Collaboration diagram for SymmetricSparse:

Public Member Functions
	SymmetricSparse ()
	Constructor without parameters.
	SymmetricSparse (long MatrixOrder, long Diagonal, long Columns, double InitValue=0.0)
	Constructor with parameters.
	SymmetricSparse (long MatrixOrder, long Diagonal, long Columns, double *Elements)
	Constructor with parameters.
	SymmetricSparse (FILE *File)
	Constructor.
	SymmetricSparse (SymmetricSparse &Instance)
	Copy constructor.
	~SymmetricSparse ()
	Destructor.
SymmetricSparse &	operator= (SymmetricSparse &Instance)
	Assignment operator '='.
SymmetricSparse &	Equal (SymmetricSparse &Instance)
	This method assigns a matrix to another.
	operator SSM_T *const ()
	Convert 'SymmetricSparse' to 'double*'.
	operator long *const ()
	Convert 'SymmetricSparse' to 'long*'.
double &	operator() (long Row, long Col)
	Overlap Opertor '()'.
double &	GetElement (unsigned long Row, unsigned long Column)
	Returns the element value and checks if it belongs to matrix.
void	SetOrder (unsigned long MatrixOrder, long *Diagonal=NULL, double Value=0.0)
	Initializes the matrix order through the elements number per row.
void	SetOrder (unsigned long MatrixOrder, long NumberElements, double Value=0.0)
	Initializes the matrix order and the elements number of the matrix.
unsigned long	GetOrder ()
	Returns the matrix order.
unsigned long	GetNumberElements ()
	Returns the elements number of the matrix.
long	GetAverElementsLine ()
	Returns average elements number per row.
void	GetMaxMinDiagElems (double &MinDiagValue, double &MaxDiagValue)
void	GetTopology (long &MatrixOrder, long &NumberElements, long &MaxNumberElLine, long &MinNumberElLine, long &AverNumberElLine)
long	SizeOf ()
	Returns the bytes number occupied by the matrix.
void	Reset (double Value=0.0)
	Reinitializes the elements with the value contained in the "Value" parameter.
void	Insert (SymmetricMatrix &SubMatrix, long *LineCol)
	Inserts a sparse matrix from the rows and columns given in the 'LineCol' vector.
void	Insert (SymmetricMatrix &SubMatrix, long *LineCol, unsigned long InitLineCol)
	Inserts a sparse matrix from the rows and columns given in the 'LineCol' vector.
void	Multiplicate (double X, double Y)
	Multiplies matrix by a vector: x = A.y.
void	Multiplicate (double scalar)
	Multiplies matrix by a scalas: A = A*scalar.
void	Add (SymmetricSparse &A, SymmetricSparse &B)
	Performs the addition between two symmetric sparse matrices. this = A + B.
void	Add (SymmetricSparse &A)
	Performs the addition between two symmetric sparse matrices. this = this + A.
void	Save (FILE *File)
	Saves the matrix in a binary file.
void	Save (char TabName, int Version, char Filename)
	Saves the matrix in the 'acdp' database.
void	Restore (FILE *File)
	Reads the matrix from a binary file.
void	Restore (char TabName, int Version, char Filename)
	Reads the matrix from the 'acdp' database.
void	Print (FILE *File, int Mant, int Dec)
	Prints the matrix in a ASCII file in the suitable format for the SymmetricSparse(FILE*) constructor.
void	Print (FILE File=stdout, const char Message="", int NumMaxCols=5, int Mant=14, int Dec=7)
	Prints the matrix in ASCII format.
void	PrintTopology (FILE *File)
	Prints the matrix topology.
void	Free ()
	Frees memory area allocated by the sparse matrix.
void	GaussSolver (double U, double B, double Precision)
	Method to solve system of linear equations.
void	GaussSolver (double Precision=1e-12)
	Method to solve system of linear equations.
void	GaussSolver (double U, double B)
	Method to solve system of linear equations.
long	Jacobi (double U, double B, unsigned long MaxNumberIter, double &NormRes, double Precision=1e-4, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
long	Jacobi (double U, double B, unsigned long MaxNumberIter)
	Method to solve system of linear equations.
long	GaussSeidel (double U, double B, unsigned long MaxNumberIter, double &NormRes, double Precision=1e-4, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
long	GaussSeidel (double U, double B, unsigned long NumberIter)
	Method to solve system of linear equations.
long	SOR (double U, double B, long MaxNumberIter, double &NormRes, double Precision=1e-4, double Omega=1.81, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
long	CG (double U, double B, long MaxNumberIter, double &NormRes, double Precision=1e-4, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
long	CGDiagonal (double U, double B, long MaxNumberIter, double &NormRes, double Precision=1e-4, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
long	CGPreConditionated (double U, double B, long MaxNumberIter, double &NormRes, double Precision=1e-4, double Omega=1.81, PreMatrix_E PreMatrix=SGS, ConvCriterion_E Crit=WEIGHT_RESIDUE, NormDS_E Norm=EUCLID)
	Method to solve system of linear equations.
Protected Member Functions
void	FreeAuxVectors ()
	Deallocates memory area used by auxiliary vectors.
void	AllocDiag (long *Diagonal=NULL)
	Allocates vector with the height of the diagonal elements .
void	Alloc (double InitValue=0.0)
	Allocates memory area to the sparse matrix.
unsigned long	GetNumElLine (long Lin)
	Returns elements number of the row.
void	TopologyEvaluation ()
	Calculates the maximum and minimum number of elements for each row.
void	MaxMinDiagEvaluation ()
	Returns the maximum and minimum absolute value of the diagonal.
int	ConvCriteria (double U, double Un, double *b, double Prec, double Normb, ConvCriterion_E Crit, NormDS_E Norm, double &NormRes)
	Checks the convergence of iterative methods by criterias following:
void	prodauxsk (double X, double Y, double cte)
	Auxiliary method used in the conjugate gradient method.
void	bcksbstsk (double U, double B, double omega)
	Auxiliary method used in the conjugate gradient method.
void	fwdsbstsk (double U, double B, double omega)
	Auxiliary method used in the conjugate gradient method.
void	prodfx (double Z, double X, double Omega)
	Auxiliary method used in the conjugate gradient method.
void	prauxsk (double U, double B)
	Auxiliary method used in the conjugate gradient method.
double	norinfv (double *u, long dim)
	Calculates the infinity norm of the array.
double	scalv (double x, double y, unsigned long dim)
	Calculates the product scalar between two vectors.
double	normaev (double *u, long dim)
	Calculate the euclidean norm of a vector.
Protected Attributes
unsigned long	PolyOrder
	The matrix order.
unsigned long	MaxNumElLine
	Maximum number of the elements by row.
unsigned long	MinNumElLine
	Minimum number of the elements by row.
long *	Pos
	Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.
long *	Diag
	Pointers for beginning of rows.
double *	d
	Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.
double *	f
	Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.
double *	g
	Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.
double *	h
	Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.
SSM_T *	SSMatrix
	Matrix elements stored as a array.
double	MaxDiagElem
	Maximum value of the diagonal.
double	MinDiagElem
	Minimum value of the diagonal.
Static Protected Attributes
static double	Zero
	Coefficient null or nonexistent.

Constructor & Destructor Documentation

SymmetricSparse::SymmetricSparse	(	long	MatrixOrder,
		long *	Diagonal,
		long *	Columns,
		double	InitValue = `0.0`
	)

Constructor with parameters.

Initializes attributes, allocates memory area to the matrix, initializes the columns with values in "Columns" and the elements with value in "InitValue".

Parameters:

[in]	MatrixOrder	- The order matrix.
[in]	Diagonal	- Elements of the diagonal.
[in]	Columns.
[in]	InitValue	- The initial value to the matrix.

SymmetricSparse::SymmetricSparse	(	long	MatrixOrder,
		long *	Diagonal,
		long *	Columns,
		double *	Elements
	)

Constructor with parameters.

Initializes attributes, allocates memory area to the matrix, Initializes columns with values in "Columns" and elements with values in "Elements".

Parameters:

[in]	MatrixOrder	- The order matrix.
[in]	Diagonal	- Elements of the diagonal.
[in]	Columns.
[in]	Elements.

SymmetricSparse::SymmetricSparse ( FILE * File )

Constructor.

Reads from a ASCII file with the elements of the matrix. The ASCII file contains the matrix order (the long type), the diagonal (a vector the long type) and elements of matrix (vector the SSM_T type).

Parameters:

[in] File - A binary file.

SymmetricSparse::SymmetricSparse ( SymmetricSparse & Instance )

Copy constructor.

Parameters:

[in] Instance - symmetric sparse type.

Member Function Documentation

void SymmetricSparse::Add	(	SymmetricSparse &	A,
		SymmetricSparse &	B
	)

Performs the addition between two symmetric sparse matrices. this = A + B.

Parameters:

[in]	A	- Symmetric sparse matrix;
[in]	B	- Symmetric sparse matrix;

void SymmetricSparse::Add ( SymmetricSparse & A )

Performs the addition between two symmetric sparse matrices. this = this + A.

Parameters:

[in] A - Symmetric sparse matrix;

void SymmetricSparse::Alloc ( double InitValue = 0.0 ) [protected]

Allocates memory area to the sparse matrix.

Parameters:

[in] InitValue - Value that will be assigned for all elements of matrix.

void SymmetricSparse::AllocDiag ( long * Diagonal = NULL ) [protected]

Allocates vector with the height of the diagonal elements .

Parameters:

[in] Diagonal - diagonal elements of matrix.

void SymmetricSparse::bcksbstsk	(	double *	U,
		double *	B,
		double	omega
	)		`[protected]`

Auxiliary method used in the conjugate gradient method.

Makes the step of "back substitution" of the system: A * U = B.

Parameters:

[in]	B	- the residue vector.
[in]	omega.
[out]	U	- the answer vector.

long SymmetricSparse::CG	(	double *	U,
		double *	B,
		long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

Standard conjugate gradient method.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- Maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- constant for the convergence test.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

long SymmetricSparse::CGDiagonal	(	double *	U,
		double *	B,
		long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

Conjugate gradient with the diagonal conditioning sparse matrix.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- Maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- constant for the convergence test.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

long SymmetricSparse::CGPreConditionated	(	double *	U,
		double *	B,
		long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		double	Omega = `1.81`,
		PreMatrix_E	PreMatrix = `SGS`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

Conjugte gradient method with preconditioning SSOR (Symmetric Successive Overrelaxation Method) and GS (Gauss-Seidel Method) symmetric.

This method returns zero or the iterations number. if the calculus didn't converge and the maximum iterations number was overcome this method will returns zero. And if the calculus converged it will returns iterations number.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- Maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- constant for the convergence test.
[in]	Omega	- parameter of relaxation.
[in]	PreMatrix	- It is a enum type that indicates the preconditioning matrix type.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

int SymmetricSparse::ConvCriteria	(	double *	U,
		double *	Un,
		double *	b,
		double	Prec,
		double	Normb,
		ConvCriterion_E	Crit,
		NormDS_E	Norm,
		double &	NormRes
	)		`[protected]`

Checks the convergence of iterative methods by criterias following:

Crit = 1: ||un+1 - un|| Crit = 2: ||rn+1|| = ||Aun+1 - b|| Crit = 3: ||rn+1|| = ||Aun+1 - b||/||b|| Norm = 1: ||.|| = euclidean norm Norm = 2: ||.|| = infinity norm

Parameters:

[in]	U	- pointer to vector operand.
[in]	b.
[in]	Prec.
[in]	Normb.
[in]	Crit	- criterion of convergence.
[in]	Norm.
[in]	NormRes.
[out]	Un	- Pointer to vector that contains the result.

Returns:: 0 or 1 - indicating if converges or not.

SymmetricSparse & SymmetricSparse::Equal ( SymmetricSparse & Instance )

This method assigns a matrix to another.

Parameters:

[in] Instance - symmetric sparse type.

void SymmetricSparse::fwdsbstsk	(	double *	U,
		double *	B,
		double	omega
	)		`[protected]`

Auxiliary method used in the conjugate gradient method.

Makes the step of "forward substitution" of the system: A * U = B.

Parameters:

[in]	B	- the residue vector.
[in]	omega.
[out]	U	- the answer vector.

long SymmetricSparse::GaussSeidel	(	double *	U,
		double *	B,
		unsigned long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

Gauss-Seidel iterative method.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- The maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- precision desired.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

long SymmetricSparse::GaussSeidel	(	double *	U,
		double *	B,
		unsigned long	NumberIter
	)

Method to solve system of linear equations.

Gauss-Seidel iterative method. Runs a fixed number of iterations without verify the convergence.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	NumberIter	- The number of iterations.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

void SymmetricSparse::GaussSolver	(	double *	U,
		double *	B,
		double	Precision
	)

Method to solve system of linear equations.

Gauss elimination method. U - solution and B - independent term)

Parameters:

[in]	B	- double pointer to independent terms.
[in]	Precision.
[out]	U	- double pointer to answer vector.

void SymmetricSparse::GaussSolver ( double Precision = 1e-12 )

Method to solve system of linear equations.

Matrix triangularization by the gauss method.

Parameters:

[in] Precision

void SymmetricSparse::GaussSolver	(	double *	U,
		double *	B
	)

Method to solve system of linear equations.

Solves system by the Gauss method - reduces vector and back substitution.

Parameters:

[in]	B	- double pointer to independent terms.
[out]	U	- double pointer to answer vector.

long SymmetricSparse::GetAverElementsLine ( )

Returns average elements number per row.

Returns:: a long type with the value of the average elements number per row.

double & SymmetricSparse::GetElement	(	unsigned long	Row,
		unsigned long	Column
	)

Returns the element value and checks if it belongs to matrix.

Parameters:

[in]	Row	- Indicates the row number of the sparse matrix.
[in]	Col	- Indicates the columns number of the sparse matrix.

Returns:: a double type contains a element of sparse matrix.

void SymmetricSparse::GetMaxMinDiagElems	(	double &	MinDiagValue,
		double &	MaxDiagValue
	)

the maximum and minimum absolute values of the diagonal.

Parameters:

[out]	MinDiagValue	- The minimum value of the diagonal.
[out]	MaxDiagValue	- The maximum value of the diagonal.

unsigned long SymmetricSparse::GetNumberElements ( )

Returns the elements number of the matrix.

Returns:: a long type with elements number.

unsigned long SymmetricSparse::GetNumElLine ( long Lin ) [protected]

Returns elements number of the row.

Parameters:

[in] Lin - the number of row.

Returns:: the size of row.

unsigned long SymmetricSparse::GetOrder ( )

Returns the matrix order.

Returns:: a long type with the matrix order.

void SymmetricSparse::GetTopology	(	long &	MatrixOrder,
		long &	NumberElements,
		long &	MaxNumberElLine,
		long &	MinNumberElLine,
		long &	AverNumberElLine
	)

the matrix topology.

Parameters:

[out]	MatrixOrder	- The order of matrix.
[out]	NumberElements	- Number of elements not null of the matrix.
[out]	MaxNumberElLine	- Maximum number of elements per row.
[out]	MinNumberElLine	- Minimum number of elements per row.
[out]	AverNumberElLine	- Average number of elements per row.

void SymmetricSparse::Insert	(	SymmetricMatrix &	SubMatrix,
		long *	LineCol
	)

Inserts a sparse matrix from the rows and columns given in the 'LineCol' vector.

Parameters:

[in]	SubMatrix	- Sub-Matrix to be inserted in the sparse matrix.
[in]	LineCol	- The row and column where will be inserted the sub-matrix.

void SymmetricSparse::Insert	(	SymmetricMatrix &	SubMatrix,
		long *	LineCol,
		unsigned long	InitLineCol
	)

Inserts a sparse matrix from the rows and columns given in the 'LineCol' vector.

Parameters:

[in]	SubMatrix	- Sub-Matrix to be inserted in the sparse matrix.
[in]	LineCol	- The row and column where will be inserted the sub-matrix.
[in]	InitLineCol	- A variable used for verification of free degrees of freedom. For example, if the current column or line is lesser than this variable, it inserts the current value.

long SymmetricSparse::Jacobi	(	double *	U,
		double *	B,
		unsigned long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

Jacobi's iterative method.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- The maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- precision desired.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

long SymmetricSparse::Jacobi	(	double *	U,
		double *	B,
		unsigned long	MaxNumberIter
	)

Method to solve system of linear equations.

Jacobi's iterative method. Runs a fixed number of iterations without verify the convergence.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- The maximum number of iterations.
[out]	U	- double pointer to answer vector.

void SymmetricSparse::Multiplicate	(	double *	X,
		double *	Y
	)

Multiplies matrix by a vector: x = A.y.

Parameters:

[in]	Y	- Entry vector to be multiplied by the sparse matrix.
[out]	X	- Result vector of multiplication the sparse matrix by the vector "Y".

void SymmetricSparse::Multiplicate ( double scalar )

Multiplies matrix by a scalas: A = A*scalar.

Parameters:

[in] scalar - Scalar;

double SymmetricSparse::norinfv	(	double *	u,
		long	dim
	)		`[protected]`

Calculates the infinity norm of the array.

Parameters:

[in]	*u	- array of double.
[in]	dim	- dimension of the array u.

Returns:: a double - the infinity norm of an array.

double SymmetricSparse::normaev	(	double *	u,
		long	dim
	)		`[protected]`

Calculate the euclidean norm of a vector.

Parameters:

[in]	*u	- the vector which will be calculated the euclidean norm.
[in]	dim	- dimension of vector u.

Returns:: a double - the value of the euclidean norm of a vector.

SymmetricSparse::operator long *const ( )

Convert 'SymmetricSparse' to 'long*'.

Returns:: a double pointer to the vector with the indexes of the initial elements of each row of matrix.

SymmetricSparse::operator SSM_T *const ( )

Convert 'SymmetricSparse' to 'double*'.

Returns:: a double pointer to the data struct with the elements and indexes of columns.

double & SymmetricSparse::operator()	(	long	Row,
		long	Col
	)

Overlap Opertor '()'.

Returns the value of the position '(Row, Col)'

Parameters:

[in]	Row	- Indicates the row number of the sparse matrix.
[in]	Col	- Indicates the columns number of the sparse matrix.

Returns:: a double type contains a element of sparse matrix.

SymmetricSparse & SymmetricSparse::operator= ( SymmetricSparse & Instance )

Assignment operator '='.

Ex.: A = B;

Parameters:

[in] Instance - symmetric sparse type.

void SymmetricSparse::prauxsk	(	double *	U,
		double *	B
	)		`[protected]`

Auxiliary method used in the conjugate gradient method.

Solve the system D.x = b (where D is a diagonal matrix).

Parameters:

[in]	B.
[out]	U.

void SymmetricSparse::Print	(	FILE *	File,
		int	Mant,
		int	Dec
	)

Prints the matrix in a ASCII file in the suitable format for the SymmetricSparse(FILE*) constructor.

This format is following: the matrix order, diagonal vector and the coeficients (row, columns, value). Example: 3 0 2 4 5 0 0 1.2 0 1 2.0 1 1 3.2 1 2 5.4 2 2 10.84

Parameters:

[in]	File	- the file where will be printed the matrix data.
[in]	Title	- this a header to indicate a operation type.
[in]	Format	- the data format to be printed.

void SymmetricSparse::Print	(	FILE *	File = `stdout`,
		const char *	Message = `""`,
		int	NumMaxCols = `5`,
		int	Mant = `14`,
		int	Dec = `7`
	)

Prints the matrix in ASCII format.

Prints the band of matrix by row.

Parameters:

[in]	File	- the file where will be printed the sparse matrix data. (the default is stdout)
[in]	Message	- Indicates the operation type that was run.
[in]	NumMaxCols	- the maximum value of columns that can be printed by row.
[in]	Mant	- Indicates the space size that should have between columns.
[in]	Dec	- Indicates the size of the decimal part to be printed.

void SymmetricSparse::PrintTopology ( FILE * File )

Prints the matrix topology.

Parameters:

[in] File - the file where will be printed the topology of sparse matrix. (the default is stdout)

void SymmetricSparse::prodauxsk	(	double *	X,
		double *	Y,
		double	cte
	)		`[protected]`

Auxiliary method used in the conjugate gradient method.

This method makes the product between the element of main diagonal of the matrix of system and the element correspondent of the output 'y' vector provided by fwdsbstsk function.

Parameters:

[in]	Y.
[in]	cte.
[out]	X.

void SymmetricSparse::prodfx	(	double *	Z,
		double *	X,
		double	Omega
	)		`[protected]`

Auxiliary method used in the conjugate gradient method.

Makes the product "z = F.x", where F is upper triangular matrix.

Parameters:

[in]	X.
[in]	omega.
[out]	Z.

void SymmetricSparse::Reset ( double Value = 0.0 )

Reinitializes the elements with the value contained in the "Value" parameter.

Parameters:

[in] Value - The value to reinitialize the sparse matrix.

void SymmetricSparse::Restore ( FILE * File )

Reads the matrix from a binary file.

Parameters:

[in] File - the binary file where the sparse matrix will be read.

void SymmetricSparse::Restore	(	char *	TabName,
		int	Version,
		char *	Filename
	)

Reads the matrix from the 'acdp' database.

Parameters:

[in]	TabName	- the table name of the database to read sparse matrix data.
[in]	Version	- the version of the read data.
[in]	Filename	- the database name.

void SymmetricSparse::Save ( FILE * File )

Saves the matrix in a binary file.

Parameters:

[in] File - the file to save the sparse matrix data.

void SymmetricSparse::Save	(	char *	TabName,
		int	Version,
		char *	Filename
	)

Saves the matrix in the 'acdp' database.

Parameters:

[in]	TabName	- the table name of the database to saved sparse matrix data.
[in]	Version	- the version of the update data.
[in]	Filename	- Stores the database name.

double SymmetricSparse::scalv	(	double *	x,
		double *	y,
		unsigned long	dim
	)		`[protected]`

Calculates the product scalar between two vectors.

Parameters:

[in]	*x	- first vector.
[in]	*y	- second vector.
[in]	dim	- dimension of vector x and y.

Returns:: a double - the value of the product scalar between vectors x and y.

void SymmetricSparse::SetOrder	(	unsigned long	MatrixOrder,
		long *	Diagonal = `NULL`,
		double	Value = `0.0`
	)

Initializes the matrix order through the elements number per row.

Parameters:

[in]	MatrixOrder	- The order sparse matrix.
[in]	Diagonal	- Diagonal elements of the sparse matrix.
[in]	Value	- The initial value to the sparse matrix.

void SymmetricSparse::SetOrder	(	unsigned long	MatrixOrder,
		long	NumberElements,
		double	Value = `0.0`
	)

Initializes the matrix order and the elements number of the matrix.

Parameters:

[in]	MatrixOrder	- The order sparse matrix.
[in]	NumberElements.
[in]	Value	- The initial value to the sparse matrix.

long SymmetricSparse::SizeOf ( )

Returns the bytes number occupied by the matrix.

Returns:: a long type with the bytes value that matrix occupies.

long SymmetricSparse::SOR	(	double *	U,
		double *	B,
		long	MaxNumberIter,
		double &	NormRes,
		double	Precision = `1e-4`,
		double	Omega = `1.81`,
		ConvCriterion_E	Crit = `WEIGHT_RESIDUE`,
		NormDS_E	Norm = `EUCLID`
	)

Method to solve system of linear equations.

SOR iterative method.

Parameters:

[in]	B	- double pointer to independent terms.
[in]	MaxNumberIter	- The maximum number of iterations.
[in]	NormRes	- the residue norm of solution.
[in]	Precision	- constant for the convergence test.
[in]	Omega	- parameter of relaxation.
[in]	Crit	- convergence criterion.
[in]	Norm	- Norm type. 1 - Infinity norm 2 - Euclidean norm.
[out]	U	- double pointer to answer vector.

Returns:: a long type indicating if there was convergence.

Member Data Documentation

double* SymmetricSparse::d [protected]

Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.

Vector used in the iterative methods.

double* SymmetricSparse::f [protected]

Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.

Vector used in the iterative methods.

double* SymmetricSparse::g [protected]

Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.

Vector used in the iterative methods.

double* SymmetricSparse::h [protected]

Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.

Vector used in the iterative methods.

long* SymmetricSparse::Pos [protected]

Auxiliary vector used in the methods for solve system of linear equations and for the "Insert" method.

Stores the position of the elements along a row.

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

include/ds/SymmetricSparse.h
src/ds/SymmetricSparse.cpp

Public Member Functions

Protected Member Functions

Protected Attributes

Static Protected Attributes

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation