1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2011 Gael Guennebaud <gael.guennebaud@inria.fr>
5 // Copyright (C) 2012 Désiré Nuentsa-Wakam <desire.nuentsa_wakam@inria.fr>
6 //
7 // This Source Code Form is subject to the terms of the Mozilla
8 // Public License v. 2.0. If a copy of the MPL was not distributed
9 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
10
11 #ifndef EIGEN_BICGSTAB_H
12 #define EIGEN_BICGSTAB_H
13
14 namespace Eigen {
15
16 namespace internal {
17
18 /** \internal Low-level bi conjugate gradient stabilized algorithm
19 * \param mat The matrix A
20 * \param rhs The right hand side vector b
21 * \param x On input and initial solution, on output the computed solution.
22 * \param precond A preconditioner being able to efficiently solve for an
23 * approximation of Ax=b (regardless of b)
24 * \param iters On input the max number of iteration, on output the number of performed iterations.
25 * \param tol_error On input the tolerance error, on output an estimation of the relative error.
26 * \return false in the case of numerical issue, for example a break down of BiCGSTAB.
27 */
28 template<typename MatrixType, typename Rhs, typename Dest, typename Preconditioner>
bicgstab(const MatrixType & mat,const Rhs & rhs,Dest & x,const Preconditioner & precond,int & iters,typename Dest::RealScalar & tol_error)29 bool bicgstab(const MatrixType& mat, const Rhs& rhs, Dest& x,
30 const Preconditioner& precond, int& iters,
31 typename Dest::RealScalar& tol_error)
32 {
33 using std::sqrt;
34 using std::abs;
35 typedef typename Dest::RealScalar RealScalar;
36 typedef typename Dest::Scalar Scalar;
37 typedef Matrix<Scalar,Dynamic,1> VectorType;
38 RealScalar tol = tol_error;
39 int maxIters = iters;
40
41 int n = mat.cols();
42 VectorType r = rhs - mat * x;
43 VectorType r0 = r;
44
45 RealScalar r0_sqnorm = r0.squaredNorm();
46 RealScalar rhs_sqnorm = rhs.squaredNorm();
47 if(rhs_sqnorm == 0)
48 {
49 x.setZero();
50 return true;
51 }
52 Scalar rho = 1;
53 Scalar alpha = 1;
54 Scalar w = 1;
55
56 VectorType v = VectorType::Zero(n), p = VectorType::Zero(n);
57 VectorType y(n), z(n);
58 VectorType kt(n), ks(n);
59
60 VectorType s(n), t(n);
61
62 RealScalar tol2 = tol*tol;
63 RealScalar eps2 = NumTraits<Scalar>::epsilon()*NumTraits<Scalar>::epsilon();
64 int i = 0;
65 int restarts = 0;
66
67 while ( r.squaredNorm()/rhs_sqnorm > tol2 && i<maxIters )
68 {
69 Scalar rho_old = rho;
70
71 rho = r0.dot(r);
72 if (abs(rho) < eps2*r0_sqnorm)
73 {
74 // The new residual vector became too orthogonal to the arbitrarily choosen direction r0
75 // Let's restart with a new r0:
76 r0 = r;
77 rho = r0_sqnorm = r.squaredNorm();
78 if(restarts++ == 0)
79 i = 0;
80 }
81 Scalar beta = (rho/rho_old) * (alpha / w);
82 p = r + beta * (p - w * v);
83
84 y = precond.solve(p);
85
86 v.noalias() = mat * y;
87
88 alpha = rho / r0.dot(v);
89 s = r - alpha * v;
90
91 z = precond.solve(s);
92 t.noalias() = mat * z;
93
94 RealScalar tmp = t.squaredNorm();
95 if(tmp>RealScalar(0))
96 w = t.dot(s) / tmp;
97 else
98 w = Scalar(0);
99 x += alpha * y + w * z;
100 r = s - w * t;
101 ++i;
102 }
103 tol_error = sqrt(r.squaredNorm()/rhs_sqnorm);
104 iters = i;
105 return true;
106 }
107
108 }
109
110 template< typename _MatrixType,
111 typename _Preconditioner = DiagonalPreconditioner<typename _MatrixType::Scalar> >
112 class BiCGSTAB;
113
114 namespace internal {
115
116 template< typename _MatrixType, typename _Preconditioner>
117 struct traits<BiCGSTAB<_MatrixType,_Preconditioner> >
118 {
119 typedef _MatrixType MatrixType;
120 typedef _Preconditioner Preconditioner;
121 };
122
123 }
124
125 /** \ingroup IterativeLinearSolvers_Module
126 * \brief A bi conjugate gradient stabilized solver for sparse square problems
127 *
128 * This class allows to solve for A.x = b sparse linear problems using a bi conjugate gradient
129 * stabilized algorithm. The vectors x and b can be either dense or sparse.
130 *
131 * \tparam _MatrixType the type of the sparse matrix A, can be a dense or a sparse matrix.
132 * \tparam _Preconditioner the type of the preconditioner. Default is DiagonalPreconditioner
133 *
134 * The maximal number of iterations and tolerance value can be controlled via the setMaxIterations()
135 * and setTolerance() methods. The defaults are the size of the problem for the maximal number of iterations
136 * and NumTraits<Scalar>::epsilon() for the tolerance.
137 *
138 * This class can be used as the direct solver classes. Here is a typical usage example:
139 * \code
140 * int n = 10000;
141 * VectorXd x(n), b(n);
142 * SparseMatrix<double> A(n,n);
143 * // fill A and b
144 * BiCGSTAB<SparseMatrix<double> > solver;
145 * solver.compute(A);
146 * x = solver.solve(b);
147 * std::cout << "#iterations: " << solver.iterations() << std::endl;
148 * std::cout << "estimated error: " << solver.error() << std::endl;
149 * // update b, and solve again
150 * x = solver.solve(b);
151 * \endcode
152 *
153 * By default the iterations start with x=0 as an initial guess of the solution.
154 * One can control the start using the solveWithGuess() method. Here is a step by
155 * step execution example starting with a random guess and printing the evolution
156 * of the estimated error:
157 * * \code
158 * x = VectorXd::Random(n);
159 * solver.setMaxIterations(1);
160 * int i = 0;
161 * do {
162 * x = solver.solveWithGuess(b,x);
163 * std::cout << i << " : " << solver.error() << std::endl;
164 * ++i;
165 * } while (solver.info()!=Success && i<100);
166 * \endcode
167 * Note that such a step by step excution is slightly slower.
168 *
169 * \sa class SimplicialCholesky, DiagonalPreconditioner, IdentityPreconditioner
170 */
171 template< typename _MatrixType, typename _Preconditioner>
172 class BiCGSTAB : public IterativeSolverBase<BiCGSTAB<_MatrixType,_Preconditioner> >
173 {
174 typedef IterativeSolverBase<BiCGSTAB> Base;
175 using Base::mp_matrix;
176 using Base::m_error;
177 using Base::m_iterations;
178 using Base::m_info;
179 using Base::m_isInitialized;
180 public:
181 typedef _MatrixType MatrixType;
182 typedef typename MatrixType::Scalar Scalar;
183 typedef typename MatrixType::Index Index;
184 typedef typename MatrixType::RealScalar RealScalar;
185 typedef _Preconditioner Preconditioner;
186
187 public:
188
189 /** Default constructor. */
190 BiCGSTAB() : Base() {}
191
192 /** Initialize the solver with matrix \a A for further \c Ax=b solving.
193 *
194 * This constructor is a shortcut for the default constructor followed
195 * by a call to compute().
196 *
197 * \warning this class stores a reference to the matrix A as well as some
198 * precomputed values that depend on it. Therefore, if \a A is changed
199 * this class becomes invalid. Call compute() to update it with the new
200 * matrix A, or modify a copy of A.
201 */
202 BiCGSTAB(const MatrixType& A) : Base(A) {}
203
204 ~BiCGSTAB() {}
205
206 /** \returns the solution x of \f$ A x = b \f$ using the current decomposition of A
207 * \a x0 as an initial solution.
208 *
209 * \sa compute()
210 */
211 template<typename Rhs,typename Guess>
212 inline const internal::solve_retval_with_guess<BiCGSTAB, Rhs, Guess>
213 solveWithGuess(const MatrixBase<Rhs>& b, const Guess& x0) const
214 {
215 eigen_assert(m_isInitialized && "BiCGSTAB is not initialized.");
216 eigen_assert(Base::rows()==b.rows()
217 && "BiCGSTAB::solve(): invalid number of rows of the right hand side matrix b");
218 return internal::solve_retval_with_guess
219 <BiCGSTAB, Rhs, Guess>(*this, b.derived(), x0);
220 }
221
222 /** \internal */
223 template<typename Rhs,typename Dest>
224 void _solveWithGuess(const Rhs& b, Dest& x) const
225 {
226 bool failed = false;
227 for(int j=0; j<b.cols(); ++j)
228 {
229 m_iterations = Base::maxIterations();
230 m_error = Base::m_tolerance;
231
232 typename Dest::ColXpr xj(x,j);
233 if(!internal::bicgstab(*mp_matrix, b.col(j), xj, Base::m_preconditioner, m_iterations, m_error))
234 failed = true;
235 }
236 m_info = failed ? NumericalIssue
237 : m_error <= Base::m_tolerance ? Success
238 : NoConvergence;
239 m_isInitialized = true;
240 }
241
242 /** \internal */
243 template<typename Rhs,typename Dest>
244 void _solve(const Rhs& b, Dest& x) const
245 {
246 // x.setZero();
247 x = b;
248 _solveWithGuess(b,x);
249 }
250
251 protected:
252
253 };
254
255
256 namespace internal {
257
258 template<typename _MatrixType, typename _Preconditioner, typename Rhs>
259 struct solve_retval<BiCGSTAB<_MatrixType, _Preconditioner>, Rhs>
260 : solve_retval_base<BiCGSTAB<_MatrixType, _Preconditioner>, Rhs>
261 {
262 typedef BiCGSTAB<_MatrixType, _Preconditioner> Dec;
263 EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs)
264
265 template<typename Dest> void evalTo(Dest& dst) const
266 {
267 dec()._solve(rhs(),dst);
268 }
269 };
270
271 } // end namespace internal
272
273 } // end namespace Eigen
274
275 #endif // EIGEN_BICGSTAB_H
276