Improve tests consistency for convergence on GitHub Actions (#450)

* rebase smt 2.0

* improve EGO tests consistency for convergence on GitHub Actions

* Update test_ego.py

Author: Paul-Saves

smt/applications/tests/test_ego.py

@@ -115,13 +115,13 @@ def test_rosenbrock_2D(self):
         criterion = "LCB"  #'EI' or 'SBO' or 'LCB'
         random_state = 42
         design_space = DesignSpace(xlimits, seed=random_state)
-
         xdoe = FullFactorial(xlimits=xlimits)(10)
         ego = EGO(
+            n_start=30,
             xdoe=xdoe,
             n_iter=n_iter,
             criterion=criterion,
-            surrogate=KRG(design_space=design_space, print_global=False),
+            surrogate=KRG(design_space=design_space, n_start=25, print_global=False),
             random_state=random_state,
         )
 
@@ -156,7 +156,8 @@ def test_rosenbrock_2D_parallel(self):
         fun = Rosenbrock(ndim=2)
         xlimits = fun.xlimits
         criterion = "LCB"  #'EI' or 'SBO' or 'LCB'
-        design_space = DesignSpace(xlimits)
+        random_state = 42
+        design_space = DesignSpace(xlimits, seed=random_state)
 
         xdoe = FullFactorial(xlimits=xlimits)(10)
         qEI = "KB"
@@ -168,7 +169,7 @@ def test_rosenbrock_2D_parallel(self):
             n_parallel=n_parallel,
             qEI=qEI,
             evaluator=ParallelEvaluator(),
-            random_state=42,
+            random_state=random_state,
         )
 
         x_opt, y_opt, _, _, _ = ego.optimize(fun=fun)
@@ -244,8 +245,7 @@ def test_branin_2D_mixed_parallel(self):
             ],
             seed=random_state,
         )
-        sm = KRG(design_space=design_space, print_global=False)
-
+        sm = KRG(design_space=design_space, print_global=False, n_start=25)
         mixint = MixedIntegerContext(design_space)
         sampling = mixint.build_sampling_method(FullFactorial)
         xdoe = sampling(10)
@@ -256,6 +256,7 @@ def test_branin_2D_mixed_parallel(self):
             criterion=criterion,
             n_parallel=n_parallel,
             qEI=qEI,
+            n_start=30,
             evaluator=ParallelEvaluator(),
             surrogate=sm,
             random_state=random_state,
@@ -275,11 +276,13 @@ def test_branin_2D_mixed(self):
         n_iter = 20
         fun = Branin(ndim=2)
         xlimits = fun.xlimits
+        random_state = 42
         design_space = DesignSpace(
             [
                 IntegerVariable(*xlimits[0]),
                 FloatVariable(*xlimits[1]),
-            ]
+            ],
+            seed=random_state,
         )
         criterion = "EI"  #'EI' or 'SBO' or 'LCB'
 
@@ -293,7 +296,7 @@ def test_branin_2D_mixed(self):
             criterion=criterion,
             surrogate=sm,
             enable_tunneling=False,
-            random_state=42,
+            random_state=random_state,
         )
 
         x_opt, y_opt, _, _, _ = ego.optimize(fun=fun)
@@ -414,10 +417,12 @@ def test_ego_mixed_integer_gower_distance(self):
 
         criterion = "EI"  #'EI' or 'SBO' or 'LCB'
         ego = EGO(
+            n_start=30,
             n_iter=n_iter,
             criterion=criterion,
             xdoe=xdoe,
             surrogate=KRG(
+                n_start=25,
                 design_space=design_space,
                 categorical_kernel=MixIntKernelType.GOWER,
                 print_global=False,
@@ -431,6 +436,8 @@ def test_ego_mixed_integer_gower_distance(self):
 
     @unittest.skipIf(int(os.getenv("RUN_SLOW", 0)) < 1, "too slow")
     def test_ego_mixed_integer_hierarchical_NN(self):
+        random_state = 42
+
         def f_neu(x1, x2, x3, x4):
             if x4 == 0:
                 return 2 * x1 + x2 - 0.5 * x3
@@ -652,6 +659,7 @@ def f_hv(X):
                     raise ValueError
             return np.array(y)
 
+        random_state = 0
         ds = DesignSpace(
             [
                 CategoricalVariable(values=[0, 1, 2, 3]),  # meta
@@ -665,7 +673,8 @@ def f_hv(X):
                 IntegerVariable(0, 2),
                 IntegerVariable(0, 2),
                 IntegerVariable(0, 2),
-            ]
+            ],
+            seed=random_state,
         )
 
         # x4 is acting if meta == 1, 3
@@ -678,7 +687,6 @@ def f_hv(X):
         ds.declare_decreed_var(decreed_var=8, meta_var=0, meta_value=[0, 1])
 
         n_doe = 25
-        random_state = 0
         ds.seed = random_state
         Xt, x_is_active = ds.sample_valid_x(n_doe)
 
@@ -712,20 +720,22 @@ def f_hv(X):
 
     def test_ego_mixed_integer_homo_gaussian(self):
         n_iter = 15
+        random_state = 42
         design_space = DesignSpace(
             [
                 FloatVariable(-5, 5),
                 CategoricalVariable(["blue", "red", "green"]),
                 CategoricalVariable(["large", "small"]),
                 IntegerVariable(0, 2),
-            ]
+            ],
+            seed=random_state,
         )
         n_doe = 5
         sampling = MixedIntegerSamplingMethod(
             LHS,
             design_space,
             criterion="ese",
-            random_state=42,
+            random_state=random_state,
             output_in_folded_space=True,
         )
         xdoe = sampling(n_doe)
@@ -741,7 +751,7 @@ def test_ego_mixed_integer_homo_gaussian(self):
                 print_global=False,
             ),
             enable_tunneling=False,
-            random_state=42,
+            random_state=random_state,
         )
         _, y_opt, _, _, _ = ego.optimize(fun=TestEGO.function_test_mixed_integer)
 
@@ -783,7 +793,7 @@ def test_ego_mixed_integer_homo_gaussian_pls(self):
             xdoe=xdoe,
             surrogate=sm,
             enable_tunneling=False,
-            random_state=42,
+            random_state=random_state,
         )
         _, y_opt, _, _, _ = ego.optimize(fun=TestEGO.function_test_mixed_integer)
 
@@ -798,7 +808,6 @@ def test_ydoe_option(self):
         design_space = DesignSpace(xlimits, seed=random_state)
         xdoe = FullFactorial(xlimits=xlimits)(10)
         ydoe = fun(xdoe)
-
         ego = EGO(
             xdoe=xdoe,
             ydoe=ydoe,

smt/applications/tests/test_mixed_integer.py

@@ -34,34 +34,6 @@
 
 
 class TestMixedInteger(unittest.TestCase):
-    def test_krg_mixed_3D_INT(self):
-        design_space = DesignSpace(
-            [
-                FloatVariable(-10, 10),
-                CategoricalVariable(["blue", "red", "green"]),
-                IntegerVariable(-10, 10),
-            ]
-        )
-
-        mixint = MixedIntegerContext(design_space)
-
-        sm = mixint.build_kriging_model(KRG(print_prediction=False))
-        sampling = mixint.build_sampling_method(LHS, criterion="m")
-
-        fun = Sphere(ndim=3)
-        xt = sampling(20)
-        yt = fun(xt)
-        sm.set_training_values(xt, yt)
-        sm.train()
-
-        eq_check = True
-        for i in range(xt.shape[0]):
-            if abs(float(xt[i, :][2]) - int(float(xt[i, :][2]))) > 10e-8:
-                eq_check = False
-            if not (xt[i, :][1] == 0 or xt[i, :][1] == 1 or xt[i, :][1] == 2):
-                eq_check = False
-        self.assertTrue(eq_check)
-
     def test_krg_mixed_3D(self):
         design_space = DesignSpace(
             [