1 #-*-coding:iso-8859-1-*-
3 # Copyright (C) 2008-2011 EDF R&D
5 # This library is free software; you can redistribute it and/or
6 # modify it under the terms of the GNU Lesser General Public
7 # License as published by the Free Software Foundation; either
8 # version 2.1 of the License.
10 # This library is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 # Lesser General Public License for more details.
15 # You should have received a copy of the GNU Lesser General Public
16 # License along with this library; if not, write to the Free Software
17 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19 # See http://www.salome-platform.org/ or email : webmaster.salome@opencascade.com
23 from daCore import BasicObjects, PlatformInfo
24 m = PlatformInfo.SystemUsage()
29 if logging.getLogger().level < 30:
31 message = scipy.optimize.tnc.MSG_ALL
35 message = scipy.optimize.tnc.MSG_NONE
38 # ==============================================================================
39 class ElementaryAlgorithm(BasicObjects.Algorithm):
41 BasicObjects.Algorithm.__init__(self)
43 logging.debug("%s Initialisation"%self._name)
45 def run(self, Xb=None, Y=None, H=None, M=None, R=None, B=None, Q=None, Parameters=None):
47 Calcul de l'estimateur 3D-VAR
49 logging.debug("%s Lancement"%self._name)
50 logging.debug("%s Taille mémoire utilisée de %.1f Mo"%(self._name, m.getUsedMemory("Mo")))
52 # Opérateur d'observation
53 # -----------------------
54 Hm = H["Direct"].appliedTo
55 Ht = H["Adjoint"].appliedInXTo
57 # Utilisation éventuelle d'un vecteur H(Xb) précalculé
58 # ----------------------------------------------------
59 if H["AppliedToX"] is not None and H["AppliedToX"].has_key("HXb"):
60 logging.debug("%s Utilisation de HXb"%self._name)
61 HXb = H["AppliedToX"]["HXb"]
63 logging.debug("%s Calcul de Hm(Xb)"%self._name)
65 HXb = numpy.asmatrix(HXb).flatten().T
67 # Calcul de l'innovation
68 # ----------------------
70 logging.debug("%s Innovation d = %s"%(self._name, d))
72 # Précalcul des inversion appellée dans les fonction-coût et gradient
73 # -------------------------------------------------------------------
77 # Définition de la fonction-coût
78 # ------------------------------
80 _X = numpy.asmatrix(x).flatten().T
81 logging.info("%s CostFunction X = %s"%(self._name, numpy.asmatrix( _X ).flatten()))
83 _HX = numpy.asmatrix(_HX).flatten().T
84 Jb = 0.5 * (_X - Xb).T * BI * (_X - Xb)
85 Jo = 0.5 * (Y - _HX).T * RI * (Y - _HX)
86 J = float( Jb ) + float( Jo )
87 logging.info("%s CostFunction Jb = %s"%(self._name, Jb))
88 logging.info("%s CostFunction Jo = %s"%(self._name, Jo))
89 logging.info("%s CostFunction J = %s"%(self._name, J))
90 self.StoredVariables["CurrentState"].store( _X.A1 )
91 self.StoredVariables["CostFunctionJb"].store( Jb )
92 self.StoredVariables["CostFunctionJo"].store( Jo )
93 self.StoredVariables["CostFunctionJ" ].store( J )
96 def GradientOfCostFunction(x):
97 _X = numpy.asmatrix(x).flatten().T
98 logging.info("%s GradientOfCostFunction X = %s"%(self._name, numpy.asmatrix( _X ).flatten()))
100 _HX = numpy.asmatrix(_HX).flatten().T
101 GradJb = BI * (_X - Xb)
102 GradJo = - Ht( (_X, RI * (Y - _HX)) )
103 GradJ = numpy.asmatrix( GradJb ).flatten().T + numpy.asmatrix( GradJo ).flatten().T
104 logging.debug("%s GradientOfCostFunction GradJb = %s"%(self._name, numpy.asmatrix( GradJb ).flatten()))
105 logging.debug("%s GradientOfCostFunction GradJo = %s"%(self._name, numpy.asmatrix( GradJo ).flatten()))
106 logging.debug("%s GradientOfCostFunction GradJ = %s"%(self._name, numpy.asmatrix( GradJ ).flatten()))
109 # Point de démarrage de l'optimisation : Xini = Xb
110 # ------------------------------------
111 if type(Xb) is type(numpy.matrix([])):
112 Xini = Xb.A1.tolist()
115 logging.debug("%s Point de démarrage Xini = %s"%(self._name, Xini))
117 # Paramètres de pilotage
118 # ----------------------
119 # Potentiels : "Bounds", "Minimizer", "MaximumNumberOfSteps", "ProjectedGradientTolerance", "GradientNormTolerance", "InnerMinimizer"
120 if Parameters.has_key("Bounds") and (type(Parameters["Bounds"]) is type([]) or type(Parameters["Bounds"]) is type(())) and (len(Parameters["Bounds"]) > 0):
121 Bounds = Parameters["Bounds"]
124 MinimizerList = ["LBFGSB","TNC", "CG", "NCG", "BFGS"]
125 if Parameters.has_key("Minimizer") and (Parameters["Minimizer"] in MinimizerList):
126 Minimizer = str( Parameters["Minimizer"] )
129 logging.warning("%s Unknown or undefined minimizer, replaced by the default one \"%s\""%(self._name,Minimizer))
130 logging.debug("%s Minimiseur utilisé = %s"%(self._name, Minimizer))
131 if Parameters.has_key("MaximumNumberOfSteps") and (Parameters["MaximumNumberOfSteps"] > -1):
132 maxiter = int( Parameters["MaximumNumberOfSteps"] )
135 logging.debug("%s Nombre maximal de pas d'optimisation = %s"%(self._name, str(maxiter)))
136 if Parameters.has_key("CostDecrementTolerance") and (Parameters["CostDecrementTolerance"] > 0):
137 ftol = float(Parameters["CostDecrementTolerance"])
142 logging.debug("%s Diminution relative minimale du cout lors de l'arret = %s"%(self._name, str(1./factr)))
143 if Parameters.has_key("ProjectedGradientTolerance") and (Parameters["ProjectedGradientTolerance"] > -1):
144 pgtol = float(Parameters["ProjectedGradientTolerance"])
147 logging.debug("%s Maximum des composantes du gradient projete lors de l'arret = %s"%(self._name, str(pgtol)))
148 if Parameters.has_key("GradientNormTolerance") and (Parameters["GradientNormTolerance"] > -1):
149 gtol = float(Parameters["GradientNormTolerance"])
152 logging.debug("%s Maximum des composantes du gradient lors de l'arret = %s"%(self._name, str(gtol)))
153 InnerMinimizerList = ["CG", "NCG", "BFGS"]
154 if Parameters.has_key("InnerMinimizer") and (Parameters["InnerMinimizer"] in InnerMinimizerList):
155 InnerMinimizer = str( Parameters["Minimizer"] )
157 InnerMinimizer = "BFGS"
158 logging.debug("%s Minimiseur interne utilisé = %s"%(self._name, InnerMinimizer))
159 logging.debug("%s Norme du gradient lors de l'arret = %s"%(self._name, str(gtol)))
161 # Minimisation de la fonctionnelle
162 # --------------------------------
163 if Minimizer == "LBFGSB":
164 Minimum, J_optimal, Informations = scipy.optimize.fmin_l_bfgs_b(
167 fprime = GradientOfCostFunction,
175 nfeval = Informations['funcalls']
176 rc = Informations['warnflag']
177 elif Minimizer == "TNC":
178 Minimum, nfeval, rc = scipy.optimize.fmin_tnc(
181 fprime = GradientOfCostFunction,
189 elif Minimizer == "CG":
190 Minimum, fopt, nfeval, grad_calls, rc = scipy.optimize.fmin_cg(
193 fprime = GradientOfCostFunction,
200 elif Minimizer == "NCG":
201 Minimum, fopt, nfeval, grad_calls, hcalls, rc = scipy.optimize.fmin_ncg(
204 fprime = GradientOfCostFunction,
211 elif Minimizer == "BFGS":
212 Minimum, fopt, gopt, Hopt, nfeval, grad_calls, rc = scipy.optimize.fmin_bfgs(
215 fprime = GradientOfCostFunction,
223 raise ValueError("Error in Minimizer name: %s"%Minimizer)
225 # Correction pour pallier a un bug de TNC sur le retour du Minimum
226 # ----------------------------------------------------------------
227 StepMin = numpy.argmin( self.StoredVariables["CostFunctionJ"].valueserie() )
228 MinJ = self.StoredVariables["CostFunctionJ"].valueserie(step = StepMin)
229 Minimum = self.StoredVariables["CurrentState"].valueserie(step = StepMin)
231 logging.debug("%s %s Step of min cost = %s"%(self._name, Minimizer, StepMin))
232 logging.debug("%s %s Minimum cost = %s"%(self._name, Minimizer, MinJ))
233 logging.debug("%s %s Minimum state = %s"%(self._name, Minimizer, Minimum))
234 logging.debug("%s %s Nb of F = %s"%(self._name, Minimizer, nfeval))
235 logging.debug("%s %s RetCode = %s"%(self._name, Minimizer, rc))
237 # Calcul de l'analyse
238 # --------------------
239 Xa = numpy.asmatrix(Minimum).T
240 logging.debug("%s Analyse Xa = %s"%(self._name, Xa))
242 self.StoredVariables["Analysis"].store( Xa.A1 )
243 self.StoredVariables["Innovation"].store( d.A1 )
245 logging.debug("%s Taille mémoire utilisée de %.1f Mo"%(self._name, m.getUsedMemory("MB")))
246 logging.debug("%s Terminé"%self._name)
250 # ==============================================================================
251 if __name__ == "__main__":
252 print '\n AUTODIAGNOSTIC \n'