def setUp(self):
"""Test Problem: Assume you can only carry 3 items to sell in your bag.
Pick a combination of items that maximises your profit. The profit for
each item is listed.
"""
self.seed_data = [('apple', 15), ('banana', 10), ('carrot', 12),
('pear', 5), ('mango', 8)]
member_1 = pyeasyga.Chromosome([0, 1, 0, 1, 1])
member_2 = pyeasyga.Chromosome([1, 1, 0, 0, 1])
member_3 = pyeasyga.Chromosome([1, 1, 1, 1, 0])
member_4 = pyeasyga.Chromosome([0, 1, 1, 1, 1])
self.population = []
self.population.append(member_1)
self.population.append(member_2)
self.population.append(member_3)
self.population.append(member_4)
self.ga = pyeasyga.GeneticAlgorithm(self.seed_data)
self.ga.population_size = 10
self.ga.generations = 10
self.ga.fitness_function = lambda member, data: sum([
profit for (selected, (fruit, profit)) in zip(member, data)
if selected and member.count(1) == 3
])
self.ga.selection_function = self.ga.tournament_selection
def create_initial_population(self):
initial_population = []
if len(self.seed_data) > 2:
initial_population.append(pyeasyga.Chromosome(self.seed_data[2]))
while len(initial_population) < self.population_size:
genes = self.create_individual(self.seed_data)
individual = pyeasyga.Chromosome(genes)
initial_population.append(individual)
self.current_generation = initial_population
def test_chromosome_initialisation_2(self):
chromosome = pyeasyga.Chromosome(['d', 'e', 'f'])
chromosome.fitness = 20
assert chromosome.genes == ['d', 'e', 'f']
assert chromosome.fitness == 20
assert str(chromosome) == "(20, ['d', 'e', 'f'])"
def create_individual(prob):
individual = []
for p in prob:
if random() < p:
individual.append(1)
else:
individual.append(0)
return pyeasyga.Chromosome(individual)
def rn_create_individual(prob):
individual = []
for (mean, stdev) in prob:
# value = random.uniform(mean - stdev, mean + stdev)
# individual.append(value)
individual.append(np.random.normal(mean, stdev))
# Return a new individual
return pyeasyga.Chromosome(individual)
def test_create_next_generation(self):
pop_size = self.ga.population_size
self.ga.create_first_generation()
self.ga.create_next_generation()
current_gen = self.ga.current_generation
assert len(current_gen) == pop_size
assert isinstance(current_gen[0], type(pyeasyga.Chromosome([1])))
def test_create_initial_population(self):
pop_size = self.ga.population_size
self.ga.create_initial_population()
initial_pop = self.ga.current_generation
assert len(initial_pop) == pop_size
assert isinstance(initial_pop[0], type(pyeasyga.Chromosome([1])))
assert sum([member.fitness for member in initial_pop]) == 0
def test_run(self):
self.ga.run()
current_gen = self.ga.current_generation
last_generation = self.ga.last_generation()
assert len(current_gen) == self.ga.population_size
assert isinstance(current_gen[0], type(pyeasyga.Chromosome([1])))
assert isinstance(next(last_generation), type((1, [1, 1, 1, 1, 1])))
assert len(next(last_generation)) == 2
assert len(next(last_generation)[1]) == 5
def create_individual(prob):
individual = []
for p in prob:
mean, stdev = p
value = random.uniform(mean - stdev, mean + stdev)
individual.append(value)
mean = numpy.mean(individual)
stdev = numpy.std(individual, ddof=1)
individual = numpy.random.normal(mean, stdev, len(individual))
# Return a new individual
return pyeasyga.Chromosome(individual)
def test_create_new_population(self):
""" Write more functional test """
pop_size = self.ga.population_size
self.ga.create_initial_population()
self.ga.calculate_population_fitness()
self.ga.rank_population()
self.ga.create_new_population()
current_gen = self.ga.current_generation
assert len(current_gen) == pop_size
assert isinstance(current_gen[0], type(pyeasyga.Chromosome([1])))
def create_new_population(self):
new_population = []
for i in range(
int(CONSTANT.POPULATION_BEST_PORTION * self.population_size)):
elite = copy.deepcopy(self.current_generation[i])
new_population.append(elite)
selection = self.selection_function
while len(new_population) < int(self.population_size *
(CONSTANT.POPULATION_BEST_PORTION +
CONSTANT.POPULATION_REMAIN_PORTION)):
parent_1 = copy.deepcopy(selection(self.current_generation))
parent_2 = copy.deepcopy(selection(self.current_generation))
child_1, child_2 = parent_1, parent_2
child_1.fitness, child_2.fitness = 0, 0
can_crossover = random.random() < self.crossover_probability
can_mutate = random.random() < self.mutation_probability
if can_crossover:
child_1.genes, child_2.genes = self.crossover_function(
parent_1.genes, parent_2.genes)
if can_mutate:
self.mutate_function(child_1.genes)
self.mutate_function(child_2.genes)
new_population.append(child_1)
if len(new_population) < self.population_size:
new_population.append(child_2)
remain_num = self.population_size - len(new_population)
for i in range(remain_num):
new_population.append(
pyeasyga.Chromosome(self.create_individual(0)))
self.current_generation = new_population
def test_chromosome_initialisation_1(self):
chromosome = pyeasyga.Chromosome(['a', 'b', 'c'])
assert chromosome.genes == ['a', 'b', 'c']
assert chromosome.fitness == 0
assert str(chromosome) == "(0, ['a', 'b', 'c'])"
