XC::Matrix Class Reference

Matrix of floats. More...

#include <Matrix.h>

Inheritance diagram for XC::Matrix:

Collaboration diagram for XC::Matrix:

Public Member Functions
	Matrix (void)
	Default constructor.
	Matrix (int nrows, int ncols)
	Constructor. More...
	Matrix (double *data, int nrows, int ncols)
	Constructor. More...
	Matrix (const Vector &v)
	Construct from vector.
	Matrix (const boost::python::list &l)
	Constructor (Python interface).
int	setData (double *newData, int nRows, int nCols)
const double *	getDataPtr (void) const
double *	getDataPtr (void)
bool	isEmpty (void) const
int	getDataSize (void) const
int	getNumBytes (void) const
int	noRows () const
	Returns the number of rows, numRows, of the Matrix.
int	noCols () const
	Returns the number of columns, numCols, of the Matrix.
void	Zero (void)
	Zero's out the Matrix. More...
void	Identity (void)
int	resize (int numRow, int numCol)
int	Assemble (const Matrix &, const ID &rows, const ID &cols, double fact=1.0)
	Assembles the argument into the current Matrix. More...
int	Solve (const Vector &V, Vector &res) const
	Solve the equation {\(Ax=V\)} for the Vector x, which is returned. More...
int	Solve (const Matrix &M, Matrix &res) const
int	Invert (Matrix &res) const
	Return the inverse matrix in the argument.
Matrix	getInverse (void) const
double	rowSum (int i) const
	Returns the sum of the i-th row.
double	columnSum (int j) const
	Returns the sum of the j-th row.
double	rowNorm (void) const
	Returns the maximum value of the elements of the vector that results that contains the row sums. More...
double	columnNorm (void) const
	Returns the maximum value of the elements of the vector that results that contains the column sums. More...
double	Norm2 (void) const
	Returns the squared modulus (euclidean norm) of the matrix.
double	Norm (void) const
	Returns the modulus (euclidean norm) of the matrix.
double	OneNorm (void) const
	Returns the value of the one norm.
double	RCond (void) const
	Return an estimation of the reciprocal of the condition number. More...
int	addMatrix (double factThis, const Matrix &other, double factOther)
	Add a factor fact times the Matrix other to the current Matrix. More...
int	addMatrixTranspose (double factThis, const Matrix &other, double factOther)
int	addMatrixProduct (double factThis, const Matrix &A, const Matrix &B, double factOther)
int	addMatrixTransposeProduct (double factThis, const Matrix &A, const Matrix &B, double factOther)
int	addMatrixTripleProduct (double factThis, const Matrix &A, const Matrix &B, double factOther)
	to perform this += T' * B * T
int	addMatrixTripleProduct (double factThis, const Matrix &A, const Matrix &B, const Matrix &C, double otherFact)
	to perform this += At * B * C
double &	operator() (int row, int col)
	Returns a reference to the data at location({row,col}). More...
const double &	operator() (int row, int col) const
	Returns a constant reference to the data at location({row,col}). More...
Matrix	operator() (const ID &rows, const ID &cols) const
	Returns a new Matrix of dimension (rows.Size(), {cols.Size()}). More...
Vector	getRow (int row) const
	Return the row which index being passed as parameter.
void	putRow (int, const Vector &)
	Put the vector at the i-th row.
Vector	getCol (int col) const
	Return the column which index being passed as parameter.
void	putCol (int, const Vector &)
	Put the vector at the i-th col.
boost::python::list	getPyList (void) const
	Return the matrix values in a Python list.
void	setPyList (const boost::python::list &)
	Populate the matrix with the values of the given list.
template<class TNSR >
Matrix &	operator= (const TNSR &)
bool	isRow (void) const
bool	isColumn (void) const
Matrix &	operator+= (double fact)
	A method to add fact to each component of the current Matrix. More...
Matrix &	operator-= (double fact)
	A method to subtract fact from each component of the current Matrix. More...
Matrix &	operator*= (double fact)
	A method to multiply each component of the current Matrix by fact. More...
Matrix &	operator/= (double fact)
	A method which will divide each component of the current Matrix by fact. More...
Matrix	operator+ (double fact) const
	A method to return a new Matrix, whose components are equal to the components of the current Matrix plus the value fact. More...
Matrix	operator- (double fact) const
	A method to return a new Matrix, whose components are equal to the components of the current Matrix minus the value fact. More...
Matrix	operator* (double fact) const
	A method to return a new Matrix, whose components are equal to the components of the current Matrix times the value fact. More...
Matrix	operator/ (double fact) const
	A method to return a new Matrix whose components are equal to the components of the current Matrix divided the value fact. More...
Vector	operator* (const Vector &V) const
	A method to return a new Vector, of size numRows, whose components are equal to the product of the current Matrix times the Vector {V}. More...
Vector	operator^ (const Vector &V) const
	A method to return a new Vector, of size numCols, whose components are equal to the product of the transpose of the current Matrix times the Vector V. More...
Matrix	operator- (void) const
	Unary minus operator.
Matrix	operator+ (const Matrix &M) const
	A method to return a new Matrix equal to the sum of the current Matrix and the Matrix M. More...
Matrix	operator- (const Matrix &M) const
	A method to return a new Matrix equal to the the current Matrix minus the Matrix M. More...
Matrix	operator* (const Matrix &M) const
	A method to return a new Matrix equal to the product of the current Matrix and the Matrix M. More...
Matrix	operator^ (const Matrix &M) const
	A method to return a new Matrix equal to the product of the transpose of the current Matrix and the Matrix M. More...
Matrix &	operator+= (const Matrix &M)
	+= operator.
Matrix &	operator-= (const Matrix &M)
	-= operator.
Matrix	getTrn (void) const
	Return the transpuesta.
void	Output (std::ostream &s) const
	Write to the argument stream.
void	Input (const std::string &)
	Read from string.
void	write (std::ofstream &)
	Write to a binary file.
void	read (std::ifstream &)
	Read from a bynary file.
int	Assemble (const Matrix &V, int init_row, int init_col, double fact=1.0)
	Assemble.
int	AssembleTranspose (const Matrix &V, int init_row, int init_col, double fact=1.0)
int	Extract (const Matrix &V, int init_row, int init_col, double fact=1.0)
std::string	toString (void) const
template<class TNSR >
XC::Matrix &	operator= (const TNSR &V)
	Assignment operator.
Public Member Functions inherited from CommandEntity
	CommandEntity (CommandEntity *owr=nullptr)
	Default constructor.
CommandEntity *	Owner (void)
	Return a pointer to the object owner.
const CommandEntity *	Owner (void) const
	Return un puntero al objeto propietario de ESTE.
const StandardOutputWrapper &	getStandardOutputWrapper (void) const
	Return the regular output stream wrapper.
StandardOutputWrapper &	getStandardOutputWrapper (void)
	Return the regular output stream wrapper.
const std::string &	getLogFileName (void) const
	Returns log file name.
void	setLogFileName (const std::string &)
	Sets log file name.
const std::string &	getErrFileName (void) const
	Returns err file name.
void	setErrFileName (const std::string &)
	Sets error file name.
const std::string &	getOutputFileName (void) const
	Returns regular output file name.
void	setOutputFileName (const std::string &)
	Sets regular output file name.
boost::python::object	evalPy (boost::python::object dict, const std::string &)
	Return the Python object that results from evaluating the argument.
boost::python::object	execPy (boost::python::object dict, const std::string &)
	Return the Python objects that results from executing the code in the string argument.
boost::python::object	execFilePy (boost::python::object dict, const std::string &)
	Return the Python object that results from executing the code in the file.
Public Member Functions inherited from EntityWithProperties
	EntityWithProperties (EntityWithProperties *owr=nullptr)
	Default constructor.
void	clearPyProps (void)
	Clear python properties map.
bool	hasPyProp (const std::string &)
	Returns true if property exists.
boost::python::object	getPyProp (const std::string &str)
	Return the Python object with the name being passed as parameter.
void	setPyProp (std::string str, boost::python::object val)
	Sets/appends a value tho the Python object's dictionary.
void	copyPropsFrom (const EntityWithProperties &)
	Copy the properties from the argument.
boost::python::list	getPropNames (void) const
	Return the names of the object properties weightings.
const PythonDict &	getPropertiesDict (void) const
	Return a std::map container with the properties of the object.
boost::python::dict	getPyDict (void) const
	Return a Python dictionary containing the object members values.
void	setPyDict (const boost::python::dict &)
	Set the values of the object members from a Python dictionary.
Public Member Functions inherited from EntityWithOwner
	EntityWithOwner (EntityWithOwner *owr=nullptr)
	Default constructor.
	EntityWithOwner (const EntityWithOwner &)
	Copy constructor.
EntityWithOwner &	operator= (const EntityWithOwner &)
	Assignment operator.
virtual bool	operator== (const EntityWithOwner &) const
	Comparison operator.
virtual	~EntityWithOwner (void)
	Virtual destructor.
virtual std::string	getClassName (void) const
	Returns demangled class name.
void	set_owner (EntityWithOwner *owr)
	Assigns the owner of the object.
EntityWithOwner *	Owner (void)
const EntityWithOwner *	Owner (void) const

Friends
class	Vector
class	Message
class	TCP_Socket
class	TCP_SocketNoDelay
class	UDP_Socket
class	MPI_Channel
std::ostream &	operator<< (std::ostream &, const Matrix &)
std::string	to_string (const Matrix &)
	Returns a string that represents the matrix.
Matrix	operator* (double, const Matrix &)

Additional Inherited Members
Public Types inherited from EntityWithProperties
typedef std::map< std::string, boost::python::object >	PythonDict
Static Public Member Functions inherited from CommandEntity
static void	resetStandardOutput (void)
	Reset the standard output streams to its defaults buffers.
Static Public Member Functions inherited from EntityWithOwner
static int	getVerbosityLevel (void)
	Get the value of the verbosity level.
static void	setVerbosityLevel (const int &)
	Set the value of the verbosity level.
Protected Member Functions inherited from CommandEntity
template<class T >
void	string_to (T &, const std::string &) const
Protected Member Functions inherited from EntityWithProperties
bool	isEqual (const EntityWithProperties &) const
	Return true if both objects are equal.
Protected Member Functions inherited from EntityWithOwner
virtual bool	isEqual (const EntityWithOwner &) const
	Return true if both objects are equal.
Static Protected Member Functions inherited from CommandEntity
static CommandEntity *	entcmd_cast (boost::any &data)
Static Protected Attributes inherited from EntityWithOwner
static int	verbosity = 1
	Object that owns THIS ONE. More...

Detailed Description

Matrix of floats.

The Matrix class provides the matrix abstraction. A matrix of order numRows X numCols is an ordered 2d array of numbers arranged in numRows rows and numCols columns. For example, a matrix \(A\) of order 2 X 3: \( A = \left[ \begin{array}{ccc} a_{0,0} & a_{0,1} & a_{0,2}\\ a_{1,0} & a_{1,1} & a_{1,2} \end{array} \right] \)

The Matrix class provides the implementation of a general unsymmetric matrix. The data for the matrix is stored in a 1d double array of size numRows*numCols with the data for \(a_{i,j}\) located at j*numRows + i in the 1d array. This is similar to the ordering of a Fortran 2d array and will permit calls to numerical Fortran libraries, e.g. BLAS, for certain method calls at a future stage. At present no subclassing is permitted (THIS MAY CHANGE), the reason for this is that the Matrix objects are envisioned to be small scale matrices primarily used for the passing of data between objects in the system. To allow subclassing could reduce the efficiency of the program due to the manner in which virtual functions are implemented.

Constructor & Destructor Documentation

Matrix() [1/2]

XC::Matrix::Matrix	(	int	nRows,
		int	nCols
	)

Constructor.

To construct a Matrix with nRows and numberOfColumns and all coefficients equal to \(0\). Allocates memory for the storage of the data. If numRows \(<\) \(0\), numCols \(<\) \(0\) or not enough memory is available available, an error message is printed and a Matrix with numRows = \(0\) and numCols = \(0\) is returned. Before the Matrix is returned, Zero() is called.

Parameters

nRows	number of matrix rows.
numberOfColumns	number of matrix columns.

Matrix() [2/2]

XC::Matrix::Matrix	(	double *	theData,
		int	nRows,
		int	nCols
	)

Constructor.

Wrapper for values contained in theData.

To construct a Matrix with nRow and nCols. The memory for storage of the data is found at the location pointed to by data. Note that this memory must have been previously allocated and it must be of appropriate size, erroneous results and segmentation faults can occur otherwise. No additional memory is allocated for the storage of the data.

Parameters

theData	values.
nRows	number of matrix rows.
nCols	number of matrix columns.

Member Function Documentation

addMatrix()

int XC::Matrix::addMatrix	(	double	factThis,
		const Matrix &	other,
		double	factOther
	)

Add a factor fact times the Matrix other to the current Matrix.

To add a factor fact times the Matrix other to the current Matrix. The size of the other Matrix is first checked to see sizes are compatible; if size are not compatible nothing is done and a warning message is printed. The method tests for the type of other, to see whether the performance can be improved by avoiding having to call other's overloaded (i,j) operators, if other is of type genMatrix.

Assemble()

int XC::Matrix::Assemble	(	const Matrix &	V,
		const ID &	rows,
		const ID &	cols,
		double	fact = 1.0
	)

Assembles the argument into the current Matrix.

Assembles into the current Matrix the Matrix M. The contents of the current matrix at location ({rows(i),cols(j)}) is set equal to the current value plus fact times the value of the matrix M at location ({i,j }). A warning is printed for every {rows(i), cols(j)} specified which is outside the bounds of the matrix.

columnNorm()

double XC::Matrix::columnNorm

(

void

)

const

Returns the maximum value of the elements of the vector that results that contains the column sums.

operator()() [1/3]

double & XC::Matrix::operator() ( int  row, int  col  )

inline

Returns a reference to the data at location({row,col}).

Returns a reference to the data at location({row,col}) in the Matrix. Assumes ({row,col}) is a valid location in the Matrix, a segmentation fault or erroneous results can occur if this is not the case.

{inline double &operator()(int row, int col);} Used to set the data at location({row,col}) in the Matrix. Assumes ({row,col}) is a valid location in the Matrix, a segmentation fault or erroneous results can occur if this is not the case.

operator()() [2/3]

const double & XC::Matrix::operator() ( int  row, int  col  ) const

inline

Returns a constant reference to the data at location({row,col}).

Returns a constant reference to the data at location({row,col}) in the Matrix. Assumes ({row,col}) is a valid location in the Matrix, a segmentation fault or erroneous results can occur if this is not the case.

{inline double &operator()(int row, int col);} Used to set the data at location({row,col}) in the Matrix. Assumes ({row,col}) is a valid location in the Matrix, a segmentation fault or erroneous results can occur if this is not the case.

operator()() [3/3]

XC::Matrix XC::Matrix::operator()	(	const ID &	rows,
		const ID &	cols
	)		const

Returns a new Matrix of dimension (rows.Size(), {cols.Size()}).

The contents of the new matrix are given by the contents of the current matrix at the locations given by the {rows} and cols objects. For example the contents of the new matrix at location ({i,j}) are equal to the contents of the current matrix at location ({rows(i),cols(j)}). Assumes ({row,col}) is a valid location in the Matrix, a segmentation fault or erroneous results can occur if this is not the case.

operator*() [1/3]

XC::Matrix XC::Matrix::operator*

(

double

fact

)

const

A method to return a new Matrix, whose components are equal to the components of the current Matrix times the value fact.

A method to return a new Matrix, whose components are equal to the components of the current Matrix times the value fact. A new Matrix object is constructed, using the current Matrix as the argument to the constructor. The {*=} operator is then invoked on this Matrix with fact as the argument, and this new Matrix is then returned.

operator*() [2/3]

XC::Vector XC::Matrix::operator*

(

const Vector &

V

)

const

A method to return a new Vector, of size numRows, whose components are equal to the product of the current Matrix times the Vector {V}.

A method to return a new Vector, of size numRows, whose components are equal to the product of the current Matrix times the Vector {V}. If the current Matrix and Vector V are not compatible, i.e. V.Size() is not equal to numCols, an error message is printed and a zero Vector of size equal to the number of rows in the current Matrix is returned. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator*() [3/3]

XC::Matrix XC::Matrix::operator*

(

const Matrix &

M

)

const

A method to return a new Matrix equal to the product of the current Matrix and the Matrix M.

A method to return a new Matrix equal to the product of the current Matrix and the Matrix M. It does this by first creating a new Matrix of size numRows and M.numCols. The contents of this new Matrix are then determined and the resulting Matrix is returned. If the two matrices are of incompatible sizes, a warning message is printed and a zeroed Matrix is returned. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if M is of type genMatrix.

operator*=()

XC::Matrix & XC::Matrix::operator*=

(

double

fact

)

A method to multiply each component of the current Matrix by fact.

A method to multiply each component of the current Matrix by fact. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator+() [1/2]

XC::Matrix XC::Matrix::operator+

(

double

fact

)

const

A method to return a new Matrix, whose components are equal to the components of the current Matrix plus the value fact.

A method to return a new Matrix, whose components are equal to the components of the current Matrix plus the value fact. A new Matrix object is constructed, using the current Matrix as the argument to the constructor. The {+=} operator is then invoked on this Matrix with fact as the argument, and the new Matrix is then returned.

operator+() [2/2]

XC::Matrix XC::Matrix::operator+

(

const Matrix &

M

)

const

A method to return a new Matrix equal to the sum of the current Matrix and the Matrix M.

A method to return a new Matrix equal to the sum of the current Matrix and the Matrix M. It does this by creating a new matrix passing itself as an argument to the constructor. The addMatrix() method is then invoked on this new Matrix with \(M\) and \(-1\) as the arguments. The new Matrix is then returned.

operator+=()

XC::Matrix & XC::Matrix::operator+=

(

double

fact

)

A method to add fact to each component of the current Matrix.

A method to add fact to each component of the current Matrix. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator-() [1/2]

XC::Matrix XC::Matrix::operator-

(

double

fact

)

const

A method to return a new Matrix, whose components are equal to the components of the current Matrix minus the value fact.

A method to return a new Matrix, whose components are equal to the components of the current Matrix minus the value fact. A new Matrix object is constructed, using the current Matrix as the argument to the constructor. The {-=} operator is then invoked on this Matrix with fact as the argument, and this new Matrix is then returned.

operator-() [2/2]

XC::Matrix XC::Matrix::operator-

(

const Matrix &

M

)

const

A method to return a new Matrix equal to the the current Matrix minus the Matrix M.

A method to return a new Matrix equal to the the current Matrix minus the Matrix M. It does this by creating a new matrix passing itself as an argument to the constructor. The addMatrix() method is then invoked on this new Matrix with \(M\) and \(-1\) as the arguments. The new Matrix is then returned.

operator-=()

XC::Matrix & XC::Matrix::operator-=

(

double

fact

)

A method to subtract fact from each component of the current Matrix.

A method to subtract fact from each component of the current Matrix. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator/()

XC::Matrix XC::Matrix::operator/

(

double

fact

)

const

A method to return a new Matrix whose components are equal to the components of the current Matrix divided the value fact.

A method to return a new Matrix whose components are equal to the components of the current Matrix divided the value fact. A new Matrix object is constructed by using the current Matrix as the argument to the constructor. The {/=} operator is then invoked on this Matrix with fact as the argument, and this new Matrix is then returned.

operator/=()

XC::Matrix & XC::Matrix::operator/=

(

double

fact

)

A method which will divide each component of the current Matrix by fact.

A method which will divide each component of the current Matrix by fact. If fact is equal to zero, an error message is printed and the contents of the Matrix are set to MATRIX_VERY_LARGE_VALUE (defined in \(<\)Matrix.h \(>\)). The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator^() [1/2]

XC::Vector XC::Matrix::operator^

(

const Vector &

V

)

const

A method to return a new Vector, of size numCols, whose components are equal to the product of the transpose of the current Matrix times the Vector V.

A method to return a new Vector, of size numCols, whose components are equal to the product of the transpose of the current Matrix times the Vector V. If the current Matrix and Vector V are not compatible, i.e. V.Size() is not equal to numRows, an error message is printed and a zero Vector of size equal to the number of columns in the current Matrix is returned. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if the current Matrix is of type genMatrix.

operator^() [2/2]

XC::Matrix XC::Matrix::operator^

(

const Matrix &

M

)

const

A method to return a new Matrix equal to the product of the transpose of the current Matrix and the Matrix M.

A method to return a new Matrix equal to the product of the transpose of the current Matrix and the Matrix M. It does this by first creating a new Matrix of size numRows and M.numRows. The contents of this new Matrix are then determined and the resulting Matrix is returned. If the two matrices are of incompatible sizes, a warning message is printed and a zeroed Matrix is returned. The method tests for the type of the current Matrix, to see whether the performance can be improved by avoiding having to call the overloaded (i,j) operators, if M is of type genMatrix.

RCond()

double XC::Matrix::RCond

(

void

)

const

Return an estimation of the reciprocal of the condition number.

estimates the reciprocal of the condition number of the matrix using the 1-norm. An estimate is obtained for norm(inv(A)), and the reciprocal of the condition number is computed as RCond = 1 / ( norm(A) * norm(inv(A)) ).

rowNorm()

double XC::Matrix::rowNorm

(

void

)

const

Returns the maximum value of the elements of the vector that results that contains the row sums.

Solve()

int XC::Matrix::Solve	(	const Vector &	b,
		Vector &	x
	)		const

Solve the equation {\(Ax=V\)} for the Vector x, which is returned.

Will solve the equation {\(Ax=V\)} for the Vector x, which is returned. At the moment the current matrix is assumed to be symmetric positive definite. THIS IS TO CHANGE. A Vector is created using V. If this Vector is not of the correct size or if an error occurs during factorization a warning message is printed and the Vector x is returned.

Zero()

void XC::Matrix::Zero

(

void

)

Zero's out the Matrix.

Zero's out the Matrix, i.e. sets all the components of the matrix to \(0\).

---

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

• src/utility/matrix/Matrix.h
• src/utility/matrix/Matrix.cpp