def test_recombine_two_parents(self):
ind1 = Individual(None, [0.11, 0.12, 0.13], [0.011, 0.012, 0.013])
ind2 = Individual(None, [0.21, 0.22, 0.23], [0.021, 0.022, 0.023])
newborn = StandardRecombination.recombine([ind1, ind2])
self.assertTrue(np.equal([0.016, 0.017, 0.018], newborn.sigma).all())
def test_recombine_threes_parents(self):
error = 1e-10
ind1 = Individual(None, [0.11, 0.12, 0.13, 0.14], [0.011, 0.012, 0.013, 0.014])
ind2 = Individual(None, [0.21, 0.22, 0.23, 0.24], [0.021, 0.022, 0.023, 0.024])
ind3 = Individual(None, [0.31, 0.32, 0.33, 0.3], [0.031, 0.032, 0.033, 0.034])
newborn = StandardRecombination.recombine([ind1, ind2, ind3])
self.assertTrue(np.abs(np.linalg.norm(newborn.sigma - [0.021, 0.022, 0.023, 0.024])) < error)
def test_remove_last_decrease_table_size_by_one(self):
# insert two individuals
self.table.add_individual(Individual(15.0, [2.0], [0.2]))
self.table.add_individual(Individual(-15.0, [3.0], [0.3]))
self.assertEquals(2, len(self.table))
self.table.remove_last()
self.assertEquals(1, len(self.table))
def test_clear_sets_tablesize_to_zero(self):
# insert two individuals
self.table.add_individual(Individual(15.0, [2.0], [0.2]))
self.table.add_individual(Individual(-15.0, [3.0], [0.3]))
self.assertEquals(2, len(self.table))
self.table.clear()
self.assertEquals(0, len(self.table))
def test_inserted_elements_are_sorted(self):
# insert four individuals
self.table.add_individual(Individual(15.0, [2.0], [0.2]))
self.table.add_individual(Individual(-15.0, [3.0], [0.3]))
self.table.add_individual(Individual(-30.0, [4.0], [0.4]))
self.table.add_individual(Individual(15.0, [5.0], [0.5]))
self.assertEqual(4, len(self.table))
for i in range(1, len(self.table)):
self.assertTrue(self.table[i].fitness <= self.table[i-1].fitness)
def table_random_initialization(self, fitness_object):
n_params = self.params_ngb.num_params
initial_sigma = self.params_ngb.initial_sigma
sigma_dim = n_params if self.params_ngb.is_sigma_array else 1
for i in range(0, self.params_ngb.initial_table_size):
rnd_individ = Individual(None, np.random.uniform(0.0, 1.0, n_params),
initial_sigma * np.ones(sigma_dim))
mapped_params = self.map_parameters(rnd_individ.params) # map params from (0.0, 1.0) -> search_space
rnd_individ.fitness = fitness_object.compute(mapped_params)
self.fitness_table.add_individual(rnd_individ)
def test_add_remove_individual_increases_size(self):
size_initial = len(self.table)
self.table.add_individual(Individual(5.0, [1.0], [0.1]))
assert size_initial == 0
assert len(self.table) == 1
def test_after_mutation_parameters_are_within_zero_to_one_range(self):
individual = Individual(None, 0.5 * np.ones(5), 0.3 * np.ones(5))
sigma_min = 0.01
sigma_max = 0.35
m = StandardMutation(len(individual.params), sigma_min, sigma_max)
mutated_individual = m.mutate(individual)
self.assertTrue((mutated_individual.params <= 1.0).all())
self.assertTrue((mutated_individual.params >= 0.0).all())
def mutate(self, individual):
# mutate sigma array
cmf = np.exp(
self.tau0 *
self.rnd_std_norm(1)) # common mutation factor for sigma vector
sigma = cmf * individual.sigma * np.exp(
self.tau1 * self.rnd_std_norm(len(individual.sigma)))
# mutate parameters, use newly mutated sigma (important)
params = individual.params + sigma * self.rnd_std_norm(
len(individual.params))
# in case of overshooting: return parameters inside the range [0.0, 1.0]
params[params < 0.0] = 0.0
params[params > 1.0] = 1.0
# in case of overshooting sigmas: return parameters inside the range [sigma_min, sigma_max]
sigma[sigma < self.sigma_min] = self.sigma_min
sigma[sigma > self.sigma_max] = self.sigma_max
return Individual(individual.fitness, params, sigma)
def recombine(individuals):
"""As a result of the parents recombination a new child is produced.
Each parameter of the newborn(recombined) individual represents a copy
of one of the parent's. Parent for each parameter is chosen randomly.
Sigma of the recombined individual is a simple average of all parents"""
parents = len(individuals)
params = np.zeros(len(individuals[0].params))
sigma = np.zeros(len(individuals[0].sigma))
# recombine parameters array
for i in range(0, len(params)):
idx = random.randint(0, parents - 1)
params[i] = individuals[idx].params[i]
# recombine sigma array
for i in range(0, len(sigma)):
for j in range(0, parents):
sigma[i] += individuals[j].sigma[i]
sigma /= parents
return Individual(None, params, sigma)
def test_creation_of_individuals(self):
individual = Individual(0.5, [-10.0, 20.0, 1000.0], [1.0, 0.5, 0.4])
self.assertEqual(0.5, individual.fitness)
self.assertTrue(np.equal(np.array([-10.0, 20.0, 1000.0]), individual.params).all())
self.assertFalse(np.equal(np.array([-10.0, 20.1, 1000.0]), individual.params).all())
self.assertTrue(np.equal(np.array([1.0, 0.5, 0.4]),individual.sigma).all())
def test_create_individual_with_empty_sigma_throws(self):
with self.assertRaises(ValueError):
Individual(-1.0, [1.0, 2.0], [0.1, 0.2, 0.3])
def test_create_individual_with_empty_parameters_throws(self):
with self.assertRaises(ValueError):
Individual(-1.0, [], [1.0])
def test_create_individual_with_complex_fitness_throws(self):
with self.assertRaises(ValueError):
Individual(complex(1.0, 1.0), [1.0], [1.0])
def test_create_individual_with_nan_fitness_throws(self):
with self.assertRaises(ValueError):
Individual(math.inf, [1.0], [1.0])
