Brian Silverman | 72890c2 | 2015-09-19 14:37:37 -0400 | [diff] [blame^] | 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 | // |
| 6 | // This Source Code Form is subject to the terms of the Mozilla |
| 7 | // Public License v. 2.0. If a copy of the MPL was not distributed |
| 8 | // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. |
| 9 | |
| 10 | #ifndef EIGEN_BASIC_PRECONDITIONERS_H |
| 11 | #define EIGEN_BASIC_PRECONDITIONERS_H |
| 12 | |
| 13 | namespace Eigen { |
| 14 | |
| 15 | /** \ingroup IterativeLinearSolvers_Module |
| 16 | * \brief A preconditioner based on the digonal entries |
| 17 | * |
| 18 | * This class allows to approximately solve for A.x = b problems assuming A is a diagonal matrix. |
| 19 | * In other words, this preconditioner neglects all off diagonal entries and, in Eigen's language, solves for: |
| 20 | * \code |
| 21 | * A.diagonal().asDiagonal() . x = b |
| 22 | * \endcode |
| 23 | * |
| 24 | * \tparam _Scalar the type of the scalar. |
| 25 | * |
| 26 | * This preconditioner is suitable for both selfadjoint and general problems. |
| 27 | * The diagonal entries are pre-inverted and stored into a dense vector. |
| 28 | * |
| 29 | * \note A variant that has yet to be implemented would attempt to preserve the norm of each column. |
| 30 | * |
| 31 | */ |
| 32 | template <typename _Scalar> |
| 33 | class DiagonalPreconditioner |
| 34 | { |
| 35 | typedef _Scalar Scalar; |
| 36 | typedef Matrix<Scalar,Dynamic,1> Vector; |
| 37 | typedef typename Vector::Index Index; |
| 38 | |
| 39 | public: |
| 40 | // this typedef is only to export the scalar type and compile-time dimensions to solve_retval |
| 41 | typedef Matrix<Scalar,Dynamic,Dynamic> MatrixType; |
| 42 | |
| 43 | DiagonalPreconditioner() : m_isInitialized(false) {} |
| 44 | |
| 45 | template<typename MatType> |
| 46 | DiagonalPreconditioner(const MatType& mat) : m_invdiag(mat.cols()) |
| 47 | { |
| 48 | compute(mat); |
| 49 | } |
| 50 | |
| 51 | Index rows() const { return m_invdiag.size(); } |
| 52 | Index cols() const { return m_invdiag.size(); } |
| 53 | |
| 54 | template<typename MatType> |
| 55 | DiagonalPreconditioner& analyzePattern(const MatType& ) |
| 56 | { |
| 57 | return *this; |
| 58 | } |
| 59 | |
| 60 | template<typename MatType> |
| 61 | DiagonalPreconditioner& factorize(const MatType& mat) |
| 62 | { |
| 63 | m_invdiag.resize(mat.cols()); |
| 64 | for(int j=0; j<mat.outerSize(); ++j) |
| 65 | { |
| 66 | typename MatType::InnerIterator it(mat,j); |
| 67 | while(it && it.index()!=j) ++it; |
| 68 | if(it && it.index()==j) |
| 69 | m_invdiag(j) = Scalar(1)/it.value(); |
| 70 | else |
| 71 | m_invdiag(j) = 0; |
| 72 | } |
| 73 | m_isInitialized = true; |
| 74 | return *this; |
| 75 | } |
| 76 | |
| 77 | template<typename MatType> |
| 78 | DiagonalPreconditioner& compute(const MatType& mat) |
| 79 | { |
| 80 | return factorize(mat); |
| 81 | } |
| 82 | |
| 83 | template<typename Rhs, typename Dest> |
| 84 | void _solve(const Rhs& b, Dest& x) const |
| 85 | { |
| 86 | x = m_invdiag.array() * b.array() ; |
| 87 | } |
| 88 | |
| 89 | template<typename Rhs> inline const internal::solve_retval<DiagonalPreconditioner, Rhs> |
| 90 | solve(const MatrixBase<Rhs>& b) const |
| 91 | { |
| 92 | eigen_assert(m_isInitialized && "DiagonalPreconditioner is not initialized."); |
| 93 | eigen_assert(m_invdiag.size()==b.rows() |
| 94 | && "DiagonalPreconditioner::solve(): invalid number of rows of the right hand side matrix b"); |
| 95 | return internal::solve_retval<DiagonalPreconditioner, Rhs>(*this, b.derived()); |
| 96 | } |
| 97 | |
| 98 | protected: |
| 99 | Vector m_invdiag; |
| 100 | bool m_isInitialized; |
| 101 | }; |
| 102 | |
| 103 | namespace internal { |
| 104 | |
| 105 | template<typename _MatrixType, typename Rhs> |
| 106 | struct solve_retval<DiagonalPreconditioner<_MatrixType>, Rhs> |
| 107 | : solve_retval_base<DiagonalPreconditioner<_MatrixType>, Rhs> |
| 108 | { |
| 109 | typedef DiagonalPreconditioner<_MatrixType> Dec; |
| 110 | EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs) |
| 111 | |
| 112 | template<typename Dest> void evalTo(Dest& dst) const |
| 113 | { |
| 114 | dec()._solve(rhs(),dst); |
| 115 | } |
| 116 | }; |
| 117 | |
| 118 | } |
| 119 | |
| 120 | /** \ingroup IterativeLinearSolvers_Module |
| 121 | * \brief A naive preconditioner which approximates any matrix as the identity matrix |
| 122 | * |
| 123 | * \sa class DiagonalPreconditioner |
| 124 | */ |
| 125 | class IdentityPreconditioner |
| 126 | { |
| 127 | public: |
| 128 | |
| 129 | IdentityPreconditioner() {} |
| 130 | |
| 131 | template<typename MatrixType> |
| 132 | IdentityPreconditioner(const MatrixType& ) {} |
| 133 | |
| 134 | template<typename MatrixType> |
| 135 | IdentityPreconditioner& analyzePattern(const MatrixType& ) { return *this; } |
| 136 | |
| 137 | template<typename MatrixType> |
| 138 | IdentityPreconditioner& factorize(const MatrixType& ) { return *this; } |
| 139 | |
| 140 | template<typename MatrixType> |
| 141 | IdentityPreconditioner& compute(const MatrixType& ) { return *this; } |
| 142 | |
| 143 | template<typename Rhs> |
| 144 | inline const Rhs& solve(const Rhs& b) const { return b; } |
| 145 | }; |
| 146 | |
| 147 | } // end namespace Eigen |
| 148 | |
| 149 | #endif // EIGEN_BASIC_PRECONDITIONERS_H |