def test_statistics(self):
sample1, cost1 = np.array([[0, 1, 1, 1, 0]]), 3.0
sample2, cost2 = np.array([[0, 0, 0, 1, 1]]), 2.0
sample3, cost3 = np.array([[0, 1, 1, 1, 1]]), 4.0
func = GbCostFunction(OneMax)
exp_cost1 = func(sample1)
exp_cost2 = func(sample2)
exp_cost3 = func(sample3)
evals, argmin, argmax, min, max, mean, stdev = func.statistics()
self.assertEqual((evals, argmin, argmax, min, max, mean),
(3, 2, 3, cost2, cost3, 3.0))
self.assertAlmostEqual(stdev, np.sqrt(2.0 / 3.0), 5)
def test_basic(self):
cga = Cga()
cga.setup(20)
cga.cost_func = GbCostFunction(OneMax, var_size=10)
result = cga.search()
self.assertTrue(result.params.all())
self.assertGreaterEqual(result.cost, 8)
def test_basic(self):
pbil = Pbil()
pbil.setup(20)
pbil.cost_func = GbCostFunction(OneMax, var_size=10)
result = pbil.search()
self.assertTrue(result.params.all())
self.assertGreaterEqual(result.cost, 8)
def test_basic_search_zero(self):
"""Test class for the Bmda algorithm"""
bmda = Bmda()
bmda.setup(40)
bmda.cost_func = GbCostFunction(ZeroMax, var_size=10)
result = bmda.search()
self.assertTrue((result.params + 1).all())
self.assertGreaterEqual(result.cost, 8.0)
def test_basic_search_onemax(self):
"""Test class for the Bmda algorithm"""
bmda = Bmda()
bmda.setup(10, 40)
bmda.cost_func = GbCostFunction(OneMax)
result = bmda.search()
self.assertGreaterEqual(result.params.sum(), 8.0)
self.assertGreaterEqual(result.cost, 8.0)
def test_basic_search_onemax_mi_method(self):
"""Test class for the Bmda algorithm"""
bmda = Bmda()
bmda.setup(40, dependency_method=DependencyMethod.mi)
bmda.cost_func = GbCostFunction(OneMax, var_size=10)
result = bmda.search()
self.assertGreaterEqual(result.params.sum(), 8.0)
self.assertGreaterEqual(result.cost, 8.0)
def test_reset(self):
cga = Cga()
cga.setup(20)
cga.cost_func = GbCostFunction(OneMax, var_size=10)
cga.search()
cga.reset()
self.assertEqual(cga.iters, 0)
self.assertNotEqual(cga.distr, None)
self.assertEqual(cga.cost_func.evals, 0)
def test_basic_search_onemax(self):
"""Test class for the Bmda algorithm"""
var_size = 10
bmda = Bmda()
bmda.setup(40)
bmda.cost_func = GbCostFunction(OneMax, var_size=var_size)
result = bmda.search()
self.assertTrue((result.params + 1).all())
self.assertGreaterEqual(result.cost, var_size * 0.8)
def test_reset(self):
pbil = Pbil()
pbil.setup(20)
pbil.cost_func = GbCostFunction(OneMax, var_size=10)
pbil.search()
pbil.reset()
self.assertEqual(pbil.iters, 0)
self.assertNotEqual(pbil.distr, None)
self.assertEqual(pbil.cost_func.evals, 0)
def test_reset(self):
tilda = Tilda()
tilda.setup(20)
tilda.cost_func = GbCostFunction(OneMax, var_size=10)
result = tilda.search()
tilda.reset()
self.assertEqual(tilda.iters, 0)
self.assertNotEqual(tilda.distr, None)
self.assertEqual(tilda.cost_func.evals, 0)
def test_reset(self):
"""Test if the reset function allows a new algorithm execution."""
bmda = Bmda()
bmda.setup(40)
bmda.cost_func = GbCostFunction(OneMax, var_size=10)
bmda.search()
bmda.reset()
self.assertEqual(bmda.iters, 0)
self.assertIsNotNone(bmda.distr)
self.assertEqual(bmda.cost_func.evals, 0)
def test_basic(self):
total_runs = 10
opttester = GbBlackBoxTester()
optimizer = Cga()
optimizer.setup(20)
optimizer.cost_func = GbCostFunction(OneMax, var_size=10)
run_results, test_results, candidates = opttester.test(
optimizer, total_runs)
self.assertEqual(len(run_results), total_runs)
self.assertEqual(12, len(test_results))
def test_ready(self):
pbil = Pbil()
pbil.setup(20)
self.assertFalse(pbil.ready())
pbil.cost_func = GbCostFunction(OneMax, var_size=10)
self.assertTrue(pbil.ready())
def test_custom_function(self):
sample = np.array([[0, 1, 0, 1, 0]])
cust = GbCostFunction(
script=
"def mycustomfunction(solution):\n\treturn solution.sum() - 1.0")
self.assertEqual(cust.cost(sample), 1.0)
def test_onemax(self):
sample = np.array([[0, 1, 0, 1, 0]])
func = GbCostFunction(OneMax)
self.assertEqual(func(sample), 2)
def test_ready(self):
cga = Cga()
cga.setup(20)
self.assertFalse(cga.ready())
cga.cost_func = GbCostFunction(OneMax, var_size=10)
self.assertTrue(cga.ready())
def test_basic(self):
pbilc = Pbilc()
pbilc.setup(30, learning_rate=0.5, max_evals=1000)
pbilc.cost_func = GbCostFunction(OneMax, var_size=10)
result = pbilc.search()
self.assertGreaterEqual(result.cost, 8)
def test_zero(self):
sample = np.array([[0, 1, 0, 1, 0]])
func = GbCostFunction(ZeroMax)
self.assertEqual(func(sample), 3)
def test_basic(self):
tilda = Tilda()
tilda.setup(40, learning_rate=0.3, max_evals=10000)
tilda.cost_func = GbCostFunction(OneMax, var_size=10)
result = tilda.search()
self.assertGreaterEqual(result.cost, 8.0)
