[str, any]] =\
[(elem[0], fitness_calculator.converter.convert(elem[1][-1].solution)) for elem in name_to_reports.items()]
print("Found solutions:")
for elem in solutions:
print("From", elem[0])
print("Generated word:", elem[1])
def produce_offspring(self, population: List[ByteGenotype]) -> Tuple[List[ByteGenotype], List[ByteGenotype]]:
children = ByteGenotype.generate_random_population(20, len(self._best_genotype))
return [self._best_genotype], children