def test_eval_node_probs():
"""Test eval_node_probs method"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
pop = np.array([[0, 0, 0, 0, 1],
[1, 0, 1, 0, 1],
[1, 1, 1, 1, 0],
[1, 0, 0, 0, 1],
[0, 0, 0, 0, 0],
[1, 1, 1, 1, 1]])
problem.keep_sample = pop
problem.eval_node_probs()
parent = np.array([2, 0, 1, 0])
probs = np.array([[[0.33333, 0.66667],
[0.33333, 0.66667]],
[[1.0, 0.0],
[0.33333, 0.66667]],
[[1.0, 0.0],
[0.25, 0.75]],
[[1.0, 0.0],
[0.0, 1.0]],
[[0.5, 0.5],
[0.25, 0.75]]])
assert (np.allclose(problem.node_probs, probs, atol=0.00001)
and np.array_equal(problem.parent_nodes, parent))
def test_random():
"""Test random method"""
problem = DiscreteOpt(5, OneMax(), maximize=True, max_val=5)
rand = problem.random()
assert (len(rand) == 5 and max(rand) >= 0 and min(rand) <= 4)
def test_reproduce_mut1_max2():
"""Test reproduce method when mutation_prob is 1 and max_val is 2"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
father = np.array([0, 0, 0, 0, 0])
mother = np.array([1, 1, 1, 1, 1])
child = problem.reproduce(father, mother, mutation_prob=1)
assert (len(child) == 5 and sum(child) > 0 and sum(child) <= 5)
def test_find_sample_order():
"""Test find_sample_order method"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
problem.parent_nodes = np.array([2, 0, 1, 0])
order = np.array([0, 2, 4, 1, 3])
problem.find_sample_order()
assert np.array_equal(np.array(problem.sample_order), order)
def generate(seed, size=60, t_pct=0.1):
np.random.seed(seed)
fitness = ContinuousPeaks(t_pct=t_pct)
problem = DiscreteOpt(length=size, fitness_fn=fitness)
crossover = one_point_crossover.OnePointCrossOver(problem)
problem = DiscreteOpt(length=size,
fitness_fn=fitness,
crossover=crossover)
return problem
def test_random_neighbor_max2():
"""Test random_neighbor method when max_val is equal to 2"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
x = np.array([0, 0, 1, 1, 1])
problem.set_state(x)
neigh = problem.random_neighbor()
sum_diff = np.sum(np.abs(x - neigh))
assert (len(neigh) == 5 and sum_diff == 1)
def test_reproduce_mut1_max_gt2():
"""Test reproduce method when mutation_prob is 1 and max_val is
greater than 2"""
problem = DiscreteOpt(5, OneMax(),
maximize=True,
max_val=3)
problem._crossover = OnePointCrossOver(problem)
father = np.array([0, 0, 0, 0, 0])
mother = np.array([2, 2, 2, 2, 2])
child = problem.reproduce(father, mother, mutation_prob=1)
assert (len(child) == 5 and sum(child) > 0 and sum(child) < 10)
def test_random_neighbor_max_gt2():
"""Test random_neighbor method when max_val is greater than 2"""
problem = DiscreteOpt(5, OneMax(), maximize=True, max_val=5)
x = np.array([0, 1, 2, 3, 4])
problem.set_state(x)
neigh = problem.random_neighbor()
abs_diff = np.abs(x - neigh)
abs_diff[abs_diff > 0] = 1
sum_diff = np.sum(abs_diff)
assert (len(neigh) == 5 and sum_diff == 1)
def test_hill_climb_discrete_min():
"""Test hill_climb function for a discrete minimization problem"""
problem = DiscreteOpt(5, OneMax(), maximize=False)
best_state, best_fitness, _ = hill_climb(problem, restarts=20)
x = np.array([0, 0, 0, 0, 0])
assert (np.array_equal(best_state, x) and best_fitness == 0)
def test_mimic_discrete_min():
"""Test mimic function for a discrete minimization problem"""
problem = DiscreteOpt(5, OneMax(), maximize=False)
best_state, best_fitness, _ = mimic(problem, max_attempts=50)
x = np.array([0, 0, 0, 0, 0])
assert (np.array_equal(best_state, x) and best_fitness == 0)
def test_genetic_alg_discrete_max():
"""Test genetic_alg function for a discrete maximization problem"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
best_state, best_fitness, _ = genetic_alg(problem, max_attempts=50)
x = np.array([1, 1, 1, 1, 1])
assert (np.array_equal(best_state, x) and best_fitness == 5)
def test_hill_climb_max_iters():
"""Test hill_climb function with max_iters less than infinite"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
x = np.array([0, 0, 0, 0, 0])
best_state, best_fitness, _ = hill_climb(problem,
max_iters=1,
restarts=0,
init_state=x)
assert best_fitness == 1
def test_random_hill_climb_discrete_max():
"""Test random_hill_climb function for a discrete maximization
problem"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
best_state, best_fitness, _ = random_hill_climb(problem,
max_attempts=10,
restarts=20)
x = np.array([1, 1, 1, 1, 1])
assert (np.array_equal(best_state, x) and best_fitness == 5)
def test_simulated_annealing_max_iters():
"""Test simulated_annealing function with max_iters less than
infinite"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
x = np.array([0, 0, 0, 0, 0])
best_state, best_fitness, _ = simulated_annealing(problem,
max_attempts=1,
max_iters=1,
init_state=x)
assert best_fitness == 1
def test_random_pop():
"""Test random_pop method"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
problem.random_pop(100)
pop = problem.get_population()
pop_fitness = problem.get_pop_fitness()
assert (np.shape(pop)[0] == 100 and np.shape(pop)[1] == 5
and np.sum(pop) > 0 and np.sum(pop) < 500
and len(pop_fitness) == 100)
def test_find_neighbors_max2():
"""Test find_neighbors method when max_val is equal to 2"""
problem = DiscreteOpt(5, OneMax(), maximize=True, max_val=2)
x = np.array([0, 1, 0, 1, 0])
problem.set_state(x)
problem.find_neighbors()
neigh = np.array([[1, 1, 0, 1, 0],
[0, 0, 0, 1, 0],
[0, 1, 1, 1, 0],
[0, 1, 0, 0, 0],
[0, 1, 0, 1, 1]])
assert np.array_equal(np.array(problem.neighbors), neigh)
def test_find_top_pct_min():
"""Test find_top_pct method for a minimization problem"""
problem = DiscreteOpt(5, OneMax(), maximize=False)
pop = np.array([[0, 0, 0, 0, 1],
[1, 0, 1, 0, 1],
[1, 1, 1, 1, 0],
[1, 0, 0, 0, 1],
[100, 0, 0, 0, 0],
[0, 0, 0, 0, 0],
[1, 1, 1, 1, 1],
[0, 0, 0, 0, -50]])
problem.set_population(pop)
problem.find_top_pct(keep_pct=0.25)
x = np.array([[0, 0, 0, 0, 0],
[0, 0, 0, 0, -50]])
assert np.array_equal(problem.get_keep_sample(), x)
def test_sample_pop():
"""Test sample_pop method"""
problem = DiscreteOpt(5, OneMax(), maximize=True)
pop = np.array([[0, 0, 0, 0, 1],
[1, 0, 1, 0, 1],
[1, 1, 1, 1, 0],
[1, 0, 0, 0, 1],
[0, 0, 0, 0, 0],
[1, 1, 1, 1, 1]])
problem.keep_sample = pop
problem.eval_node_probs()
sample = problem.sample_pop(100)
assert (np.shape(sample)[0] == 100 and np.shape(sample)[1] == 5
and np.sum(sample) > 0 and np.sum(sample) < 500)
def generate(seed, size=20, t_pct=0.15):
np.random.seed(seed)
fitness = FourPeaks(t_pct=t_pct)
problem = DiscreteOpt(length=size, fitness_fn=fitness)
return problem
# plt.figure()
fit_func = SixPeaks()
# plot time over problem sizes
lens = [len(init_state) for init_state in init_states]
RHC_vals = []
SA_vals = []
GA_vals = []
MIMIC_vals = []
RHC_times = []
SA_times = []
GA_times = []
MIMIC_times = []
plt.figure()
for init_state in init_states:
problem = DiscreteOpt(length=len(init_state), fitness_fn=fit_func)
fit, duration, _ = run_RHC_2(problem, init_state)
RHC_vals.append(fit)
RHC_times.append(duration)
fit, duration, _ = run_SA_2(problem, init_state)
SA_vals.append(fit)
SA_times.append(duration)
fit, duration, _ = run_GA_2(problem, init_state)
GA_vals.append(fit)
GA_times.append(duration)
fit, duration, _ = run_MIMIC_2(problem, init_state)
MIMIC_vals.append(fit)
MIMIC_times.append(duration)
plt.plot(lens, RHC_vals, label="rhc")
plt.plot(lens, SA_vals, label="sa")
plt.plot(lens, GA_vals, label="ga")
