def test_partially_match_crossover():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([4, 3, 5, 7, 8, 1, 2, 0, 9, 6]))
i2 = pg.Individual(genotype=np.array([8, 9, 1, 3, 6, 5, 7, 4, 2, 0]))
o1, o2 = pg.partially_match_crossover(i1, i2)
assert np.array_equal(o1.genotype, np.array([4, 7, 8, 3, 6, 5, 1, 2, 0,
9]))
assert np.array_equal(o2.genotype, np.array([9, 3, 6, 7, 8, 1, 5, 4, 2,
0]))
def test_equal_individual():
ind1 = pg.Individual(fitness=pg.Fitness(value=3), genotype=[0, 1, 0, 1, 1])
ind2 = ind1.clone()
assert ind1.equal(ind2) is True
ind3 = pg.Individual(fitness=pg.Fitness(value=3), genotype=[0, 1, 1, 1, 1])
assert ind1.equal(ind3) is False
ind4 = pg.Individual(fitness=pg.Fitness(value=2), genotype=[0, 1, 0, 1, 1])
assert ind1.equal(ind4) is False
ind5 = pg.Individual(fitness=pg.Fitness(value=3), genotype=[0, 1, 0, 1, 1])
ind5.run_eval = False
assert ind1.equal(ind5) is False
def test_uniform_crossover():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array(
[0, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1]))
i2 = pg.Individual(genotype=np.array(
[1, 1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 1]))
o1, o2 = pg.uniform_crossover(i1, i2)
assert np.array_equal(
o1.genotype,
np.array([0, 1, 0, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1]))
assert np.array_equal(
o2.genotype,
np.array([1, 1, 1, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1]))
def test_flip_mutation3():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([4, 3, 5, 7, 8, 1, 2, 0, 9, 6]))
m1 = pg.flip_mutation(i1, gene_rate=None, low=0, high=9)
assert m1.run_eval is True
assert np.array_equal(m1.genotype, np.array([4, 3, 5, 7, 8, 1, 7, 0, 9,
6]))
def discrete_crossover_adaptive(i1, i2):
'''
Discrete Crossover
Args:
i1 (Individual): indivdiual 1 fixed genome
i2 (Individual): indivdiual 2 fixed genome
Returns:
single offspring from 2 parents
'''
o1 = pg.Individual(
genotype=np.empty(i1.genotype.size, dtype=i1.genotype.dtype))
o1.parameters = np.empty(i1.parameters.size, dtype=i1.parameters.dtype)
for i in range(o1.genotype.size):
if np.random.uniform() < 0.5:
o1.genotype[i] = i1.genotype[i]
else:
o1.genotype[i] = i2.genotype[i]
for i in range(o1.parameters.size):
if np.random.uniform() < 0.5:
o1.parameters[i] = i1.parameters[i]
else:
o1.parameters[i] = i2.parameters[i]
return o1
def test_swap_mutation2():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([4, 3, 5, 7, 8, 1, 2, 0, 9, 6]))
m1 = pg.swap_mutation(i1, gene_rate=1.0)
assert m1.run_eval is True
assert np.array_equal(m1.genotype, np.array([0, 2, 3, 4, 9, 5, 8, 1, 7,
6]))
def test_disrete_crossover():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
i2 = pg.Individual(genotype=np.array([
0.60754485, 0.17052412, 0.06505159, 0.94888554, 0.96563203, 0.80839735,
0.30461377, 0.09767211, 0.68423303, 0.44015249
]))
o1 = pg.discrete_crossover(i1, i2)
assert np.isclose(
o1.genotype.all(),
np.array([
0.61185289, 0.17052412, 0.06505159, 0.94888554, 0.45606998,
0.78517596, 0.19967378, 0.09767211, 0.68423303, 0.44015249
]).all())
def test_intermediary_crossover():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
i2 = pg.Individual(genotype=np.array([
0.60754485, 0.17052412, 0.06505159, 0.94888554, 0.96563203, 0.80839735,
0.30461377, 0.09767211, 0.68423303, 0.44015249
]))
o1 = pg.intermediary_crossover(i1, i2)
assert np.isclose(
o1.genotype.all(),
np.array([
0.60969887, 0.15500899, 0.17859812, 0.65762369, 0.710851,
0.79678666, 0.25214378, 0.30595327, 0.6383238, 0.24330145
]).all())
def test_binary_flip_mutation1():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array(
[0, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1]))
m1 = pg.binary_flip_mutation(i1, gene_rate=0.0)
assert m1.run_eval is True
assert np.array_equal(
m1.genotype,
np.array([0, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1]))
def test_uncorrelated_one_step_mutation():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
m1 = pg.uncorrelated_one_step_mutation(i1)
assert np.isclose(
m1.genotype.all(),
np.array([
0.60427399, 0.17499665, 0.37562893, 0.35352682, 0.44323586,
0.87173989, 0.24174041, 0.48850039, 0.62215478, 0.05481461
]).all())
def test_clone_individual():
ind1 = pg.Individual(fitness=pg.Fitness(value=3), genotype=[0, 1, 0, 1, 1])
ind2 = ind1.clone()
assert ind1.fitness.equal(ind2.fitness) is True
assert ind1.genotype == ind2.genotype
assert ind1.run_eval == ind2.run_eval
ind1.fitness = 2
ind1.genotype = [1, 0, 1, 0, 0]
ind1.run_eval = False
assert ind2.fitness.value == 3
assert ind2.genotype == [0, 1, 0, 1, 1]
assert ind2.run_eval is True
def test_uncorrelated_n_steps_mutation():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
m1 = pg.uncorrelated_n_steps_mutation(i1)
assert np.isclose(
m1.genotype.all(),
np.array([
1.97088495, 0.38304596, -0.28830838, 0.70303098, 0.43352264,
0.86057563, 0.31735872, 1.23638915, 0.6205196, 0.04865446
]).all())
def test_uniform_mutation3():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
m1 = pg.uniform_mutation(i1, gene_rate=None, low=0.0, high=1.0)
assert m1.run_eval is True
assert np.isclose(
m1.genotype.all(),
np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998,
0.78517596, 0.86617615, 0.51423444, 0.59241457, 0.96990985
]).all())
def test_uniform_mutation2():
np.random.seed(42)
i1 = pg.Individual(genotype=np.array([
0.61185289, 0.13949386, 0.29214465, 0.36636184, 0.45606998, 0.78517596,
0.19967378, 0.51423444, 0.59241457, 0.04645041
]))
m1 = pg.uniform_mutation(i1, gene_rate=1.0, low=0.0, high=1.0)
assert m1.run_eval is True
assert np.isclose(
m1.genotype.all(),
np.array([
0.95071431, 0.59865848, 0.15599452, 0.86617615, 0.70807258,
0.96990985, 0.21233911, 0.18340451, 0.52475643, 0.29122914
]).all())
def test_negative_tournament_selection():
np.random.seed(42)
individuals = [
pg.Individual(fitness=pg.Fitness(value=np.random.randint(10)))
for _ in range(10)
]
pop = pg.Population(individuals=np.array(individuals))
# [6, 3, 7, 4, 6, 9, 2, 6, 7, 4]
np.random.seed(42)
inds1 = [pg.negative_tournament_selection(pop, size=5) for _ in range(10)]
assert [inds1[i].fitness.value
for i in range(10)] == [6, 7, 6, 9, 9, 9, 7, 7, 7, 7]
np.random.seed(24)
inds2 = [pg.negative_tournament_selection(pop, size=5) for _ in range(10)]
assert [inds2[i].fitness.value
for i in range(10)] == [7, 6, 7, 7, 6, 6, 3, 4, 6, 7]
def test_random_selection():
np.random.seed(42)
individuals = [
pg.Individual(fitness=pg.Fitness(value=np.random.randint(10)))
for _ in range(10)
]
pop = pg.Population(individuals=np.array(individuals))
# [6, 3, 7, 4, 6, 9, 2, 6, 7, 4]
np.random.seed(42)
inds1 = [pg.random_selection(pop) for _ in range(10)]
assert [inds1[i].fitness.value
for i in range(10)] == [2, 4, 6, 6, 2, 4, 7, 2, 6, 6]
np.random.seed(24)
inds2 = [pg.random_selection(pop) for _ in range(10)]
assert [inds2[i].fitness.value
for i in range(10)] == [7, 4, 6, 6, 3, 3, 3, 6, 6, 4]
def intermediary_crossover(i1, i2):
'''
Intermediary Crossover
Args:
i1 (Individual): indivdiual 1 fixed genome
i2 (Individual): indivdiual 2 fixed genome
Returns:
single offspring from 2 parents
'''
o1 = pg.Individual(
genotype=np.empty(i1.genotype.size, dtype=i1.genotype.dtype))
for i in range(o1.genotype.size):
o1.genotype[i] = (i1.genotype[i] + i2.genotype[i]) / 2.0
return o1
def make_generic_population(size, make_individual_fn, *args, **kargs):
'''
Make Generic Population
Args:
size (int): number of individuals in the population
make_individual_fn (function): function that returns a random individual
Return:
random population of fixed size without fitness evaluation
'''
pop = make_empty_population(size)
for i in range(size):
pop.individuals[i] = pg.Individual()
pop.individuals[i].genotype = make_individual_fn(*args, **kargs)
return pop
def apply_global_mutation(pop, pool_size, mt_op, **kargs):
'''
Apply Global Mutation
Args:
pop (Population): individuals to apply mutation, randomly picked until pool_size
pool_size (int): number of individuals that will be applied mutation
mt_op (function): mutation operator
**kargs: keyword arguments for the mutation oeprator
Return:
population of mutated individuals
'''
new_pop = pg.make_empty_population(pool_size)
for i in range(0, pool_size):
index = np.random.randint(pop.size)
new_pop.individuals[i] = pg.Individual()
new_pop.individuals[i] = mt_op(pop.individuals[index], **kargs)
new_pop.individuals[i].run_eval = True
return new_pop
def test_make_individual1():
ind = pg.Individual()
assert type(ind) is pg.Individual
assert ind.fitness is None
assert ind.genotype is None
assert ind.run_eval is True
def test_make_individual2():
ind = pg.Individual(fitness=pg.Fitness(value=3), genotype=[0, 1, 0, 1, 1])
assert type(ind) is pg.Individual
assert ind.fitness.value == 3
assert ind.genotype == [0, 1, 0, 1, 1]
assert ind.run_eval is True
