1 # -*- coding: utf-8 -*-
3 # Copyright (C) 2008-2024 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
21 # Author: Jean-Philippe Argaud, jean-philippe.argaud@edf.fr, EDF R&D
26 __author__ = "Jean-Philippe ARGAUD"
28 import numpy, scipy, scipy.optimize, scipy.version
29 from daCore.NumericObjects import HessienneEstimation, QuantilesEstimations
30 from daCore.PlatformInfo import vt, vfloat
32 # ==============================================================================
33 def psas3dvar(selfA, Xb, Y, U, HO, CM, R, B, __storeState = False):
40 Hm = HO["Direct"].appliedTo
42 if HO["AppliedInX"] is not None and "HXb" in HO["AppliedInX"]:
43 HXb = numpy.asarray(Hm( Xb, HO["AppliedInX"]["HXb"] ))
45 HXb = numpy.asarray(Hm( Xb ))
46 HXb = HXb.reshape((-1, 1))
47 if Y.size != HXb.size:
48 raise ValueError("The size %i of observations Y and %i of observed calculation H(X) are different, they have to be identical."%(Y.size, HXb.size)) # noqa: E501
49 if max(Y.shape) != max(HXb.shape):
50 raise ValueError("The shapes %s of observations Y and %s of observed calculation H(X) are different, they have to be identical."%(Y.shape, HXb.shape)) # noqa: E501
52 Ht = HO["Tangent"].asMatrix(Xb)
53 Ht = Ht.reshape(Y.size, Xb.size) # ADAO & check shape
58 if selfA._toStore("JacobianMatrixAtBackground"):
59 selfA.StoredVariables["JacobianMatrixAtBackground"].store( Ht )
61 Xini = numpy.zeros(Y.size)
63 # Définition de la fonction-coût
64 # ------------------------------
67 _W = numpy.asarray(w).reshape((-1, 1))
68 if selfA._parameters["StoreInternalVariables"] or \
69 selfA._toStore("CurrentState") or \
70 selfA._toStore("CurrentOptimum"):
71 selfA.StoredVariables["CurrentState"].store( Xb + BHT @ _W )
72 if selfA._toStore("SimulatedObservationAtCurrentState") or \
73 selfA._toStore("SimulatedObservationAtCurrentOptimum"):
74 selfA.StoredVariables["SimulatedObservationAtCurrentState"].store( Hm( Xb + BHT @ _W ) )
75 if selfA._toStore("InnovationAtCurrentState"):
76 selfA.StoredVariables["InnovationAtCurrentState"].store( Innovation )
78 Jb = vfloat( 0.5 * _W.T @ (HBHTpR @ _W) )
79 Jo = vfloat( - _W.T @ Innovation )
82 selfA.StoredVariables["CurrentIterationNumber"].store( len(selfA.StoredVariables["CostFunctionJ"]) )
83 selfA.StoredVariables["CostFunctionJb"].store( Jb )
84 selfA.StoredVariables["CostFunctionJo"].store( Jo )
85 selfA.StoredVariables["CostFunctionJ" ].store( J )
86 if selfA._toStore("IndexOfOptimum") or \
87 selfA._toStore("CurrentOptimum") or \
88 selfA._toStore("CostFunctionJAtCurrentOptimum") or \
89 selfA._toStore("CostFunctionJbAtCurrentOptimum") or \
90 selfA._toStore("CostFunctionJoAtCurrentOptimum") or \
91 selfA._toStore("SimulatedObservationAtCurrentOptimum"):
92 IndexMin = numpy.argmin( selfA.StoredVariables["CostFunctionJ"][nbPreviousSteps:] ) + nbPreviousSteps
93 if selfA._toStore("IndexOfOptimum"):
94 selfA.StoredVariables["IndexOfOptimum"].store( IndexMin )
95 if selfA._toStore("CurrentOptimum"):
96 selfA.StoredVariables["CurrentOptimum"].store( selfA.StoredVariables["CurrentState"][IndexMin] )
97 if selfA._toStore("SimulatedObservationAtCurrentOptimum"):
98 selfA.StoredVariables["SimulatedObservationAtCurrentOptimum"].store( selfA.StoredVariables["SimulatedObservationAtCurrentState"][IndexMin] ) # noqa: E501
99 if selfA._toStore("CostFunctionJbAtCurrentOptimum"):
100 selfA.StoredVariables["CostFunctionJbAtCurrentOptimum"].store( selfA.StoredVariables["CostFunctionJb"][IndexMin] ) # noqa: E501
101 if selfA._toStore("CostFunctionJoAtCurrentOptimum"):
102 selfA.StoredVariables["CostFunctionJoAtCurrentOptimum"].store( selfA.StoredVariables["CostFunctionJo"][IndexMin] ) # noqa: E501
103 if selfA._toStore("CostFunctionJAtCurrentOptimum"):
104 selfA.StoredVariables["CostFunctionJAtCurrentOptimum" ].store( selfA.StoredVariables["CostFunctionJ" ][IndexMin] ) # noqa: E501
107 def GradientOfCostFunction(w):
108 _W = numpy.asarray(w).reshape((-1, 1))
110 GradJo = - Innovation
111 GradJ = numpy.ravel( GradJb ) + numpy.ravel( GradJo )
114 # Minimisation de la fonctionnelle
115 # --------------------------------
116 nbPreviousSteps = selfA.StoredVariables["CostFunctionJ"].stepnumber()
118 if selfA._parameters["Minimizer"] == "LBFGSB":
119 if vt("0.19") <= vt(scipy.version.version) <= vt("1.4.99"):
120 import daAlgorithms.Atoms.lbfgsb14hlt as optimiseur
121 elif vt("1.5.0") <= vt(scipy.version.version) <= vt("1.7.99"):
122 import daAlgorithms.Atoms.lbfgsb17hlt as optimiseur
123 elif vt("1.8.0") <= vt(scipy.version.version) <= vt("1.8.99"):
124 import daAlgorithms.Atoms.lbfgsb18hlt as optimiseur
125 elif vt("1.9.0") <= vt(scipy.version.version) <= vt("1.10.99"):
126 import daAlgorithms.Atoms.lbfgsb19hlt as optimiseur
127 elif vt("1.11.0") <= vt(scipy.version.version) <= vt("1.11.99"):
128 import daAlgorithms.Atoms.lbfgsb111hlt as optimiseur
129 elif vt("1.12.0") <= vt(scipy.version.version) <= vt("1.12.99"):
130 import daAlgorithms.Atoms.lbfgsb112hlt as optimiseur
131 elif vt("1.13.0") <= vt(scipy.version.version) <= vt("1.13.99"):
132 import daAlgorithms.Atoms.lbfgsb113hlt as optimiseur
133 elif vt("1.14.0") <= vt(scipy.version.version) <= vt("1.14.99"):
134 import daAlgorithms.Atoms.lbfgsb114hlt as optimiseur
136 import scipy.optimize as optimiseur
137 Minimum, J_optimal, Informations = optimiseur.fmin_l_bfgs_b(
140 fprime = GradientOfCostFunction,
142 maxfun = selfA._parameters["MaximumNumberOfIterations"] - 1,
143 factr = selfA._parameters["CostDecrementTolerance"] * 1.e14,
144 pgtol = selfA._parameters["ProjectedGradientTolerance"],
145 iprint = selfA._parameters["optiprint"],
147 # nfeval = Informations['funcalls']
148 # rc = Informations['warnflag']
149 elif selfA._parameters["Minimizer"] == "TNC":
150 Minimum, nfeval, rc = scipy.optimize.fmin_tnc(
153 fprime = GradientOfCostFunction,
155 maxfun = selfA._parameters["MaximumNumberOfIterations"],
156 pgtol = selfA._parameters["ProjectedGradientTolerance"],
157 ftol = selfA._parameters["CostDecrementTolerance"],
158 messages = selfA._parameters["optmessages"],
160 elif selfA._parameters["Minimizer"] == "CG":
161 Minimum, fopt, nfeval, grad_calls, rc = scipy.optimize.fmin_cg(
164 fprime = GradientOfCostFunction,
166 maxiter = selfA._parameters["MaximumNumberOfIterations"],
167 gtol = selfA._parameters["GradientNormTolerance"],
168 disp = selfA._parameters["optdisp"],
171 elif selfA._parameters["Minimizer"] == "NCG":
172 Minimum, fopt, nfeval, grad_calls, hcalls, rc = scipy.optimize.fmin_ncg(
175 fprime = GradientOfCostFunction,
177 maxiter = selfA._parameters["MaximumNumberOfIterations"],
178 avextol = selfA._parameters["CostDecrementTolerance"],
179 disp = selfA._parameters["optdisp"],
182 elif selfA._parameters["Minimizer"] == "BFGS":
183 Minimum, fopt, gopt, Hopt, nfeval, grad_calls, rc = scipy.optimize.fmin_bfgs(
186 fprime = GradientOfCostFunction,
188 maxiter = selfA._parameters["MaximumNumberOfIterations"],
189 gtol = selfA._parameters["GradientNormTolerance"],
190 disp = selfA._parameters["optdisp"],
194 raise ValueError("Error in minimizer name: %s is unkown"%selfA._parameters["Minimizer"])
196 IndexMin = numpy.argmin( selfA.StoredVariables["CostFunctionJ"][nbPreviousSteps:] ) + nbPreviousSteps
197 MinJ = selfA.StoredVariables["CostFunctionJ"][IndexMin]
199 # Correction pour pallier a un bug de TNC sur le retour du Minimum
200 # ----------------------------------------------------------------
201 if selfA._parameters["StoreInternalVariables"] or selfA._toStore("CurrentState"):
202 Minimum = selfA.StoredVariables["CurrentState"][IndexMin]
204 Minimum = Xb + BHT @ Minimum.reshape((-1, 1))
208 selfA._setInternalState("Xn", Xa)
209 # --------------------------
211 selfA.StoredVariables["Analysis"].store( Xa )
213 if selfA._toStore("OMA") or \
214 selfA._toStore("InnovationAtCurrentAnalysis") or \
215 selfA._toStore("SigmaObs2") or \
216 selfA._toStore("SimulationQuantiles") or \
217 selfA._toStore("SimulatedObservationAtOptimum"):
218 if selfA._toStore("SimulatedObservationAtCurrentState"):
219 HXa = selfA.StoredVariables["SimulatedObservationAtCurrentState"][IndexMin]
220 elif selfA._toStore("SimulatedObservationAtCurrentOptimum"):
221 HXa = selfA.StoredVariables["SimulatedObservationAtCurrentOptimum"][-1]
224 oma = Y - numpy.asarray(HXa).reshape((-1, 1))
226 if selfA._toStore("APosterioriCovariance") or \
227 selfA._toStore("SimulationQuantiles") or \
228 selfA._toStore("JacobianMatrixAtOptimum") or \
229 selfA._toStore("KalmanGainAtOptimum"):
230 HtM = HO["Tangent"].asMatrix(ValueForMethodForm = Xa)
231 HtM = HtM.reshape(Y.size, Xa.size) # ADAO & check shape
232 if selfA._toStore("APosterioriCovariance") or \
233 selfA._toStore("SimulationQuantiles") or \
234 selfA._toStore("KalmanGainAtOptimum"):
235 HaM = HO["Adjoint"].asMatrix(ValueForMethodForm = Xa)
236 HaM = HaM.reshape(Xa.size, Y.size) # ADAO & check shape
237 if selfA._toStore("APosterioriCovariance") or \
238 selfA._toStore("SimulationQuantiles"):
241 A = HessienneEstimation(selfA, Xa.size, HaM, HtM, BI, RI)
242 if selfA._toStore("APosterioriCovariance"):
243 selfA.StoredVariables["APosterioriCovariance"].store( A )
244 if selfA._toStore("JacobianMatrixAtOptimum"):
245 selfA.StoredVariables["JacobianMatrixAtOptimum"].store( HtM )
246 if selfA._toStore("KalmanGainAtOptimum"):
247 if (Y.size <= Xb.size):
248 KG = B * HaM * (R + numpy.dot(HtM, B * HaM)).I
249 elif (Y.size > Xb.size):
250 KG = (BI + numpy.dot(HaM, RI * HtM)).I * HaM * RI
251 selfA.StoredVariables["KalmanGainAtOptimum"].store( KG )
253 # Calculs et/ou stockages supplémentaires
254 # ---------------------------------------
255 if selfA._toStore("Innovation") or \
256 selfA._toStore("SigmaObs2") or \
257 selfA._toStore("MahalanobisConsistency") or \
258 selfA._toStore("OMB"):
260 if selfA._toStore("Innovation"):
261 selfA.StoredVariables["Innovation"].store( Innovation )
262 if selfA._toStore("BMA"):
263 selfA.StoredVariables["BMA"].store( numpy.ravel(Xb) - numpy.ravel(Xa) )
264 if selfA._toStore("OMA"):
265 selfA.StoredVariables["OMA"].store( oma )
266 if selfA._toStore("InnovationAtCurrentAnalysis"):
267 selfA.StoredVariables["InnovationAtCurrentAnalysis"].store( oma )
268 if selfA._toStore("OMB"):
269 selfA.StoredVariables["OMB"].store( Innovation )
270 if selfA._toStore("SigmaObs2"):
271 TraceR = R.trace(Y.size)
272 selfA.StoredVariables["SigmaObs2"].store( vfloat( (Innovation.T @ oma) ) / TraceR )
273 if selfA._toStore("MahalanobisConsistency"):
274 selfA.StoredVariables["MahalanobisConsistency"].store( float( 2. * MinJ / Innovation.size ) )
275 if selfA._toStore("SimulationQuantiles"):
276 QuantilesEstimations(selfA, A, Xa, HXa, Hm, HtM)
277 if selfA._toStore("SimulatedObservationAtBackground"):
278 selfA.StoredVariables["SimulatedObservationAtBackground"].store( HXb )
279 if selfA._toStore("SimulatedObservationAtOptimum"):
280 selfA.StoredVariables["SimulatedObservationAtOptimum"].store( HXa )
284 # ==============================================================================
285 if __name__ == "__main__":
286 print('\n AUTODIAGNOSTIC\n')