From: Jean-Philippe ARGAUD <jean-philippe.argaud@edf.fr>
Date: Mon, 12 Nov 2012 11:48:05 +0000 (+0100)
Subject: Improving the treatment of special cases for minimum
X-Git-Tag: V6_6_0~6
X-Git-Url: http://git.salome-platform.org/gitweb/?a=commitdiff_plain;h=2cc2c16e29a010cd3eaba886e4688b46ed93cd4d;p=modules%2Fadao.git

Improving the treatment of special cases for minimum
---

diff --git a/src/daComposant/daAlgorithms/3DVAR.py b/src/daComposant/daAlgorithms/3DVAR.py
index 0cc504e..ebfd8f9 100644
--- a/src/daComposant/daAlgorithms/3DVAR.py
+++ b/src/daComposant/daAlgorithms/3DVAR.py
@@ -176,6 +176,8 @@ class ElementaryAlgorithm(BasicObjects.Algorithm):
         #
         # Minimisation de la fonctionnelle
         # --------------------------------
+        n0 = self.StoredVariables["CostFunctionJ"].stepnumber()
+        #
         if self._parameters["Minimizer"] == "LBFGSB":
             Minimum, J_optimal, Informations = scipy.optimize.fmin_l_bfgs_b(
                 func        = CostFunction,
@@ -238,7 +240,7 @@ class ElementaryAlgorithm(BasicObjects.Algorithm):
         else:
             raise ValueError("Error in Minimizer name: %s"%self._parameters["Minimizer"])
         #
-        StepMin = numpy.argmin( self.StoredVariables["CostFunctionJ"].valueserie() )
+        StepMin = numpy.argmin( self.StoredVariables["CostFunctionJ"].valueserie()[n0:] )
         MinJ    = self.StoredVariables["CostFunctionJ"].valueserie(step = StepMin)
         #
         # Correction pour pallier a un bug de TNC sur le retour du Minimum
diff --git a/src/daComposant/daAlgorithms/NonLinearLeastSquares.py b/src/daComposant/daAlgorithms/NonLinearLeastSquares.py
index 27cec60..e39c698 100644
--- a/src/daComposant/daAlgorithms/NonLinearLeastSquares.py
+++ b/src/daComposant/daAlgorithms/NonLinearLeastSquares.py
@@ -206,6 +206,8 @@ class ElementaryAlgorithm(BasicObjects.Algorithm):
         #
         # Minimisation de la fonctionnelle
         # --------------------------------
+        n0 = self.StoredVariables["CostFunctionJ"].stepnumber()
+        #
         if self._parameters["Minimizer"] == "LBFGSB":
             Minimum, J_optimal, Informations = scipy.optimize.fmin_l_bfgs_b(
                 func        = CostFunction,
@@ -280,7 +282,7 @@ class ElementaryAlgorithm(BasicObjects.Algorithm):
         else:
             raise ValueError("Error in Minimizer name: %s"%self._parameters["Minimizer"])
         #
-        StepMin = numpy.argmin( self.StoredVariables["CostFunctionJ"].valueserie() )
+        StepMin = numpy.argmin( self.StoredVariables["CostFunctionJ"].valueserie()[n0:] )
         MinJ    = self.StoredVariables["CostFunctionJ"].valueserie(step = StepMin)
         #
         # Correction pour pallier a un bug de TNC sur le retour du Minimum