UserMatrix Class Reference

Abstract class for a user defined quadratic matrix. More...

#include <usermatrix.h>

Inheritance diagram for UserMatrix:

Public Member Functions
	UserMatrix (int n=0)
	(Standard-)Constructor for the dimension.
virtual	~UserMatrix ()
	Virtual Destructor.
int	dim () const
	Gives the dimension.
virtual void	MultV (UserVector< double > &y, const UserVector< double > &x) const =0
	Multiplies with a UserVector<double>.
virtual void	MultV (double y, const double x) const
	Multiplies the matrix with a double*.
virtual double	yAx (const UserVector< double > &y, const UserVector< double > &x) const
	Multiplies from left and from right with UserVector<double>'s.
virtual double	xAx (const UserVector< double > &x) const
	Multiplies the quadratic product of this matrix and a UserVector<double>.
virtual void	AddMult (UserVector< double > &y, const UserVector< double > &x, const double val) const
	Adds to a UserVector<double> the product of a double value and this matrix, multiplied with a UserVector<double>.
dvector	operator* (const dvector &x) const
	Multiplies this matrix with a dvector.
double *	operator* (const double *x) const
	Multiplies this matrix with a double*.
Pointer< UserVector< double > >	operator* (const Pointer< UserVector< double > > x) const
virtual int	eig_ql (vector< dvector > &eig_vec, vector< double > &eig_val) const
	Calculates the smallest eigenvalues and eigenvectors with the QL-algorithm.
int	eig (vector< dvector > &eig_vec, vector< double > &eig_val, Param *param=NULL)
	Calculates the smallest eigenvalues of this matrix.
int	eig (dvector &eig_vec, double &eig_val, Param *param=NULL)
	Calculates the smallest eigenvalue of this matrix.
virtual void	print (ostream &out) const
	Prints out this matrix.
Protected Attributes
int	dim_
	The dimension of this matrix.
Friends
ostream &	operator<< (ostream &out, const UserMatrix &a)
	Prints out information about the matrix.

Detailed Description

Abstract class for a user defined quadratic matrix.

Following methods are abstract:

MultV(dvector&, dvector&)
solve(dvector&, dvector&)

Definition at line 35 of file usermatrix.h.

Constructor & Destructor Documentation

UserMatrix::UserMatrix ( int n = 0 ) [inline]

(Standard-)Constructor for the dimension.

Parameters:

n

The dimension, default is 0.

Definition at line 59 of file usermatrix.h.

virtual UserMatrix::~UserMatrix ( ) [inline, virtual]

Virtual Destructor.

Definition at line 65 of file usermatrix.h.

Member Function Documentation

int UserMatrix::dim ( ) const [inline]

Gives the dimension.

Returns:: The dimension as int.

Definition at line 70 of file usermatrix.h.

virtual void UserMatrix::MultV	(	UserVector< double > &	y,
		const UserVector< double > &	x
	)			const `[pure virtual]`

Multiplies with a UserVector<double>.

Abstract.

Parameters:

	y	The UserVector<double> to store the result in: y = this x.
	x	The UserVector<double> to multiply with.

See also:: MultV(double*, double*)

Implemented in QqpMatrix, QqpExtMatrix, HessMatrix, ShiftMatrix, MinusMatrix, SumMatrix, DiagMatrix, BlockMatrix, DenseMatrix, and SparseMatrix2.

virtual void UserMatrix::MultV	(	double *	y,
		const double *	x
	)			const `[inline, virtual]`

Multiplies the matrix with a double*.

Parameters:

	y	The double* to store the result in.
	x	The double* to multiply with this matrix.

See also:: MultV(dvector&, const dvector&)

Reimplemented in ShiftMatrix, MinusMatrix, DenseMatrix, and SparseMatrix2.

Definition at line 85 of file usermatrix.h.

virtual double UserMatrix::yAx	(	const UserVector< double > &	y,
		const UserVector< double > &	x
	)			const `[inline, virtual]`

Multiplies from left and from right with UserVector<double>'s.

Parameters:

	y	The UserVector<double> to multiply with from left.
	x	The UserVector<double> to multiply with from right.

Returns:: The double value y*this*x.

See also:: xAx(const UserVector<double>&)

Reimplemented in ShiftMatrix, MinusMatrix, DiagMatrix, BlockMatrix, and SparseMatrix2.

Definition at line 97 of file usermatrix.h.

virtual double UserMatrix::xAx ( const UserVector< double > & x ) const [inline, virtual]

Multiplies the quadratic product of this matrix and a UserVector<double>.

Parameters:

x

The UserVector<double> to multiply with.

Returns:: The double yAx(x,x).

See also:: yAx(const UserVector<double>&, const UserVector<double>&)

Reimplemented in SparseMatrix2.

Definition at line 108 of file usermatrix.h.

virtual void UserMatrix::AddMult	(	UserVector< double > &	y,
		const UserVector< double > &	x,
		const double	val
	)			const `[inline, virtual]`

Adds to a UserVector<double> the product of a double value and this matrix, multiplied with a UserVector<double>.

Parameters:

	y	The UserVector<double> to store the result in: y + val * A * x
	x	The UserVector<double> to multiply this matrix with.
	val	The double to multiply with.

Reimplemented in SparseMatrix2.

Definition at line 116 of file usermatrix.h.

dvector UserMatrix::operator* ( const dvector & x ) const [inline]

Multiplies this matrix with a dvector.

Parameters:

x

The dvector to multiply with.

Returns:: The result (this*x).

See also:: MultV(dvector&, const dvector&)

Definition at line 125 of file usermatrix.h.

double* UserMatrix::operator* ( const double * x ) const [inline]

Multiplies this matrix with a double*.

Parameters:

x

The double* to multiply with.

Returns:: The result as a new(!!) double*.

See also:: MultV(double*, double*)

Definition at line 136 of file usermatrix.h.

Pointer<UserVector<double> > UserMatrix::operator* ( const Pointer< UserVector< double > > x ) const [inline]

Definition at line 142 of file usermatrix.h.

virtual int UserMatrix::eig_ql	(	vector< dvector > &	eig_vec,
		vector< double > &	eig_val
	)			const `[virtual]`

Calculates the smallest eigenvalues and eigenvectors with the QL-algorithm.

Converts the matrix to a DenseMatrix and calls eig_ql from there.

Parameters:

	eig_vec	A vector of dvectors to store the eigenvectors in.
	eig_val	A vector of doubles to store the eigenvalues in.
	param	Some parameters, default is NULL.

Returns:: The return value from imtql2, which is 0, if all went good or the number of the eigenvalue, which wasn't founds after 30 iterations.

See also:: DenseMatrix::eig_ql(vector<dvector>&, vector<double>&, Param*)

Reimplemented in DenseMatrix.

int UserMatrix::eig	(	vector< dvector > &	eig_vec,
		vector< double > &	eig_val,
		Param *	param = `NULL`
	)

Calculates the smallest eigenvalues of this matrix.

If dim>=50, calls the lanczos-algorithm. Otherwise a DenseMatrix is created and the ql-algorithm called.

Parameters:

	eig_vec	The dvectors to store the eigenvectors in.
	eig_val	The doubles to store the eigenvalues in.
	param	Parameters for Lanczos or ARPACK, default is NULL.

Returns:: Depends on the choosen eigenvalue algorithm. It's 0, if all went good.

See also:: eig_ql(vector<dvector>&, vector<double>&); eig_lanczos(vector<dvector>&, vector<double>&, Param*); eig_arpack(vector<dvector>&, vector<double>&, Param*)

int UserMatrix::eig	(	dvector &	eig_vec,
		double &	eig_val,
		Param *	param = `NULL`
	)

Calculates the smallest eigenvalue of this matrix.

Parameters:

	eig_vec	The dvector to store the smallest eigenvector in.
	eig_val	The double to store the smallest eigenvalue in.
	param	Parameters for the lanczos-algorithm, default is NULL.

Returns:: Depends on the choosen eigenvalue algorithm. It's 0, if all went good.

See also:: eig(vector<dvector>&, vector<double>&, Param*)

virtual void UserMatrix::print ( ostream & out ) const [virtual]

Prints out this matrix.

Generates a DenseMatrix with the values of this matrix and prints this one.

Parameters:

out

The ostream to print to.

See also:: DenseMatrix::print(ostream&)

Reimplemented in HessMatrix, ShiftMatrix, MinusMatrix, SumMatrix, DiagMatrix, BlockMatrix, DenseMatrix, and SparseMatrix2.

Friends And Related Function Documentation

ostream& operator<<	(	ostream &	out,
		const UserMatrix &	a
	)			`[friend]`

Prints out information about the matrix.

Calls a.print(out).

Parameters:

	out	The ostream to print to.
	a	The UserMatrix to get informations about.

Returns:: The ostream.

See also:: print(ostream&)

Definition at line 47 of file usermatrix.h.

Member Data Documentation

int UserMatrix::dim_ [protected]

The dimension of this matrix.

Definition at line 48 of file usermatrix.h.

The documentation for this class was generated from the following file:

usermatrix.h

UserMatrix Class Reference

Public Member Functions

Protected Attributes

Friends

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Friends And Related Function Documentation

Member Data Documentation