def test_sca_run():
def square(x):
return np.sum(x**2)
def hook(optimizer, space, function):
return
new_function = function.Function(pointer=square)
hyperparams = {
'r_min': 0,
'r_max': 2,
'a': 3
}
new_sca = sca.SCA(hyperparams=hyperparams)
search_space = search.SearchSpace(n_agents=10, n_iterations=100,
n_variables=2, lower_bound=[0, 0],
upper_bound=[10, 10])
history = new_sca.run(search_space, new_function, pre_evaluation_hook=hook)
assert len(history.agents) > 0
assert len(history.best_agent) > 0
best_fitness = history.best_agent[-1][1]
assert best_fitness <= constants.TEST_EPSILON, 'The algorithm sca failed to converge.'
def test_sca_hyperparams():
hyperparams = {
'r_min': 0,
'r_max': 2,
'a': 3,
}
new_sca = sca.SCA(hyperparams=hyperparams)
assert new_sca.r_min == 0
assert new_sca.r_max == 2
assert new_sca.a == 3
def test_sca_hyperparams_setter():
new_sca = sca.SCA()
try:
new_sca.r_min = 'a'
except:
new_sca.r_min = 0.1
try:
new_sca.r_min = -1
except:
new_sca.r_min = 0.1
assert new_sca.r_min == 0.1
try:
new_sca.r_max = 'b'
except:
new_sca.r_max = 2
try:
new_sca.r_max = -1
except:
new_sca.r_max = 2
try:
new_sca.r_max = 0
except:
new_sca.r_max = 2
assert new_sca.r_max == 2
try:
new_sca.a = 'c'
except:
new_sca.a = 0.5
try:
new_sca.a = -1
except:
new_sca.a = 0.5
assert new_sca.a == 0.5
def test_sca_update_position():
new_sca = sca.SCA()
position = new_sca._update_position(1, 1, 0.5, 0.5, 0.5, 0.5)
assert position > 0
def test_sca_build():
new_sca = sca.SCA()
assert new_sca.built == True