def test_gene_value_first_bits(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 1
end_pos = 3
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertEqual(3, individual.gene_value(start_pos, end_pos))
def test_gene_value_all(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 1
end_pos = 7
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertEqual(67, individual.gene_value(start_pos, end_pos))
def __create_next_generation__(self):
next_generation = []
while len(next_generation) != self.population_size:
# select parents
parents = self.select_parents_callback(self.current_individuals)
if len(parents) != 2:
raise Exception('Parent selection must return two individuals')
# create progeny
chromosomes = self.crossover_callback(*parents)
current_index = 0
while current_index < len(chromosomes) and len(
next_generation) != self.population_size:
# mutation (defaulted to flip mutation)
chromosome = self.mutate_callback(chromosomes[current_index])
if not 0 <= chromosome <= utils.max_binary_value(
self.chromosome_size):
raise Exception(
'Chromosomes generated must be between within max chromosome value'
)
# create individual
individual = Individual(self.chromosome_size, chromosome)
individual.set_parents(*parents)
next_generation.append(individual)
current_index += 1
return next_generation
def setUp(self) -> None:
self.target = Individual(5, [1, 1, 1, 1, 1])
self.individual1 = Individual(5, [0, 0, 0, 0, 0])
self.individual2 = Individual(5, [1, 0, 1, 0, 0])
self.individual3 = Individual(5, [0, 1, 0, 1, 1])
self.individual4 = Individual(5, [1, 1, 1, 1, 1])
self.population = Population(self.target, 4, [self.individual1, self.individual2, self.individual3,
self.individual4])
self.population.calc_fitness_scores()
self.ga_engine = GAEngine(5, 4)
def setUp(self) -> None:
self.target = Individual(5, [1, 1, 1, 1, 1])
self.individual1 = Individual(5, [0, 0, 0, 0, 0])
self.individual2 = Individual(5, [1, 0, 1, 0, 0])
self.individual3 = Individual(5, [0, 1, 0, 1, 1])
self.individual4 = Individual(5, [1, 1, 1, 1, 1])
self.population = Population(self.target, 4, [
self.individual1, self.individual2, self.individual3,
self.individual4
])
def test_center_cross_over_uneven(self):
size = 5
chromosome_1 = int('10001', 2)
chromosome_2 = int('11100', 2)
individual_1 = Individual(chromosome_size=size,
chromosome=chromosome_1)
individual_2 = Individual(chromosome_size=size,
chromosome=chromosome_2)
expected_1 = int('10100', 2)
expected_2 = int('11001', 2)
cross_1, cross_2 = CenterCross.crossover(individual_1, individual_2)
self.assertEqual(cross_2, expected_2)
self.assertEqual(cross_1, expected_1)
def test_gene_value_at_zero(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 0
end_pos = 4
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertRaises(ValueError, individual.gene_value, start_pos, end_pos)
def test_gene_value_past_limit(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 6
end_pos = 8
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertRaises(ValueError, individual.gene_value, start_pos, end_pos)
def __default_initialize__(algorithm):
individuals = [
Individual(algorithm.chromosome_size,
generate_random_chromosome(algorithm.chromosome_size))
for _ in range(algorithm.population_size)
]
return individuals
def random_population(self, city2travel, population_size):
"""Generate a random population"""
self.size = population_size
for i in range(self.size):
self.list.append(Individual(city_list=city2travel))
self.compute_stats()
self.compute_breed_probability()
def test_add_individual(self):
temp_individual = Individual(5, [1, 0, 0, 0, 0])
self.population.add_individual(temp_individual)
self.assertEqual(self.population.individuals, [
temp_individual, self.individual2, self.individual3,
self.individual4
], "post addition scores wrong")
self.assertEqual(self.population.fitness_scores, [1, 2, 3, 5],
"post addition scores wrong")
temp_individual = Individual(5, [1, 0, 1, 1, 1])
self.population.add_individual(temp_individual)
self.assertEqual(self.population.individuals, [
temp_individual, self.individual2, self.individual3,
self.individual4
], "post addition scores wrong")
self.assertEqual(self.population.fitness_scores, [4, 2, 3, 5],
"post addition scores wrong")
def test_select_parents_two_elements(self):
individual_1 = Individual(4, 1)
individual_1.fitness = 1
individual_2 = Individual(4, 1)
individual_2.fitness = 2
child_1, child_2 = RouletteWheel.select_parents(
[individual_1, individual_2])
self.assertFalse(child_1 == child_2)
def next_generation(self):
"""Generate the next generation based on the current population
The next generation's individuals are sons of the current population.
The function return a new instance of a Population class.
"""
next_gen = Population()
i = 0
while i < self.size:
father, mother = self.pick_parents()
next_gen = next_gen + Individual(father=father, mother=mother)
i += 1
next_gen.compute_stats()
next_gen.compute_breed_probability()
return next_gen
def test_add_individual(self):
print("scores_b4_adding", self.population.get_fitness_scores())
new_individual = Individual(choose_from_search_space(
self.search_space))
new_individual.set_fitness_score(0.01978)
worst_index = np.argmin(self.population.get_fitness_scores())
print(worst_index)
print("new_params", new_individual.get_nn_params())
print("worst_param",
self.population.individuals[worst_index].get_nn_params())
self.population.add_individual(new_individual,
new_individual.get_fitness_score())
print("scores_after_adding", self.population.get_fitness_scores())
self.assertEqual(self.population.individuals[worst_index],
new_individual)
def setUp(self):
self.individual = Individual(chromosome_size=10, chromosome=0)
def generate_individual(self, number: int) -> Individual:
return Individual(
uid=self.generate_uid(number),
yob=self.year,
iq=self.get_random_iq_from_current_distribution()
)
class TestIndividual(unittest.TestCase):
def setUp(self):
self.individual = Individual(chromosome_size=10, chromosome=0)
def test_add_metric_value(self):
key = 'metric'
val = 5
self.individual.add_metric(key, val)
self.assertEqual(1, len(self.individual.metrics))
self.assertEqual(5, self.individual.get_metric(key))
def test_add_metric_overrides(self):
key = 'metric'
val_1 = 5
val_2 = 3
self.individual.add_metric(key, val_1)
self.individual.add_metric(key, val_2)
self.assertEqual(1, len(self.individual.metrics))
self.assertEqual(3, self.individual.get_metric(key))
def test_add_metric_averaging(self):
key = 'metric'
val_1 = 5
val_2 = 3
self.individual.add_metric(key, val_1, average=True)
self.individual.add_metric(key, val_2, average=True)
self.assertEqual(1, len(self.individual.metrics))
self.assertEqual(4, self.individual.get_metric(key))
def test_gene_value_first_bits(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 1
end_pos = 3
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertEqual(3, individual.gene_value(start_pos, end_pos))
def test_gene_value_end_bits(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 6
end_pos = 7
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertEqual(2, individual.gene_value(start_pos, end_pos))
def test_gene_value_past_limit(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 6
end_pos = 8
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertRaises(ValueError, individual.gene_value, start_pos, end_pos)
def test_gene_value_at_zero(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 0
end_pos = 4
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertRaises(ValueError, individual.gene_value, start_pos, end_pos)
def test_gene_value_all(self):
chromosome = 67 # 1000011
chromosome_size = 7
start_pos = 1
end_pos = 7
individual = Individual(chromosome_size=chromosome_size, chromosome=chromosome)
self.assertEqual(67, individual.gene_value(start_pos, end_pos))
def tearDown(self):
self.individual.clear_metrics()