MathTL
 All Classes Functions Variables Typedefs Enumerations
Public Types | Public Member Functions | Protected Attributes
MathTL::QuasiStationaryMatrix< C > Class Template Reference

#include <qs_matrix.h>

List of all members.

Public Types

typedef Vector< C >::size_type size_type

Public Member Functions

 QuasiStationaryMatrix ()
 default constructor, yields (2^j)x(2^j) zero matrix
 QuasiStationaryMatrix (const int j0, const size_type mj0, const size_type nj0, const Matrix< C > &ML, const Matrix< C > &MR, const Vector< C > &bandL, const Vector< C > &bandR, const size_type offsetL, const size_type offsetR, const double factor=1.0)
QuasiStationaryMatrix< C > & operator= (const QuasiStationaryMatrix< C > &M)
 assignment
const size_type row_dimension () const
const size_type column_dimension () const
const size_type ML_column_dimension () const
const size_type MR_column_dimension () const
void set_level (const int j) const
const C get_entry (const size_type row, const size_type column) const
void to_sparse (SparseMatrix< C > &S) const
template<class VECTOR >
void apply (const VECTOR &x, VECTOR &Mx, const size_type x_offset=0, const size_type Mx_offset=0, const bool add_to=false) const
void apply (const std::map< size_type, C > &x, std::map< size_type, C > &Mx, const size_type x_offset=0, const size_type Mx_offset=0, const bool add_to=false) const
template<class VECTOR >
void apply_transposed (const VECTOR &x, VECTOR &Mtx, const size_type x_offset=0, const size_type Mtx_offset=0, const bool add_to=false) const
void apply_transposed (const std::map< size_type, C > &x, std::map< size_type, C > &Mtx, const size_type x_offset=0, const size_type Mtx_offset=0, const bool add_to=false) const
void apply_central_block (const std::map< size_type, C > &x, std::map< size_type, C > &Mx, const size_type x_offset=0, const size_type Mx_offset=0, const bool add_to=false) const
void apply_central_block_transposed (const std::map< size_type, C > &x, std::map< size_type, C > &Mtx, const size_type x_offset=0, const size_type Mtx_offset=0, const bool add_to=false) const
void apply_central_columns (const std::map< size_type, C > &x, std::map< size_type, C > &Mx, const size_type x_offset=0, const size_type Mx_offset=0, const bool add_to=false) const
void apply_central_columns_transposed (const std::map< size_type, C > &x, std::map< size_type, C > &Mtx, const size_type x_offset=0, const size_type Mtx_offset=0, const bool add_to=false) const
void print (std::ostream &os, const unsigned int tabwidth=8, const unsigned int precision=3) const

Protected Attributes

int j0_
int j_
size_type mj0_
size_type nj0_
Matrix< C > ML_
Matrix< C > MR_
Vector< C > bandL_
Vector< C > bandR_
size_type offsetL_
size_type offsetR_
double factor_

Detailed Description

template<class C>
class MathTL::QuasiStationaryMatrix< C >

This class models quasi-stationary matrices like M_{j,0}, M_{j,1} or their dual counterparts. A quasi-stationary matrix is given by an upper left and a lower right corner block, and two bands for the left and right half, respectively. The bands are given by two stencil column vectors. For an odd number of columns, by convention, the central band stems from the _right_ half.

Member Typedef Documentation

template<class C>
typedef Vector<C>::size_type MathTL::QuasiStationaryMatrix< C >::size_type

type of indexes and size type (cf. STL containers)

Constructor & Destructor Documentation

template<class C>
MathTL::QuasiStationaryMatrix< C >::QuasiStationaryMatrix ( const int  j0,
const size_type  mj0,
const size_type  nj0,
const Matrix< C > &  ML,
const Matrix< C > &  MR,
const Vector< C > &  bandL,
const Vector< C > &  bandR,
const size_type  offsetL,
const size_type  offsetR,
const double  factor = 1.0 
)

constructor from the matrix ingredients j0 : coarsest level mj0,nj0 : number of rows and columns on coarsest level ML,MR : upper left and lower right corner block (may also be empty) bandL,bandR : left and right band offsetL,offsetR: left and right row offset of the bands factor : constant pre-factor

Member Function Documentation

template<class C >
template<class VECTOR >
void MathTL::QuasiStationaryMatrix< C >::apply ( const VECTOR &  x,
VECTOR &  Mx,
const size_type  x_offset = 0,
const size_type  Mx_offset = 0,
const bool  add_to = false 
) const

matrix-vector multiplication Mx = (*this) * x; it is possible to specify an offset which parts of the vectors shall be used, thus enabling in-place algorithms, the flag "add_to" toggles whether the result is added to Mx (true) or Mx is overwritten (false)

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mx,
const size_type  x_offset = 0,
const size_type  Mx_offset = 0,
const bool  add_to = false 
) const

matrix-vector multiplication Mx = (*this) * x, where the vector is modeled by std::map

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply_central_block ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mx,
const size_type  x_offset = 0,
const size_type  Mx_offset = 0,
const bool  add_to = false 
) const

apply the central block of the matrix (i.e. without first/last row/column) to a vector

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply_central_block_transposed ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mtx,
const size_type  x_offset = 0,
const size_type  Mtx_offset = 0,
const bool  add_to = false 
) const

apply the transpose of the central block of the matrix (i.e. without first/last row/column) to a vector

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply_central_columns ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mx,
const size_type  x_offset = 0,
const size_type  Mx_offset = 0,
const bool  add_to = false 
) const

apply the central columns of the matrix (i.e. without first/last column) to a vector

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply_central_columns_transposed ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mtx,
const size_type  x_offset = 0,
const size_type  Mtx_offset = 0,
const bool  add_to = false 
) const

apply the transpose of the central columns of the matrix (i.e. without first/last column) to a vector

template<class C >
template<class VECTOR >
void MathTL::QuasiStationaryMatrix< C >::apply_transposed ( const VECTOR &  x,
VECTOR &  Mtx,
const size_type  x_offset = 0,
const size_type  Mtx_offset = 0,
const bool  add_to = false 
) const

transposed matrix-vector multiplication Mtx = (*this)^T * x; again potentially with offsets for both input and output vector and with an "add_to" flag

template<class C>
void MathTL::QuasiStationaryMatrix< C >::apply_transposed ( const std::map< size_type, C > &  x,
std::map< size_type, C > &  Mtx,
const size_type  x_offset = 0,
const size_type  Mtx_offset = 0,
const bool  add_to = false 
) const

transposed matrix-vector multiplication Mtx = (*this)^T * x; where the vector is modeled by std::map

template<class C >
const QuasiStationaryMatrix< C >::size_type MathTL::QuasiStationaryMatrix< C >::column_dimension ( ) const [inline]

column dimension

template<class C >
const C MathTL::QuasiStationaryMatrix< C >::get_entry ( const size_type  row,
const size_type  column 
) const

read-only access to a single matrix entry

template<class C>
const size_type MathTL::QuasiStationaryMatrix< C >::ML_column_dimension ( ) const [inline]

column dimension of upper left block

template<class C>
const size_type MathTL::QuasiStationaryMatrix< C >::MR_column_dimension ( ) const [inline]

column dimension of lower right block

template<class C >
void MathTL::QuasiStationaryMatrix< C >::print ( std::ostream &  os,
const unsigned int  tabwidth = 8,
const unsigned int  precision = 3 
) const

stream output with user-defined tabwidth and precision (cf. deal.II)

template<class C >
const QuasiStationaryMatrix< C >::size_type MathTL::QuasiStationaryMatrix< C >::row_dimension ( ) const [inline]

row dimension

template<class C >
void MathTL::QuasiStationaryMatrix< C >::set_level ( const int  j) const

set level j

template<class C>
void MathTL::QuasiStationaryMatrix< C >::to_sparse ( SparseMatrix< C > &  S) const

construct a sparse matrix

The documentation for this class was generated from the following files:
 All Classes Functions Variables Typedefs Enumerations
Generated on Tue Dec 3 2013 10:52:38 for MathTL by   doxygen 1.7.6.1