Represents the Schur complement operator used in the solution process. More...

#include <polytropeOperator.h>

Inheritance diagram for serif::polytrope::SchurCompliment:

Public Member Functions
	SchurCompliment (const mfem::SparseMatrix &QOp, const mfem::SparseMatrix &DOp, const mfem::SparseMatrix &MOp, const mfem::Solver &GradInvOp)
	Constructs a SchurCompliment operator.

	SchurCompliment (const mfem::SparseMatrix &QOp, const mfem::SparseMatrix &DOp, const mfem::SparseMatrix &MOp)
	Constructs a SchurCompliment operator without the inverse gradient initially. The inverse gradient must be set later using updateInverseNonlinearJacobian.

	~SchurCompliment () override=default
	Destructor.

void	Mult (const mfem::Vector &x, mfem::Vector &y) const override
	Applies the Schur complement operator: y = (D - Q * GradInv * M) * x.

void	SetOperator (const mfem::SparseMatrix &QOp, const mfem::SparseMatrix &DOp, const mfem::SparseMatrix &MOp, const mfem::Solver &GradInvOp)
	Sets all operators for the Schur complement.

void	updateInverseNonlinearJacobian (const mfem::Solver &gradInv)
	Updates the inverse of the nonlinear Jacobian (GradInv).

Private Member Functions
void	updateConstantTerms (const mfem::SparseMatrix &QOp, const mfem::SparseMatrix &DOp, const mfem::SparseMatrix &MOp)
	Updates the constant matrix terms (Q, D, M).

Private Attributes
const mfem::SparseMatrix *	m_QOp = nullptr
	Pointer to the Q matrix operator.

const mfem::SparseMatrix *	m_DOp = nullptr
	Pointer to the D matrix operator.

const mfem::SparseMatrix *	m_MOp = nullptr
	Pointer to the M matrix operator.

const mfem::Solver *	m_GradInvOp = nullptr
	Pointer to the inverse of the gradient operator.

int	m_nPhi = 0
	Dimension related to the phi variable (typically number of rows/cols of D).

int	m_nTheta = 0
	Dimension related to the theta variable (typically number of rows of M).

std::unique_ptr< mfem::SparseMatrix >	m_matrixForm
	Optional: if a matrix representation is ever explicitly formed.

Detailed Description

Represents the Schur complement operator used in the solution process.

This class computes S = D - Q * GradInv * M.

Definition at line 36 of file polytropeOperator.h.

Constructor & Destructor Documentation

◆ SchurCompliment() [1/2]

serif::polytrope::SchurCompliment::SchurCompliment	(	const mfem::SparseMatrix &	QOp,
		const mfem::SparseMatrix &	DOp,
		const mfem::SparseMatrix &	MOp,
		const mfem::Solver &	GradInvOp )

Constructs a SchurCompliment operator.

Parameters

QOp	The Q matrix operator.
DOp	The D matrix operator.
MOp	The M matrix operator.
GradInvOp	The inverse of the gradient operator.

Definition at line 36 of file polytropeOperator.cpp.

◆ SchurCompliment() [2/2]

serif::polytrope::SchurCompliment::SchurCompliment	(	const mfem::SparseMatrix &	QOp,
		const mfem::SparseMatrix &	DOp,
		const mfem::SparseMatrix &	MOp )

Constructs a SchurCompliment operator without the inverse gradient initially. The inverse gradient must be set later using updateInverseNonlinearJacobian.

Parameters

QOp	The Q matrix operator.
DOp	The D matrix operator.
MOp	The M matrix operator.

Definition at line 48 of file polytropeOperator.cpp.

◆ ~SchurCompliment()

serif::polytrope::SchurCompliment::~SchurCompliment ( )

overridedefault

Destructor.

Member Function Documentation

◆ Mult()

void serif::polytrope::SchurCompliment::Mult	(	const mfem::Vector &	x,
		mfem::Vector &	y ) const

override

Applies the Schur complement operator: y = (D - Q * GradInv * M) * x.

Parameters

x	The input vector.
y	The output vector.

Definition at line 60 of file polytropeOperator.cpp.

◆ SetOperator()

void serif::polytrope::SchurCompliment::SetOperator	(	const mfem::SparseMatrix &	QOp,
		const mfem::SparseMatrix &	DOp,
		const mfem::SparseMatrix &	MOp,
		const mfem::Solver &	GradInvOp )

Sets all operators for the Schur complement.

Parameters

QOp	The Q matrix operator.
DOp	The D matrix operator.
MOp	The M matrix operator.
GradInvOp	The inverse of the gradient operator.

Definition at line 98 of file polytropeOperator.cpp.

◆ updateConstantTerms()

void serif::polytrope::SchurCompliment::updateConstantTerms	(	const mfem::SparseMatrix &	QOp,
		const mfem::SparseMatrix &	DOp,
		const mfem::SparseMatrix &	MOp )

private

Updates the constant matrix terms (Q, D, M).

Parameters

QOp	The Q matrix operator.
DOp	The D matrix operator.
MOp	The M matrix operator.

Definition at line 108 of file polytropeOperator.cpp.

◆ updateInverseNonlinearJacobian()

void serif::polytrope::SchurCompliment::updateInverseNonlinearJacobian ( const mfem::Solver & gradInv )

Updates the inverse of the nonlinear Jacobian (GradInv).

Parameters

gradInv The new inverse gradient solver.

Definition at line 103 of file polytropeOperator.cpp.

Member Data Documentation

◆ m_DOp

const mfem::SparseMatrix* serif::polytrope::SchurCompliment::m_DOp = nullptr

private

Pointer to the D matrix operator.

Definition at line 104 of file polytropeOperator.h.

◆ m_GradInvOp

const mfem::Solver* serif::polytrope::SchurCompliment::m_GradInvOp = nullptr

private

Pointer to the inverse of the gradient operator.

Definition at line 106 of file polytropeOperator.h.

◆ m_matrixForm

std::unique_ptr<mfem::SparseMatrix> serif::polytrope::SchurCompliment::m_matrixForm

mutableprivate

Optional: if a matrix representation is ever explicitly formed.

Definition at line 113 of file polytropeOperator.h.

◆ m_MOp

const mfem::SparseMatrix* serif::polytrope::SchurCompliment::m_MOp = nullptr

private

Pointer to the M matrix operator.

Definition at line 105 of file polytropeOperator.h.

◆ m_nPhi

int serif::polytrope::SchurCompliment::m_nPhi = 0

private

Dimension related to the phi variable (typically number of rows/cols of D).

Definition at line 109 of file polytropeOperator.h.

◆ m_nTheta

int serif::polytrope::SchurCompliment::m_nTheta = 0

private

Dimension related to the theta variable (typically number of rows of M).

Definition at line 110 of file polytropeOperator.h.

◆ m_QOp

const mfem::SparseMatrix* serif::polytrope::SchurCompliment::m_QOp = nullptr

private

Pointer to the Q matrix operator.

Definition at line 103 of file polytropeOperator.h.

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

src/polytrope/utils/public/polytropeOperator.h
src/polytrope/utils/private/polytropeOperator.cpp

Public Member Functions

Private Member Functions

Private Attributes

Detailed Description

Constructor & Destructor Documentation

◆ SchurCompliment() [1/2]

◆ SchurCompliment() [2/2]

◆ ~SchurCompliment()

Member Function Documentation

◆ Mult()

◆ SetOperator()

◆ updateConstantTerms()

◆ updateInverseNonlinearJacobian()

Member Data Documentation

◆ m_DOp

◆ m_GradInvOp

◆ m_matrixForm

◆ m_MOp

◆ m_nPhi

◆ m_nTheta

◆ m_QOp