SMTorg · relf · Oct 30, 2025 · Oct 24, 2025 · Oct 24, 2025 · Oct 24, 2025
diff --git a/doc/requirements.txt b/doc/requirements.txt
@@ -3,10 +3,10 @@ packaging
 numpy
 scipy
 scikit-learn
-pyDOE3
+pyDOE3 >= 1.5.0
 numpydoc
 matplotlib
 jenn >= 1.0.2, <2.0
-egobox ~= 0.23.0
+egobox ~= 0.33.0
 smt-design-space-ext
 git+https://github.com/hwangjt/sphinx_auto_embed.git  # for doc generation
diff --git a/requirements.txt b/requirements.txt
@@ -2,8 +2,8 @@ Cython
 packaging
 numpy
 scipy
-scikit-learn >=1.4
-pyDOE3
+scikit-learn
+pyDOE3 >= 1.5.0
 numba           # JIT compiler
 matplotlib      # used in examples and tests
 pytest          # tests runner

diff --git a/setup.py b/setup.py
@@ -114,7 +114,7 @@
     install_requires=[
         "packaging",
         "scikit-learn",
-        "pyDOE3",
+        "pyDOE3 >= 1.5.0",
         "scipy",
         "jenn",
     ],

diff --git a/smt/applications/cckrg.py b/smt/applications/cckrg.py
@@ -297,9 +297,7 @@ def grad_minus_cooperative_reduced_likelihood_function(active_log10t):
                 theta_all_loops = np.vstack((theta0, theta0_rand))
 
                 if self.options["n_start"] > 1:
-                    sampling = LHS(
-                        xlimits=theta_limits, criterion="maximin", random_state=41
-                    )
+                    sampling = LHS(xlimits=theta_limits, criterion="maximin", seed=41)
                     theta_lhs_loops = sampling(self.options["n_start"])
                     theta_all_loops = np.vstack((theta_all_loops, theta_lhs_loops))
 

diff --git a/smt/applications/ego.py b/smt/applications/ego.py
@@ -7,7 +7,6 @@
 
 import time
 from types import FunctionType
-from typing import Optional
 
 import numpy as np
 from scipy.optimize import minimize
@@ -29,7 +28,7 @@ class Evaluator(object):
     User can derive this interface and override the run() method to implement custom multiprocessing.
     """
 
-    def run(self, fun, x, design_space: Optional = None):
+    def run(self, fun, x, design_space=None):
         """
         Evaluates fun at x.
 
@@ -48,7 +47,7 @@ def run(self, fun, x, design_space: Optional = None):
         """
         if (design_space is not None) and np.any(design_space.is_conditionally_acting):
             x_corr, eval_is_acting = design_space.correct_get_acting(x)
-            return fun(x_corr, eval_is_acting)
+            return fun(x_corr, eval_is_acting=eval_is_acting)
 
         else:
             return fun(x)
@@ -123,9 +122,9 @@ def _initialize(self):
             desc="SMT kriging-based surrogate model used internaly",
         )
         self.options.declare(
-            "random_state",
-            types=(type(None), int, np.random.RandomState),
-            desc="Numpy RandomState object or seed number which controls random draws",
+            "seed",
+            types=(type(None), int, np.random.Generator),
+            desc="Numpy Generator object or seed number which controls random draws",
         )
 
     def optimize(self, fun):
@@ -281,7 +280,7 @@ def _setup_optimizer(self, fun):
             self.gpr.options["is_ri"] = self.options["is_ri"]
         self.design_space: BaseDesignSpace = self.gpr.design_space
         if isinstance(self.design_space, DesignSpace):
-            self.design_space.seed = self.options["random_state"]
+            self.design_space.seed = self.options["seed"]
 
         # Handle mixed integer optimization
         is_continuous = self.design_space.is_all_cont
@@ -301,7 +300,7 @@ def _setup_optimizer(self, fun):
                 work_in_folded_space=True,
             )
             self._sampling = self.mixint.build_sampling_method(
-                random_state=self.options["random_state"],
+                seed=self.options["seed"],
             )
 
         else:
@@ -310,7 +309,7 @@ def _setup_optimizer(self, fun):
                 LHS,
                 self.design_space,
                 criterion="ese",
-                random_state=self.options["random_state"],
+                seed=self.options["seed"],
             )
             self._sampling = lambda n: sampling(n)
 
@@ -467,7 +466,7 @@ def _get_virtual_point(self, x, y_data):
             conf = -3.0
 
         if qEI == "KBRand":
-            conf = np.random.randn()
+            conf = np.random.standard_normal()
         pred = self.gpr.predict_values(x)
         var = self.gpr.predict_variances(x)
 

diff --git a/smt/applications/mfck.py b/smt/applications/mfck.py
@@ -223,7 +223,7 @@ def train(self):
                 sampling = LHS(
                     xlimits=np.stack((lower_bounds, upper_bounds), axis=1),
                     criterion="ese",
-                    random_state=0,
+                    seed=0,
                 )
                 theta_lhs_loops = sampling(self.options["n_start"])
                 theta0 = np.vstack((theta_ini, theta_lhs_loops))

diff --git a/smt/applications/mfk.py b/smt/applications/mfk.py
@@ -36,7 +36,9 @@
 
 
 class NestedLHS(object):
-    def __init__(self, nlevel, xlimits=None, design_space=None, random_state=None):
+    def __init__(
+        self, nlevel, xlimits=None, design_space=None, seed=None, random_state=None
+    ):
         """
         Constructor where values of options can be passed in.
 
@@ -51,11 +53,32 @@ def __init__(self, nlevel, xlimits=None, design_space=None, random_state=None):
         design_space : DesignSpace
             The design space with type and bounds for every design variable
 
-        random_state : Numpy RandomState object or seed number which controls random draws
+        seed : Numpy Generator object or seed number which controls random draws
+
+        random_state : DEPRECATED use seed. Numpy RandomState object or seed number which controls random draws
 
         """
         self.nlevel = nlevel
         self.random_state = random_state
+        if random_state is None:
+            self.random_state = np.random.default_rng()
+        elif isinstance(random_state, np.random.RandomState):
+            raise ValueError(
+                "np.random.RandomState object is not handled anymore. Please use seed and np.random.Generator"
+            )
+        elif isinstance(random_state, int):
+            warnings.warn(
+                "Passing a seed or integer to random_state is deprecated "
+                "and will raise an error in a future version. Please "
+                "use seed parameter",
+                DeprecationWarning,
+                stacklevel=2,
+            )
+            self.random_state = np.random.default_rng(random_state)
+
+        if seed is not None:
+            self.random_state = np.random.default_rng(seed)
+
         if xlimits is None and design_space is None:
             raise ValueError(
                 "Either xlimits or design_space should be specified to have bounds for the sampling."
@@ -102,7 +125,7 @@ def __call__(self, nb_samples_hifi):
         p0 = LHS(
             xlimits=np.array(self.design_space.get_unfolded_num_bounds()),
             criterion="ese",
-            random_state=self.random_state,
+            seed=self.random_state,
         )
         p0nt0 = p0(nt[0])
         if self.design_space:
@@ -115,7 +138,7 @@ def __call__(self, nb_samples_hifi):
             p = LHS(
                 xlimits=np.array(self.design_space.get_unfolded_num_bounds()),
                 criterion="ese",
-                random_state=self.random_state,
+                seed=self.random_state,
             )
             pnti = p(nt[i])
             if self.design_space:

diff --git a/smt/applications/mixed_integer.py b/smt/applications/mixed_integer.py
@@ -41,10 +41,17 @@ def __init__(self, sampling_method_class, design_space, **kwargs):
             or not (enum masks)
         """
         self._design_space = design_space
+
         if "random_state" in kwargs:
-            self._design_space.random_state = kwargs["random_state"]
-        elif self._design_space.random_state is None:
-            self._design_space.random_state = 42
+            raise ValueError(
+                "random_state is not handled anymore. Please use seed parameter"
+            )
+
+        if "seed" in kwargs:
+            self._design_space.seed = kwargs["seed"]
+        elif self._design_space.seed is None:
+            self._design_space.seed = 42
+
         self._unfolded_xlimits = design_space.get_unfolded_num_bounds()
         self._output_in_folded_space = kwargs.get("output_in_folded_space", True)
         kwargs.pop("output_in_folded_space", None)
@@ -57,7 +64,7 @@ def _compute(self, nt, return_is_acting=False):
         x_doe, is_acting = self._design_space.sample_valid_x(
             nt,
             unfolded=not self._output_in_folded_space,
-            random_state=self._design_space.random_state,
+            seed=self._design_space.seed,
         )
         if return_is_acting:
             return x_doe, is_acting
@@ -330,15 +337,15 @@ def __init__(self, design_space, work_in_folded_space=True):
     def design_space(self) -> BaseDesignSpace:
         return self._design_space
 
-    def build_sampling_method(self, random_state=None):
+    def build_sampling_method(self, seed=None):
         """
         Build Mixed Integer LHS ESE sampler.
         """
         return_folded = self._work_in_folded_space
 
         def sample(n):
             x, _ = self._design_space.sample_valid_x(
-                n, unfolded=not return_folded, random_state=random_state
+                n, unfolded=not return_folded, seed=seed
             )
             return x
 

diff --git a/smt/applications/smfck.py b/smt/applications/smfck.py
@@ -275,7 +275,7 @@ def train(self):
                 sampling = LHS(
                     xlimits=np.stack((lower_bounds, upper_bounds), axis=1),
                     criterion="ese",
-                    random_state=0,
+                    seed=0,
                 )
                 theta_lhs_loops = sampling(self.options["n_start"])
                 theta0 = np.vstack((theta_ini, theta_lhs_loops))

diff --git a/smt/applications/tests/test_cckrg.py b/smt/applications/tests/test_cckrg.py
@@ -37,11 +37,9 @@ def run_cckrg_example():
         ncomp = 3
 
         # Initial sampling
-        samp = LHS(xlimits=prob.xlimits, random_state=42)
-        np.random.seed(0)
+        samp = LHS(xlimits=prob.xlimits, seed=42)
         xt = samp(50)
         yt = prob(xt)
-        np.random.seed(1)
 
         # Random design variable to component allocation
         comps = [*range(ncomp)]