Solver Class Reference

Generic solver algorithms that store the instance of the Model class and the common variables for all solver types,. More...

#include <Solver.h>

Inheritance diagram for Solver:

Collaboration diagram for Solver:

List of all members.

Public Member Functions
	Solver ()
	Default constructor (without parameters): the class attributes are initialized with default values.
	~Solver ()
	Destructor.
	Solver (Solver &Instance)
	Copy-initializer constructor. Copies the contents of Instance to the current object,.
Solver &	operator= (Solver &Instance)
	Copies the contents of Instance to the current object,.
void	SetSolverAttributes (SolverParam_S &SolverParameters)
	Sets the class attributes to be stored in the SolverParam_S instance.
void	SetSolverAttributes (SolverType_E SolvType, MassStiffType_E MassType, StiffnessType_E StiffType, ContactModel_E ContactMod, ContactSolver_E ContactSol, LinearSystemSolutionMethod_E LSMethod, unsigned long LSMaxIt, double LSPrec, SolutionAlgorithm_E SoluAlgorithm, unsigned long SolMaxIt, double SolPrec, unsigned long NumLoadSteps, AnalysisType_E Atype, OperatorType_E OpType, double to, double tf, unsigned long NTS, unsigned long FrictionCoef, double PenaltyParameter, unsigned long ContactInteOrder, TimeIntegrationMeth_E TimeIntMeth, unsigned long OutFreqNR, unsigned long OutFreqTime)
	Sets the class attributes as input parameters to be stored in the SolverParam_S instance.
void	SetTheoreticalSolutionParameters (TheoreticalSolution_S &TheoSol)
	Sets the pameters for theoretical solution.
LinearSystemSolutionMethod_E	GetLSMethod ()
	Returns the type of algorithm used for linear system solution.
ContactSolver_E	GetContactIntegrationMethod ()
	Returns the type of contact integration method.
ContactModel_E	GetContactFrictionCondition ()
	Returns the type of tangential contact condition (friction or frictionless)
unsigned long	GetFrictionCoef ()
	Returns the type of tangential contact condition (friction or frictionless)
SolverType_E	GetSolverType ()
	Returns the solver type.
OperatorType_E	GetOperatorType ()
	Returns the operator type.
unsigned long	GetIterativeSolutionMaxIterations ()
	Returns the maximum number of iterations for the iterative solution of linear systems.
double	GetIterativeSolutionPrecision ()
	Returns the precision required in the solution of linear systems using iterative algorithms.
SolutionAlgorithm_E	GetSolutionAlgorithm ()
	Returns the type of algorithm to be used in the solution of the discrete model.
unsigned long	GetSolutionMaxIterations ()
	Returns the maximum number of iterations for the solution of a load step in nonlinear problems.
double	GetNewtonRaphsonSolutionPrecision ()
	Returns the precision required in the solution of a load step in nonlinear problems.
unsigned long	GetNumberLoadSteps ()
	Returns the number of load steps in which a load set is subdivided.
void	PostProcessing (PostProcessorProgram_E PostProgram)
	Setls the postprocessing software and runs the specific procedure to print the output file.
void	Read (FILE *FEMFile, FILE *DEFFile, FILE *LogFile)
	Reads parameters of the analysis from the .fem and .def ASCII files.
virtual void	Run ()=0
Public Attributes
TheoreticalSolution_S *	theTheoreticalSolution_S
SolverParam_S *	theSolverParam_S
PartitionData *	thePartitionData
Protected Member Functions
void	ExportGiDMeshFile (char *FileName, FEGroups &Groups, Nodes &Nds)
	Exports the FE solution to the GID format.
void	ExportGiDResFile (char *FileName, FEGroups &Groups, Nodes &Nds)
	Exports the FE solution to the GID format.
void	ExportGMSHFile (char *FileName, FEGroups &Groups, Nodes &Nds, double *Solution)
	Exports the FE solution to the GMESH format.
void	ExportHyperMeshFile (char *FileName, FEGroups &Groups, Nodes &Nds, double *Solution)
	Exports the FE solution to the HyperMesh format.
void	ComputeErrorNorm ()
	Computes the error norms in the approximated solution when a theoretical solution is given.
void	ComputeErrorNorm_Parallel (Vector &UPart)
	Computes the error norms in the approximated solution when a theoretical solution is given.
void	ComputeErrorNorm1DMatrices ()
	Computes the error norms in the approximated solution when a theoretical solution is given.
void	ComputeErrorNorm1DMatrices_Parallel (Vector &UPart)
	Computes the error norms using 1D matrices procedure in the approximated solution when a theoretical solution is given.
void	MatrixSchurComplement (SymmetricMatrix &A, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We store the factorized internal part in the same symmetric matrix for later use.
void	MatrixSchurComplement (SymmetricMatrix &A, double *Aint, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We store the factorized internal part in auxiliary array Aint for later use. This method is used in dynamic problems.
void	MatrixMatrixSchurComplement (SymmetricMatrix &A, double *B, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given matrix using blocks AB and BB from another matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. This method is used in dynamic problems.
void	MatrixVectorSchurComplement (SymmetricMatrix &A, double *B, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given matrix and vector. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We later solve the internal part by applying a linear system solver.
void	VectorSchurComplement (double *A, double *B, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given vector. The symmetric matrix is complemented in the boundary and the internal part is factorized.
void	VectorSchurComplement (double *A, double *Aint, double *B, unsigned long NBD, unsigned long NID)
	Computes the Schur complement for the given vector.
void	RecoverInternalSolSchur (double *A, double *B, double *Usol, unsigned long NBD, unsigned long NID, double c=1)
	Computes the Schur complement for the given vector. The symmetric matrix is complementd in the boundary and its internal part is factorized.
void	RecoverInternalSolSchur (double *A, double *Aint, double *B, double *Usol, unsigned long NBD, unsigned long NID, double c=1)
	Computes the Schur complement for the given vector.
void	LinearSystemSolutionContact (SymmetricMatrix &A, double *F, unsigned long LSNum)
	Solves the linear system according to the chose algorithm.
void	LinearSystemSolutionContact (SymmetricMatrix &A, double *F, unsigned long LSNum, Vector &U)
	Solves the linear system according to the chose algorithm.
void	SetLoadIncidSchur (unsigned long &NLoadedElem)
	This function sets the load incidence used to index the element load vectors used in the schur complement.
Protected Attributes
TheoreticalSolution_S	TheoSolu
	This variable stores a struct with the atributtes related to the theoretical solution.
SolverParam_S	SolvParam
	This variable stores a struct with some atributtes of the Solver class which specifies the solution parameters (solvtype, masstype, stiffness, t0, tf, Atype, etc).
Model	GlobalModel
	Instance of the Model class that store the attributes of the finite element model.
PartitionData	PartitionInformations
	Store the informations for each partition.
OneIndexTable< double >	ElemMassMatrices
	Stores the element mass matrices,.
OneIndexTable< double >	ElemStiffMatrices
	Stores the element stiffness matrices,.
OneIndexTable< double >	ElemEffMatrices
	Stores the element effective matrices.
OneIndexTable< double >	ElemInternalMatrices
	Stores the element internal factorized matrix for the schur complement, according to the problem.
OneIndexTable< double >	ElemLoadVectors
	Stores the element load vectors,.
Vector	PatchMeas
	Store the patch measures for each node of the mesh.
Vector	AuxCoords
	Auxiliary coordinates vector.
SymmetricMatrix	LocalMassMatrix
	Local mass matrix, stored as a class variable.
SymmetricMatrix	LocalStiffMatrix
	Local stiffness matrix, stored as a class variable.
OneIndexTable< double >	U
	Stores the solution vectors for each load/time step.
Vector	Displacement
OneIndexTable< double >	Upt
	Stores the velocity/first derivatives vectors for each load/time step.
OneIndexTable< double >	U2pt
	Stores the acceleration/second derivatives vectors for each load/time step.
TwoIndexTable< double >	StrainSol
	Stores the solution for the strain and extra strain fields for each load/time step.
TwoIndexTable< double >	StressSol
	Stores the solution for the stress and extra stress fields for each load/time step.
SymmetricMatrix	ContactStiffMatrix
	Global Symmetric contact stiffness matrix.
SymmetricMatrix	Kc
	Global auxiliary Symmetric contact stiffness matrix.
TwoIndexTable< double >	ExtraSol
	Stores the solution for the extra fields for each load/time step.
Vector	incU
	Vector of primary unknown increments for the NR iterations.
Vector	Fint
OneIndexTable< double >	F
	Global vector of equivalent nodal forces.
Vector	FeAux
	Auxiliary class load vector.
OneIndexTable< double >	Fgr
	Global vector of nodal force for each time step.
OneIndexTable< double >	Press
	Global vector of equivalent nodal pressure.
Vector	Load
	Vector of unbalanced forces.
Vector	ContactResidual
	Vector of contact residual.
Vector	ContStressIP
	Vector of contact stress on integration points.
Vector	ContStressCP
	Vector of contact stress on collocation points.
Vector	ContGap
	Vector of contact gap function values on integration points.
Vector	ContCondition
	Vector of contact gap condition ContCondition = 0 - integrantion point deactivated ContCondition = 1 - integrantion point activated.
Vector	StressPosition
	Vector of contact stress position.
BuiltInArray< long >	LoadIncidSchur
	This is an auxiliary index array used to attribute the different types of loads to the variable ElemLoadVectors, and is necessary to perform the Schur complement at the element load vectors.
Vector	ElemLoad
	Element load class variable. It is allocated according to the group with the highest number of degrees of freedom.
Vector	InternalForce
	Internal force vector stored as a class variable.
Vector	Damage
	Damage vector for the integration points.
OneIndexTable< double >	React
	Vector of reaction forces.
char *	PreFileName
	Prefix of the filename.
Matrix	ErrorNorm
	Stores the error norms (L2, Linf) for simulations that have a theorical solution.
long	SchurFlag
	Flag that indicates wheter the Schur complement was calculated: 1 for Schur; 0 for non Schur.
SymmetricMatrix	Schur_AA
	Auxiliary symmetric matrix used by the schur complement method.
SymmetricMatrix	Schur_BB
	Auxiliary symmetric matrix used by the schur complement method.
Matrix	Schur_AB
	Auxiliary matrix used by the schur complement method.
Matrix	Schur_AB_Aux
	Auxiliary matrix used by the schur complement method.
Matrix	Schur_Mat_Aux
	Auxiliary matrix used by the schur complement method.
Vector	Schur_Vec_Aux
	Auxiliary vector used by the schur complement.
Vector	Schur_Vec_Aux1
	Auxiliary vector used by the schur complement.
Vector	Schur_fi
	auxiliary vector used for the Schur complement.
Vector	Schur_fb
	auxiliary vector used for the Schur complement.
Vector	LoadAux
	Auxiliary vector used for element loads using the Schur complement.
Vector	InternalMassME
	Auxiliary vector used to store the internal part of the global mass matrix.

---

Detailed Description

Generic solver algorithms that store the instance of the Model class and the common variables for all solver types,.

Author:

Marco Lucio Bittencourt/ Fabiano Fernandes Bargos

Date:

April/15/2011

---

Constructor & Destructor Documentation

Solver::Solver

(

Solver &

Instance

)

Copy-initializer constructor. Copies the contents of Instance to the current object,.

Parameters:

[in]

Instance

- instance of the solver class.

Returns:

Reference to the current object.

---

Member Function Documentation

void Solver::ComputeErrorNorm1DMatrices_Parallel

(

Vector &

UPart

)

[protected]

Computes the error norms using 1D matrices procedure in the approximated solution when a theoretical solution is given.

this error norm is computed in parallel

Parameters:

[in]

UPart

- Array solution for one partition.

void Solver::ComputeErrorNorm_Parallel

(

Vector &

UPart

)

[protected]

Computes the error norms in the approximated solution when a theoretical solution is given.

this error norm is computed in parallel

Parameters:

[in]

UPart

- Array solution for one partition.

void Solver::ExportGiDMeshFile	(	char *	FileName,
		FEGroups &	Groups,
		Nodes &	Nds
	)		[protected]

Exports the FE solution to the GID format.

Parameters:

[in]	FileName	- name of the result file.
[in]	Groups	- instance of the FEGroups class which stores information for the finite element groups for the discrete model.
[in]	Nds	- instance of the Nodes class which store the nodal coordinates of the discrete model.
[in]	Solution	- pointer to the soluction array.

void Solver::ExportGiDResFile	(	char *	FileName,
		FEGroups &	Groups,
		Nodes &	Nds
	)		[protected]

Exports the FE solution to the GID format.

Parameters:

[in]	FileName	- name of the result file.
[in]	Groups	- instance of the FEGroups class which stores information for the finite element groups for the discrete model.
[in]	Nds	- instance of the Nodes class which store the nodal coordinates of the discrete model.
[in]	Solution	- pointer to the soluction array.

void Solver::ExportGMSHFile	(	char *	FileName,
		FEGroups &	Groups,
		Nodes &	Nds,
		double *	Solution
	)		[protected]

Exports the FE solution to the GMESH format.

Parameters:

[in]	FileName	- name of the result file.
[in]	Groups	- instance of the FEGroups class which stores information for the finite element groups for the discrete model.
[in]	Nds	- instance of the Nodes class which store the nodal coordinates of the discrete model.
[in]	Solution	- pointer to the soluction array.

void Solver::ExportHyperMeshFile	(	char *	FileName,
		FEGroups &	Groups,
		Nodes &	Nds,
		double *	Solution
	)		[protected]

Exports the FE solution to the HyperMesh format.

Parameters:

[in]	FileName	- name of the result file.
[in]	Groups	- instance of the FEGroups class whichStores information for the finite element groups for the discrete model.
[in]	Nds	- instance of the Nodes class which store the nodal coordinates of the discrete model.
[in]	Solution	- pointer to the soluction array.

ContactModel_E Solver::GetContactFrictionCondition

(

)

Returns the type of tangential contact condition (friction or frictionless)

Returns:

tangential contact condition

ContactSolver_E Solver::GetContactIntegrationMethod

(

)

Returns the type of contact integration method.

Returns:

contact integration method flag

unsigned long Solver::GetFrictionCoef

(

)

Returns the type of tangential contact condition (friction or frictionless)

Returns:

tangential contact condition

unsigned long Solver::GetIterativeSolutionMaxIterations

(

)

Returns the maximum number of iterations for the iterative solution of linear systems.

Returns:

maximum number of iterations as a long.

double Solver::GetIterativeSolutionPrecision

(

)

Returns the precision required in the solution of linear systems using iterative algorithms.

Returns:

precision for linear system solution as a double.

LinearSystemSolutionMethod_E Solver::GetLSMethod

(

)

Returns the type of algorithm used for linear system solution.

Returns:

algorithm for linear system solution as a LinearSystemSolutionMethod_E enum value.

double Solver::GetNewtonRaphsonSolutionPrecision

(

)

Returns the precision required in the solution of a load step in nonlinear problems.

Returns:

precision for nonlinear problems as a double.

unsigned long Solver::GetNumberLoadSteps

(

)

Returns the number of load steps in which a load set is subdivided.

Returns:

number of load steps in which a load set is subdivided as a long

OperatorType_E Solver::GetOperatorType

(

)

Returns the operator type.

Returns:

Operator type as a OperatorType_E enum value.

SolutionAlgorithm_E Solver::GetSolutionAlgorithm

(

)

Returns the type of algorithm to be used in the solution of the discrete model.

Returns:

algorithm for solution as a SolutionAlgorithm_E type.

unsigned long Solver::GetSolutionMaxIterations

(

)

Returns the maximum number of iterations for the solution of a load step in nonlinear problems.

Returns:

maximum number of iterations as a long

SolverType_E Solver::GetSolverType

(

)

Returns the solver type.

Returns:

solver type as a SolverType_E enum value.

void Solver::LinearSystemSolutionContact	(	SymmetricMatrix &	A,
		double *	F,
		unsigned long	LSNum
	)		[protected]

Solves the linear system according to the chose algorithm.

Parameters:

[in]	A	- matrix.
[in]	F	- vector.

void Solver::LinearSystemSolutionContact	(	SymmetricMatrix &	A,
		double *	F,
		unsigned long	LSNum,
		Vector &	U
	)		[protected]

Solves the linear system according to the chose algorithm.

Parameters:

[in]	A	- matrix.
[in]	F	- vector.
[out]	U	- Solution vector

void Solver::MatrixMatrixSchurComplement	(	SymmetricMatrix &	A,
		double *	B,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given matrix using blocks AB and BB from another matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. This method is used in dynamic problems.

Parameters:

[in]	A	- Matrix to perform the schur complement.
[in]	B	- Matrix which we use blocks AB and BB to complement matrix A.
[in]	NBD	- number of boundary dofs
[in]	NID	- number of internal dofs

void Solver::MatrixSchurComplement	(	SymmetricMatrix &	A,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We store the factorized internal part in the same symmetric matrix for later use.

Parameters:

[in]	A	- matrix to perform the schur complement
[in]	NBD	- number of boundary dofs
[in]	NID	- number of internal dofs

void Solver::MatrixSchurComplement	(	SymmetricMatrix &	A,
		double *	Aint,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given matrix. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We store the factorized internal part in auxiliary array Aint for later use. This method is used in dynamic problems.

Parameters:

[in]	A	- matrix to perform the schur complement
[in]	Aint	- Internal part of A, to be stored in factorized form.
[in]	NBD	- number of boundary dofs
[in]	NID	- number of internal dofs

void Solver::MatrixVectorSchurComplement	(	SymmetricMatrix &	A,
		double *	B,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given matrix and vector. We modify the boundary DOFs but we do not perform the inverse operation on the internal dofs. We later solve the internal part by applying a linear system solver.

Parameters:

[in]	A	- Symmetric matrix to perform the schur complement
[in]	B	- Vector to perform the schur complement
[in]	NBD	- number of boundary dofs
[in]	NID	- number of internal dofs

Solver & Solver::operator=

(

Solver &

Instance

)

Copies the contents of Instance to the current object,.

Parameters:

[in]

Instance

- instance of the solver class.

Returns:

Reference to the current object.

void Solver::PostProcessing

(

PostProcessorProgram_E

PostProgram

)

Setls the postprocessing software and runs the specific procedure to print the output file.

Parameters:

[in]

PostProgram

- post-processing software.

void Solver::Read	(	FILE *	FEMFile,
		FILE *	DEFFile,
		FILE *	LogFile
	)

Reads parameters of the analysis from the .fem and .def ASCII files.

Parameters:

[in]	FEMFile	- pointer to the .fem mesh file.
[in]	DEFFile	- pointer to the .def attributes file.
[in]	LogFile	- pointer to the .log file.
[in]	DataBaseName	- database name.

void Solver::RecoverInternalSolSchur	(	double *	A,
		double *	B,
		double *	Usol,
		unsigned long	NBD,
		unsigned long	NID,
		double	c = 1
	)		[protected]

Computes the Schur complement for the given vector. The symmetric matrix is complementd in the boundary and its internal part is factorized.

Parameters:

[in]	A	- Symmetric matrix used in the Schur complement
[in]	F	- Force vector, only the internal dofs
[in]	U	- Solution vector
[in]	NBD	- Number of element boundary dofs
[in]	NID	- Number of element internal dofs
[in]	const	- Constant to multiply the element mass matrix.

void Solver::RecoverInternalSolSchur	(	double *	A,
		double *	Aint,
		double *	B,
		double *	Usol,
		unsigned long	NBD,
		unsigned long	NID,
		double	c = 1
	)		[protected]

Computes the Schur complement for the given vector.

Parameters:

[in]	A	- Symmetric matrix used in the Schur complement
[in]	Aint	- double pointer containing the internal part of the symmetric matrix
[in]	F	- Force vector, only the internal dofs
[in]	U	- Solution vector
[in]	NBD	- Number of element boundary dofs
[in]	NID	- Number of element internal dofs
[in]	const	- Constant to multiply the element mass matrix.

void Solver::SetLoadIncidSchur

(

unsigned long &

NLoadedElem

)

[protected]

This function sets the load incidence used to index the element load vectors used in the schur complement.

Parameters:

[out]

NLoadedElem

- Number of loaded elements

void Solver::SetSolverAttributes

(

SolverParam_S &

SolverParameters

)

Sets the class attributes to be stored in the SolverParam_S instance.

Parameters:

[in]

SolverParameters

- instance of the SolverParam_S struct.

void Solver::SetSolverAttributes	(	SolverType_E	SolvType,
		MassStiffType_E	MassType,
		StiffnessType_E	StiffType,
		ContactModel_E	ContactMod,
		ContactSolver_E	ContactSol,
		LinearSystemSolutionMethod_E	LSMethod,
		unsigned long	LSMaxIt,
		double	LSPrec,
		SolutionAlgorithm_E	SoluAlgorithm,
		unsigned long	SolMaxIt,
		double	SolPrec,
		unsigned long	NumLoadSteps,
		AnalysisType_E	Atype,
		OperatorType_E	OpType,
		double	to,
		double	tf,
		unsigned long	NTS,
		unsigned long	FrictionCoef,
		double	PenaltyParameter,
		unsigned long	ContactInteOrder,
		TimeIntegrationMeth_E	TimeIntMeth,
		unsigned long	OutFreqNR,
		unsigned long	OutFreqTime
	)

Sets the class attributes as input parameters to be stored in the SolverParam_S instance.

Parameters:

[in]	SolvType	(SolverType_E) : Stores the type of solver used, i.e., global or element-element.
[in]	MassType	(MassType_E) : Stores the type of mass matrix (LUMPED, CONSISTENT, EIGENVALUE).
[in]	StiffType	(StiffType_E) : Stores the type of the stiffness matrix (TRANSFORMED, CONSISTENT, EIGENVALUE)
[in]	LSMethod	(LinearSystemSolutionMethod_E) : Linear system solution method.
[in]	LSMaxIt	(unsigned long) : Maximum number of iterarions for iterative linear solution method.
[in]	LSPrec	(double): Precision for the iterative solver.
[in]	SoluAlgorithm	(SolutionAlgorithm_E) : Solution scheme to solve the discrete model (linear, newton_raphson, penalty).
[in]	SolMaxIt	(unsigned long) Maximum number of iterarions for nonlinear solution algorithms.
[in]	SolPrec	(double) : Precision for the nonlinear solution algorithms.
[in]	NumLoadSteps	(unsigned long) : Number of load/time steps.
[in]	Atype	(AnalysisType_E) : Analysis type (Reynolds, Static, Transient, Modal analysis, Projection).
[in]	OpType	(OperatorType_E) : Stores the operator type (symmetric or non_symmetric).
[in]	to	(double): Initial: time.
[in]	tf	(double): Final: time.
[in]	NTS	(unsigned long) : Stores the number of time steps (NTS) for transient analysis.
[in]	TimeIntMeth	(TimeIntegrationMeth_E) : Method for integration in time.
[in]	OutFreqNR	(unsigned long): Frequency for printing the NR solution to the output file.
[in]	OutFreqTime	(unsigned long): Frequency for printing the time integration solution to the output file.

void Solver::SetTheoreticalSolutionParameters

(

TheoreticalSolution_S &

TheoSol

)

Sets the pameters for theoretical solution.

Parameters:

[in]

TheoSol

- reference to the struct that stores attributes for the Solver class related to the theorical solution.

void Solver::VectorSchurComplement	(	double *	A,
		double *	B,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given vector. The symmetric matrix is complemented in the boundary and the internal part is factorized.

Parameters:

[in]	A	- Symmetric matrix used in the Schur complement
[in]	B	- Vector to be complemented
[in]	NBD	- Number of element boundary dofs
[in]	NID	- Number of element internal dofs

void Solver::VectorSchurComplement	(	double *	A,
		double *	Aint,
		double *	B,
		unsigned long	NBD,
		unsigned long	NID
	)		[protected]

Computes the Schur complement for the given vector.

Parameters:

[in]	A	- Symmetric matrix used in the Schur complement, only boundary and coupled parts are used
[in]	Aint	- Pointer containing the internal part of the symmetric matrix.
[in]	B	- Vector to be complemented
[in]	NBD	- Number of element boundary dofs
[in]	NID	- Number of element internal dofs

---

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

• include/solvercontrol/Solver.h
• src/solvercontrol/Solver.cpp