def init_island():
np.random.seed(10)
x = init_x_vals(START, STOP, NUM_POINTS)
y = equation_eval(x)
training_data = ExplicitTrainingData(x, y)
component_generator = ComponentGenerator(x.shape[1])
component_generator.add_operator(2)
component_generator.add_operator(3)
component_generator.add_operator(4)
crossover = AGraphCrossover()
mutation = AGraphMutation(component_generator)
agraph_generator = AGraphGenerator(STACK_SIZE, component_generator)
fitness = ExplicitRegression(training_data=training_data)
local_opt_fitness = ContinuousLocalOptimization(fitness, algorithm='lm')
evaluator = Evaluation(local_opt_fitness)
ea = AgeFitnessEA(evaluator, agraph_generator, crossover, mutation,
MUTATION_PROBABILITY, CROSSOVER_PROBABILITY, POP_SIZE)
island = Island(ea, agraph_generator, POP_SIZE)
return island
def test_evaluation_evaluates_all_list_values_per_individual(
population, fitness_function):
evaluation = Evaluation(fitness_function)
evaluation(population)
assert evaluation.eval_count == 25
for indv in population:
assert indv.fit_set
assert indv.fitness is not None
def evol_alg():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(mutation_function)
selection = Tournament(SELECTION_SIZE)
fitness = MultipleValueFitnessFunction()
evaluator = Evaluation(fitness)
return MuPlusLambda(evaluator, selection, crossover, mutation, 0.2, 0.4,
OFFSPRING_SIZE)
def test_evaluation_evaluates_all_individuals(single_value_population_of_4,
fitness_function):
evaluation = Evaluation(fitness_function)
evaluation(single_value_population_of_4)
assert evaluation.eval_count == 4
for indv in single_value_population_of_4:
assert indv.fit_set
assert indv.fitness is not None
def test_evaluation_skips_already_calculated_fitnesses(population,
fitness_function):
evaluation = Evaluation(fitness_function)
population[0].fitness = 1.0
evaluation(population)
assert evaluation.eval_count == 24
for indv in population:
assert indv.fit_set
assert indv.fitness is not None
def ev_alg():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(np.random.random)
selection = Tournament(2)
training_data = np.linspace(0.1, 1, FULL_TRAINING_DATA_SIZE)
fitness = DistanceToAverage(training_data)
evaluator = Evaluation(fitness)
return MuPlusLambda(evaluator, selection, crossover, mutation,
0., 1.0, MAIN_POPULATION_SIZE)
def test_evaluation_skips_already_calculated_fitnesses(
single_value_population_of_4, fitness_function):
evaluation = Evaluation(fitness_function)
single_value_population_of_4[0].fitness = 1.0
evaluation(single_value_population_of_4)
assert evaluation.eval_count == 3
for indv in single_value_population_of_4:
assert indv.fit_set
assert indv.fitness is not None
def island():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(mutation_function)
selection = Tournament(10)
fitness = MultipleValueFitnessFunction()
evaluator = Evaluation(fitness)
ev_alg = MuPlusLambda(evaluator, selection, crossover, mutation, 0.2, 0.4,
20)
generator = MultipleValueChromosomeGenerator(mutation_function, 10)
return Island(ev_alg, generator, 25)
def test_fitness_equals_true_value_count(fitness_function, population):
evaluation = Evaluation(fitness_function)
evaluation(population)
for indv in population:
fitness = 0
for val in indv.values:
if val == False:
fitness += 1
assert indv.fitness == fitness_function(indv)
assert fitness == indv.fitness
def create_evolutionary_algorithm():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(generate_0_or_1)
variation_phase = VarOr(crossover, mutation, crossover_probability=0.4,
mutation_probability=0.4)
fitness = OneMaxFitnessFunction()
evaluation_phase = Evaluation(fitness)
selection_phase = Tournament(tournament_size=2)
return EvolutionaryAlgorithm(variation_phase, evaluation_phase,
selection_phase)
def main():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(get_random_float)
selection = Tournament(10)
fitness = ZeroMinFitnessFunction()
local_opt_fitness = ContinuousLocalOptimization(fitness)
evaluator = Evaluation(local_opt_fitness)
ea = MuPlusLambda(evaluator, selection, crossover, mutation, 0.4, 0.4, 20)
generator = MultipleFloatChromosomeGenerator(get_random_float, 8)
island = Island(ea, generator, 25)
best_indv_values = []
best_indv_values.append(island.best_individual().values)
for i in range(500):
island.execute_generational_step()
best_indv_values.append(island.best_individual().values)
bingo.animation.animate_data(best_indv_values)
def main():
crossover = SinglePointCrossover()
mutation = SinglePointMutation(get_random_float)
selection = Tournament(10)
fitness = ZeroMinFitnessFunction()
local_opt_fitness = ContinuousLocalOptimization(fitness)
evaluator = Evaluation(local_opt_fitness)
ea = MuPlusLambda(evaluator, selection, crossover, mutation, 0.4, 0.4, 20)
generator = MultipleFloatChromosomeGenerator(get_random_float, 8)
island = Island(ea, generator, 25)
for i in range(25):
island.execute_generational_step()
print("\nGeneration #", i)
print("-" * 80, "\n")
report_max_min_mean_fitness(island.population)
print("\npopulation: \n")
for indv in island.population:
print(["{0:.2f}".format(val) for val in indv.values])
def onemax_evaluator(onemax_fitness):
return Evaluation(onemax_fitness)
def test_setting_eval_count(single_value_population_of_4, fitness_function):
evaluation = Evaluation(fitness_function)
evaluation.eval_count = -4
assert evaluation.eval_count == -4
evaluation(single_value_population_of_4)
assert evaluation.eval_count == 0