tngmath::SparseArpack Class Reference

Eigen analyses solver for sparse matrices based on ARPACK. More...

#include <sparsearpack.hpp>

Collaboration diagram for tngmath::SparseArpack:

[legend]

List of all members.

Public Types
typedef Matrix	MatrixType
	the dense matrix/vector type for communication
typedef SparseMatrix	SparseMatrixType
	the base type of sparse input
Public Member Functions
const MatrixType &	Eigenvalues () const
	returns the eigenvalues after solving them
const MatrixType &	Eigenvectors () const
	returns the eigenvectors after solving them
bool	Largest (const SparseMatrixType &A, const SparseMatrixType &B, SparseSolver &solver, const unsigned int number)
bool	Largest (const SparseMatrixType &A, const unsigned int number)
const unsigned int	MaxIterations () const
	returns m_maxIter
const double	Precision () const
	returns the current precision
void	SetMaxIterations (unsigned int it)
	sets the max. number of allowed iterations in ARPACK
void	SetPrecision (double prec)
	sets the precision of ARPACK
bool	ShiftInvert (const SparseMatrixType &A, const SparseMatrixType &B, SparseSolver &solver, const double &shift, const unsigned int number)
bool	ShiftInvert (const SparseMatrixType &A, SparseSolver &solver, const double &shift, const unsigned int number)
	SparseArpack ()
	default constructor
virtual	~SparseArpack ()
	destructor
Protected Member Functions
MatrixType &	Eigenvalues ()
MatrixType &	Eigenvectors ()
Protected Attributes
MatrixType	m_eigenvalues
	stores the eigenvalues
MatrixType	m_eigenvectors
	stores the eigenvectors
unsigned int	m_maxIter
	stores the number of iterations for ARPACK
double	m_precision
	stores the set precision for ARPACK

---

Detailed Description

Eigen analyses solver for sparse matrices based on ARPACK.

---

Member Function Documentation

bool tngmath::SparseArpack::Largest	(	const SparseMatrixType &	A,
		const SparseMatrixType &	B,
		SparseSolver &	solver,
		const unsigned int	number
	)

given a generalized symmetric eigen problem, this method computes the largest eigenvalues using a shift-inverse transformation.

Parameters:

	A	the input matrix A (stiffness)
	B	the input matrix B (mass)
	solver	the solver object used for factorizing B
	number	the number of eigenvalues to be computed

Returns:

true if successful.

bool tngmath::SparseArpack::Largest	(	const SparseMatrixType &	A,
		const unsigned int	number
	)

given a standard symmetric eigen problem, this method computes the largest eigenvalues using a shift-inverse transformation.

Parameters:

	A	the input matrix
	number	the number of eigenvalues to be computed

Returns:

true if successful.

bool tngmath::SparseArpack::ShiftInvert	(	const SparseMatrixType &	A,
		const SparseMatrixType &	B,
		SparseSolver &	solver,
		const double &	shift,
		const unsigned int	number
	)

given a generalized symmetric eigen problem, this method computes the smallest eigenvalues using a shift-inverse transformation.

Parameters:

	A	the input matrix A (stiffness)
	B	the input matrix B (mass)
	solver	the solver object used for factorizing A
	shift	the shifting parameter (lower bound)
	number	the number of eigenvalues to be computed

Returns:

true if successful.

bool tngmath::SparseArpack::ShiftInvert	(	const SparseMatrixType &	A,
		SparseSolver &	solver,
		const double &	shift,
		const unsigned int	number
	)

given a standard symmetric eigen problem, this method computes the smallest eigenvalues using a shift-inverse transformation.

Parameters:

	A	the input matrix
	solver	the solver object used for factorizing A
	shift	the shifting parameter (lower bound)
	number	the number of eigenvalues to be computed

Returns:

true if successful.

---

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

• modules/tngmath/tngmath/sparsearpack.hpp