Provide the ApplyInverse method for a Epetra_CrsMatrix object. More...

#include <ApproximatedInvertibleRowMatrix.hpp>

Inheritance diagram for ApproximatedInvertibleRowMatrix:

Collaboration diagram for ApproximatedInvertibleRowMatrix:

Public Member Functions
	ApproximatedInvertibleRowMatrix ()

virtual	~ApproximatedInvertibleRowMatrix ()

int	Compute ()

std::shared_ptr< RowMatrixPreconditioner > &	Preconditioner ()

Public Member Functions inherited from LinearOperatorAlgebra
virtual	~LinearOperatorAlgebra ()
	Destructor. More...

int	apply (const VectorEpetra &X, VectorEpetra &Y) const
	Returns the result of a LinearOperator applied to a VectorEpetra X in Y. More...

int	applyInverse (const VectorEpetra &X, VectorEpetra &Y)
	Returns the result of a LinearOperator inverse applied to an VectorEpetra X in Y. More...

Private Attributes
bool	usePreconditionerAsApproximatedInverse

rowMatrixPtr_Type	M_rowMatrix

std::shared_ptr< RowMatrixPreconditioner >	M_prec

std::shared_ptr< InvertibleOperator >	M_linSolver

pList_Type	M_pList

Public Typedef
typedef Epetra_CrsMatrix	rowMatrix_Type

typedef std::shared_ptr< rowMatrix_Type >	rowMatrixPtr_Type

typedef Teuchos::ParameterList	pList_Type

Attribute set methods
void	SetRowMatrix (const rowMatrixPtr_Type &rowMatrix)

void	SetParameterList (const pList_Type pList)

virtual int	SetUseTranspose (bool UseTranspose)
	not fully supported! More...

Mathematical functions
virtual int	Apply (const vector_Type &X, vector_Type &Y) const
	Returns the result of a raw_operator applied to a raw_vector X in Y. More...

virtual int	ApplyInverse (const vector_Type &X, vector_Type &Y) const
	Returns the result of a raw_operator inverse applied to an raw_vector X in Y. More...

virtual double	NormInf () const
	Returns the infinity norm of the global matrix. More...

Attribute access functions
virtual const char *	Label () const
	Returns a character string describing the operator. More...

virtual bool	UseTranspose () const
	Returns the current UseTranspose setting. More...

virtual bool	HasNormInf () const
	Returns true if the this object can provide an approximate Inf-norm, false otherwise. More...

virtual const comm_Type &	Comm () const
	Returns a pointer to the Epetra_Comm communicator associated with this operator. More...

virtual const map_Type &	OperatorDomainMap () const
	Returns the raw_map object associated with the domain of this operator. More...

virtual const map_Type &	OperatorRangeMap () const
	Returns the raw_map object associated with the range of this operator. More...

Additional Inherited Members
Public Types inherited from LinearOperatorAlgebra
typedef Epetra_Comm	comm_Type

typedef std::shared_ptr< comm_Type >	commPtr_Type

typedef Epetra_Map	map_Type

typedef std::shared_ptr< map_Type >	mapPtr_Type

typedef std::shared_ptr< const map_Type >	constMapPtr_Type

typedef Epetra_Operator	operator_Type

typedef std::shared_ptr< operator_Type >	operatorPtr_Type

typedef Epetra_MultiVector	vector_Type

typedef std::shared_ptr< vector_Type >	vectorPtr_Type

Detailed Description

Provide the ApplyInverse method for a Epetra_CrsMatrix object.

ApproximatedInvertibleRowMatrix is derived by the LinearOperator class, and it can be used to approximate the inverse of a Epetra_CsrMatrix, either by using the Krylov methods in Belos or AztecOO or the preconditioning techniques in ML or Ifpack. Internally ApproximatedInvertibleRowMatrix contains pointers to a Epetra_CsrMatrix, to an InvertibleOperator object (that provides the interface to AztecOO or Belos), and RowMatrixPreconditioner (the provides the interface to Ifpack or ML).

The parameters to set up the Krylov methods or the preconditioner are given by using a Teuchos Parameter List. Below how the parameter list would look like:

*         <ParameterList name="ApproximatedInvertibleRowMatrix">
*           <Parameter name="use preconditioner as approximated inverse" type="bool" value="false"/>
*           <Parameter name="preconditioner type" type="string" value="ML"/> <!-- Ifpack, ML, TwoLevel-->
*           <ParameterList name="solver">
*              <Parameter name="Linear Solver Type" type="string" value="Belos"/>
*              <ParameterList name="AztecOO">
*                 <Parameter name="conv" type="string" value="r0"/>
*                 <Parameter name="max_iter" type="int" value="1500"/>
*                 <Parameter name="output" type="string" value="warnings"/>
*                 <Parameter name="scaling" type="string" value="none"/>
*                 <Parameter name="solver" type="string" value="cg"/>
*                 <Parameter name="tol" type="double" value="1e-03"/>
*              </ParameterList>
*              <ParameterList name="Belos">
*                  <Parameter name="Solver Type" type="string" value="RCG"/>
*                  <Parameter name="Preconditioner Side" type="string" value="Left"/>
*                  <ParameterList name="options">
*                      <Parameter name="Maximum Iterations" type="int" value="50"/>
*                      <Parameter name="Num Blocks" type="int" value="50"/>
*                      <Parameter name="Num Recycled Blocks" type="int" value="10"/>
*                      <Parameter name="Convergence Tolerance" type="double" value="1e-03"/>
*                      <Parameter name="Output Frequency" type="int" value="1"/>
*                      <Parameter name="Verbosity" type="int" value="0"/>
*                  </ParameterList>
*              </ParameterList>
*           </ParameterList>
*           <ParameterList name="preconditioner">
*               <ParameterList name="ML">
*                   <ParameterList name="options">
*                       <Parameter name="default values" type="string" value="SA"/>
*                   </ParameterList>
*               </ParameterList>
*           </ParameterList>
*       </ParameterList>
*

Definition at line 87 of file ApproximatedInvertibleRowMatrix.hpp.