oofem::LDLTFactorization Class Reference

#include <ldltfact.h>

Inheritance diagram for oofem::LDLTFactorization:

Collaboration diagram for oofem::LDLTFactorization:

Public Member Functions
	LDLTFactorization (Domain *d, EngngModel *m)
	Constructor - creates new instance of LDLTFactorization, with number i, belonging to domain d and Engngmodel m.
virtual	~LDLTFactorization ()
	Destructor.
ConvergedReason	solve (SparseMtrx &A, FloatArray &b, FloatArray &x) override
const char *	giveClassName () const override
LinSystSolverType	giveLinSystSolverType () const override
SparseMtrxType	giveRecommendedMatrix (bool symmetric) const override
Public Member Functions inherited from oofem::SparseLinearSystemNM
	SparseLinearSystemNM (Domain *d, EngngModel *m)
	Constructor.
virtual	~SparseLinearSystemNM ()
	Destructor.
std::string	errorInfo (const char *func)
virtual ConvergedReason	solve (SparseMtrx &A, FloatMatrix &B, FloatMatrix &X)
Public Member Functions inherited from oofem::NumericalMethod
	NumericalMethod (Domain *d, EngngModel *m)
virtual	~NumericalMethod ()
	Destructor.
EngngModel *	giveEngngModel ()
virtual void	initializeFrom (InputRecord &ir)
virtual void	reinitialize ()
virtual void	setDomain (Domain *d)
virtual void	saveContext (DataStream &stream, ContextMode mode)
virtual void	restoreContext (DataStream &stream, ContextMode mode)

Additional Inherited Members
Protected Attributes inherited from oofem::NumericalMethod
Domain *	domain
	Pointer to domain.
EngngModel *	engngModel
	Pointer to engineering model.

Detailed Description

Implements the solution of linear system of equation in the form Ax=b using direct factorization method. Can work with any sparse matrix implementation. However, the sparse matrix implementation have to support its factorization (canBeFactorized method).

Definition at line 55 of file ldltfact.h.

Constructor & Destructor Documentation

LDLTFactorization()

oofem::LDLTFactorization::LDLTFactorization	(	Domain *	d,
		EngngModel *	m )

Constructor - creates new instance of LDLTFactorization, with number i, belonging to domain d and Engngmodel m.

References LDLTFactorization(), and solve().

Referenced by LDLTFactorization().

~LDLTFactorization()

oofem::LDLTFactorization::~LDLTFactorization ( )

virtual

Destructor.

Definition at line 46 of file ldltfact.C.

Member Function Documentation

giveClassName()

const char * oofem::LDLTFactorization::giveClassName ( ) const

inlineoverridevirtual

Implements oofem::SparseLinearSystemNM.

Definition at line 74 of file ldltfact.h.

giveLinSystSolverType()

LinSystSolverType oofem::LDLTFactorization::giveLinSystSolverType ( ) const

inlineoverridevirtual

Returns

LinSystSolverType value, corresponding to receiver.

Implements oofem::SparseLinearSystemNM.

Definition at line 75 of file ldltfact.h.

References oofem::ST_Direct.

giveRecommendedMatrix()

SparseMtrxType oofem::LDLTFactorization::giveRecommendedMatrix ( bool symmetric ) const

inlineoverridevirtual

Returns the recommended sparse matrix type for this solver.

Implements oofem::SparseLinearSystemNM.

Definition at line 76 of file ldltfact.h.

References oofem::SMT_Skyline, and oofem::SMT_SkylineU.

solve()

ConvergedReason oofem::LDLTFactorization::solve ( SparseMtrx & A, FloatArray & b, FloatArray & x )

overridevirtual

Solves the given linear system by LDL^T factorization. Implementation rely on factorization support provided by mapped sparse matrix. It calls Lhs->factorized()->backSubstitutionWith(*solutionArray). Sets solved flag to 1 if o.k.

Parameters

A	coefficient matrix
b	right hand side
x	solution array

Returns

NM_Status value

Implements oofem::SparseLinearSystemNM.

Definition at line 51 of file ldltfact.C.

References oofem::SparseMtrx::backSubstitutionWith(), oofem::SparseMtrx::canBeFactorized(), oofem::CR_CONVERGED, oofem::SparseMtrx::factorized(), and OOFEM_ERROR.

Referenced by LDLTFactorization().

---

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

• ldltfact.h
• ldltfact.C