Fix anyoptimization#207: Added DE tests and fixed another selection i…

…ssue.

pymoo/algorithms/soo/nonconvex/de.py

@@ -138,28 +138,30 @@ def _do(self, pop, n_select, n_parents, **kwargs):
         # create offsprings and add it to the data of the algorithm
         P = RandomSelection().do(pop, n_select, n_parents)
 
+        n = len(P)
+
         if variant == "best":
             P[:, 0] = 0
         elif variant == "current-to-best":
-            P[:, 0] = np.arange(len(pop))
+            P[:, 0] = np.arange(n)
             P[:, 1] = 0
-            P[:, 2] = np.arange(len(pop))
+            P[:, 2] = np.arange(n)
         elif variant == "current-to-rand":
-            P[:, 0] = np.arange(len(pop))
-            P[:, 2] = np.arange(len(pop))
+            P[:, 0] = np.arange(n)
+            P[:, 2] = np.arange(n)
         elif variant == "rand-to-best":
             P[:, 1] = 0
-            P[:, 2] = np.arange(len(pop))
+            P[:, 2] = np.arange(n)
         elif variant == "current-to-pbest":
             # best 10% of the population
-            n_pbest = int(np.ceil(0.1 * len(pop)))
+            n_pbest = max(1, int(np.ceil(0.1 * n)))
 
             # the corresponding indices to select from
             pbest = np.arange(n_pbest)
 
-            P[:, 0] = np.arange(len(pop))
-            P[:, 1] = np.random.choice(pbest, len(pop))
-            P[:, 2] = np.arange(len(pop))
+            P[:, 0] = np.arange(n)
+            P[:, 1] = np.random.choice(pbest, n)
+            P[:, 2] = np.arange(n)
 
         return P
 

tests/algorithms/test_de.py

@@ -0,0 +1,25 @@
+import pytest
+
+from pymoo.algorithms.soo.nonconvex.de import DE
+from pymoo.factory import get_problem
+from pymoo.operators.sampling.lhs import LHS
+from pymoo.optimize import minimize
+
+
+@pytest.mark.parametrize('selection', ["rand", "best", "rand-to-best", "current-to-rand", "current-to-best", "current-to-pbest"])
+@pytest.mark.parametrize('crossover', ["bin", "exp"])
+def test_de(selection, crossover):
+    problem = get_problem("ackley", n_var=10)
+
+    algorithm = DE(
+        pop_size=100,
+        sampling=LHS(),
+        variant=f"DE/{selection}/1/{crossover}")
+
+    ret = minimize(problem,
+             algorithm,
+             ('n_gen', 20),
+             seed=1,
+             verbose=True)
+
+    assert len(ret.opt) > 0

tests/scratch.py

@@ -1,19 +1,20 @@
-from pymoo.algorithms.moo.nsga2 import NSGA2
+from pymoo.algorithms.soo.nonconvex.de import DE
 from pymoo.factory import get_problem
+from pymoo.operators.sampling.lhs import LHS
 from pymoo.optimize import minimize
-from pymoo.visualization.scatter import Scatter
 
-problem = get_problem("zdt1")
+problem = get_problem("ackley", n_var=10)
 
-algorithm = NSGA2(pop_size=100)
+algorithm = DE(
+pop_size=100,
+sampling=LHS(),
+variant='DE/current-to-pbest/1/bin')
 
 res = minimize(problem,
-               algorithm,
-               ('n_gen', 200),
-               seed=1,
-               verbose=True)
+algorithm,
+seed=1,
+verbose=False)
+
+
+
 
-plot = Scatter()
-plot.add(problem.pareto_front(), plot_type="line", color="black", alpha=0.7)
-plot.add(res.F, color="red")
-plot.show()