XC Open source finite element analysis program
KrylovNewton.h
1 //----------------------------------------------------------------------------
2 // XC program; finite element analysis code
3 // for structural analysis and design.
4 //
5 // Copyright (C) Luis Claudio Pérez Tato
6 //
7 // This program derives from OpenSees <http://opensees.berkeley.edu>
8 // developed by the «Pacific earthquake engineering research center».
9 //
10 // Except for the restrictions that may arise from the copyright
11 // of the original program (see copyright_opensees.txt)
12 // XC is free software: you can redistribute it and/or modify
13 // it under the terms of the GNU General Public License as published by
14 // the Free Software Foundation, either version 3 of the License, or
15 // (at your option) any later version.
16 //
17 // This software is distributed in the hope that it will be useful, but
18 // WITHOUT ANY WARRANTY; without even the implied warranty of
19 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 // GNU General Public License for more details.
21 //
22 //
23 // You should have received a copy of the GNU General Public License
24 // along with this program.
25 // If not, see <http://www.gnu.org/licenses/>.
26 //----------------------------------------------------------------------------
27 /* ****************************************************************** **
28 ** OpenSees - Open System for Earthquake Engineering Simulation **
29 ** Pacific Earthquake Engineering Research Center **
30 ** **
31 ** **
32 ** (C) Copyright 1999, The Regents of the University of California **
33 ** All Rights Reserved. **
34 ** **
35 ** Commercial use of this program without express permission of the **
36 ** University of California, Berkeley, is strictly prohibited. See **
37 ** file 'COPYRIGHT' in main directory for information on usage and **
38 ** redistribution, and for a DISCLAIMER OF ALL WARRANTIES. **
39 ** **
40 ** Developed by: **
41 ** Frank McKenna (fmckenna@ce.berkeley.edu) **
42 ** Gregory L. Fenves (fenves@ce.berkeley.edu) **
43 ** Filip C. Filippou (filippou@ce.berkeley.edu) **
44 ** **
45 ** ****************************************************************** */
46 
47 // $Revision: 1.8 $
48 // $Date: 2005/11/29 22:42:41 $
49 // $Source: /usr/local/cvs/OpenSees/SRC/analysis/algorithm/equiSolnAlgo/KrylovNewton.h,v $
50 
51 #ifndef KrylovNewton_h
52 #define KrylovNewton_h
53 
54 // Written: MHS
55 // Created: June 2001
56 //
57 // Description: This file contains the class definition for
58 // KrylovNewton. KrylovNewton is a class which uses a Krylov
59 // subspace accelerator on the modified Newton method.
60 // The accelerator is described by Carlson and Miller in
61 // "Design and Application of a 1D GWMFE Code"
62 // from SIAM Journal of Scientific Computing (Vol. 19, No. 3,
63 // pp. 728-765, May 1998)
64 
65 #include "EquiSolnAlgo.h"
66 #include "utility/matrix/Vector.h"
67 
68 namespace XC {
69 
71 //
79  {
80  private:
81  int tangent;
82 
83  // Storage for update vectors
84  std::vector<Vector> v;
85  // Storage for subspace vectors
86  std::vector<Vector> Av;
87 
88  // Array data sent to LAPACK subroutine
89  Vector AvData;
90  Vector rData;
91  Vector work;
92 
93  // Length of work array
94  int lwork;
95 
96  // Size information
97  int numEqns;
98  int maxDimension;
99 
100  // Private lsq routine to do Krylov updates
101  // dimension is the current dimension of the subspace
102  int leastSquares(int dimension);
103 
104  friend class SoluMethod;
105  friend class FEM_ObjectBroker;
106  KrylovNewton(SoluMethod *,int tangent = CURRENT_TANGENT, int maxDim = 3);
107  virtual SolutionAlgorithm *getCopy(void) const;
108  public:
109  int solveCurrentStep(void);
110 
111  virtual int sendSelf(CommParameters &);
112  virtual int recvSelf(const CommParameters &);
113  void Print(std::ostream &s, int flag =0);
114  };
115 inline SolutionAlgorithm *KrylovNewton::getCopy(void) const
116  { return new KrylovNewton(*this); }
117 } // end of XC namespace
118 
119 #endif
120 
121 
EquiSolnAlgo is an abstract base class, i.e. no objects of it&#39;s type can be created. Its subclasses deifine the sequence of operations to be performed in the analysis by static equilibrium of a finite element model.
Definition: EquiSolnAlgo.h:85
Definition: Vector.h:82
Solution procedure for the finite element problem. The solution procedure is definde by specifiying: ...
Definition: SoluMethod.h:76
int solveCurrentStep(void)
resuelve el paso actual.
Definition: KrylovNewton.cpp:80
FEM_ObjectBroker is is an object broker class for the finite element method. All methods are virtual ...
Definition: FEM_ObjectBroker.h:138
KrylovNewton is a class which uses a Krylov subspace accelerator on the modified Newton method...
Definition: KrylovNewton.h:78
Communication parameters between processes.
Definition: CommParameters.h:65
================================================================================
Definition: ContinuaReprComponent.h:34
Base class for solution algorithms.
Definition: SolutionAlgorithm.h:86