Add parameters.reset_initial_value

pybop/optimisers/base_optimiser.py

@@ -129,7 +129,7 @@ def set_base_options(self):
         """
         # Set initial values, if x0 is None, initial values are unmodified.
         self.parameters.update(initial_values=self.unset_options.pop("x0", None))
-        self.x0 = self.parameters.initial_value()
+        self.x0 = self.parameters.reset_initial_value()
 
         # Set default bounds (for all or no parameters)
         self.bounds = self.unset_options.pop("bounds", self.parameters.get_bounds())

pybop/parameters/parameter.py

@@ -194,13 +194,10 @@ def get_initial_value(self) -> float:
                 self.update(initial_value=sample)
             else:
                 warnings.warn(
-                    "Initial value or Prior are None, proceeding without initial value.",
+                    "Initial value and prior are None, proceeding without an initial value.",
                     UserWarning,
                     stacklevel=2,
                 )
-        else:
-            # Make sure to always reset the current value as well
-            self.update(value=self.initial_value)
 
         return self.initial_value
 
@@ -424,6 +421,21 @@ def initial_value(self) -> np.ndarray:
 
         return np.asarray(initial_values)
 
+    def reset_initial_value(self) -> np.ndarray:
+        """
+        Reset and return the initial value of each parameter.
+        """
+        initial_values = []
+
+        for param in self.param.values():
+            initial_value = param.get_initial_value()
+            if initial_value is not None:
+                # Reset the current value as well
+                param.update(value=initial_value)
+            initial_values.append(initial_value)
+
+        return np.asarray(initial_values)
+
     def current_value(self) -> np.ndarray:
         """
         Return the current value of each parameter.

pybop/problems/base_problem.py

@@ -50,6 +50,8 @@ def __init__(
             )
 
         self.parameters = parameters
+        self.parameters.reset_initial_value()
+
         self._model = model
         self.check_model = check_model
         if isinstance(signal, str):

pybop/problems/design_problem.py

@@ -56,7 +56,6 @@ def __init__(
         super().__init__(parameters, model, check_model, signal, additional_variables)
         self.experiment = experiment
         self.init_soc = init_soc
-        self.parameters.initial_value()
 
         # Add an example dataset for plotting comparison
         sol = self.evaluate(self.parameters.as_dict("initial"))

pybop/problems/fitting_problem.py

@@ -45,7 +45,6 @@ def __init__(
 
         super().__init__(parameters, model, check_model, signal, additional_variables)
         self._dataset = dataset.data
-        self.parameters.initial_value()
         self._n_parameters = len(self.parameters)
         self._init_ocv = None
         if init_ocv is not None:

tests/unit/test_parameters.py

@@ -224,7 +224,7 @@ def test_initial_values_without_attributes(self):
         )
         with pytest.warns(
             UserWarning,
-            match="Initial value or Prior are None, proceeding without initial value.",
+            match="Initial value and prior are None, proceeding without an initial value.",
         ):
             sample = parameter.initial_value()