def test_random_parzen_estimators():
minimize(
function_to_optimize,
optimization_problem,
optimizer_type="parzen_estimator",
number_of_evaluation=35,
)
def test_passing_optimizer_directly():
minimize(
function_to_optimize,
optimization_problem,
optimizer_type=RandomOptimizer,
number_of_evaluation=5,
)
def test_suggestions_with_global_seed_reset():
suggestions = []
def eval_(x: float) -> float:
suggestions.append(x)
np.random.seed(0)
return 1.0
params = [{
"name": "x",
"category": "uniform",
"search_space": {
"low": 0,
"high": 1
}
}]
minimize(eval_,
params,
optimizer_type="parzen_estimator",
number_of_evaluation=5)
assert len(suggestions) == len(set(suggestions))
def test_random_uniform():
optimization_problem = [{
"name": "x",
"category": "uniform",
"search_space": {
"low": 0,
"high": np.pi
}
}]
best_sample = minimize(f,
optimization_problem,
optimizer_type="random",
number_of_evaluation=100)
assert np.abs(best_sample["x"] - (np.pi / 2)) < 0.1
def test_random_uniform():
np.random.seed(0)
minimize(function_to_optimize,
optimization_problem,
optimizer_type="random",
number_of_evaluation=5)