RWSymEigDecomp< TypeT > Class Template Reference

Encapsulates the eigenvalues and eigenvectors of a symmetric matrix. More...

#include <rw/lapack/symeig.h>

Public Member Functions
	RWSymEigDecomp ()
	RWSymEigDecomp (const RWSymBandMat< TypeT > &A, bool computeVecs=true)
	RWSymEigDecomp (const RWSymEigDecomp< TypeT > &A)
	RWSymEigDecomp (const RWSymMat< TypeT > &A, bool computeVecs=true)
unsigned	cols () const
TypeT	eigenValue (int i) const
const RWMathVec< TypeT >	eigenValues () const
const RWMathVec< TypeT >	eigenVector (int i) const
const RWGenMat< TypeT >	eigenVectors () const
void	factor (const RWSymBandMat< TypeT > &A, bool computeVecs=true)
void	factor (const RWSymMat< TypeT > &A, bool computeVecs=true)
bool	fail () const
bool	good () const
bool	inaccurate () const
unsigned	numEigenValues () const
unsigned	numEigenVectors () const
RWSymEigDecomp< TypeT > &	operator= (const RWSymEigDecomp< TypeT > &A)
unsigned	rows () const

Friends
template<class T1 >
class	RWHermEigDecomp
template<class T1 >
class	RWHermEigServer
class	RWSymEigServer< TypeT >

Detailed Description

template<class TypeT>
class RWSymEigDecomp< TypeT >

The class RWSymEigDecomp encapsulates the eigenvalues and eigenvectors of a symmetric matrix. You can construct an eigenvalue decomposition object in two ways:

• Use the decomposition class constructor. This has the advantage of simplicity, but allows only limited control over the computation. Your only choice is whether or not to compute eigenvectors.
• Use an eigenvalue server object. This gives you more precise control over the computation. See RWSymEigServer for details.

Synopsis

#include <rw/lapack/symeig.h>  // RWSymEigDecomp<T>
 
RWSymEigDecomp<double> eig(A);  // A is an RWSymMat<double>

Example

#include <rw/lapack/symeig.h>
#include <iostream>
 
int main() {
  RWSymMat<double> A;
  std::cin >> A;
  RWSymEigDecomp<double> eig(A);
  std::cout << "eigenvalues: " << eig.eigenValues() << std::endl;
  std::cout << "eigenvectors: " << eig.eigenVectors() << std::endl;
 
  return 0;
}

Constructor & Destructor Documentation

RWSymEigDecomp() [1/4]

template<class TypeT >

RWSymEigDecomp< TypeT >::RWSymEigDecomp

(

)

Default constructor. Builds a decomposition of a 0 x 0 matrix.

RWSymEigDecomp() [2/4]

template<class TypeT >

RWSymEigDecomp< TypeT >::RWSymEigDecomp

(

const RWSymEigDecomp< TypeT > &

A

)

Copy and precision conversion constructor. Where possible, data is referenced for efficiency.

RWSymEigDecomp() [3/4]

template<class TypeT >

RWSymEigDecomp< TypeT >::RWSymEigDecomp	(	const RWSymMat< TypeT > &	A,
		bool	computeVecs = true )

Constructs a representation of the eigenvalues and eigenvectors of the matrix A. The boolean parameter controls whether eigenvectors are computed.

RWSymEigDecomp() [4/4]

template<class TypeT >

RWSymEigDecomp< TypeT >::RWSymEigDecomp	(	const RWSymBandMat< TypeT > &	A,
		bool	computeVecs = true )

Constructs a representation of the eigenvalues and eigenvectors of the matrix A. The boolean parameter controls whether eigenvectors are computed.

Member Function Documentation

cols()

template<class TypeT >

unsigned RWSymEigDecomp< TypeT >::cols ( ) const

inline

Returns the number of columns in the decomposed matrix.

eigenValue()

template<class TypeT >

TypeT RWSymEigDecomp< TypeT >::eigenValue

(

int

i

)

const

Returns the i ^th eigenvalue.

eigenValues()

template<class TypeT >

const RWMathVec< TypeT > RWSymEigDecomp< TypeT >::eigenValues ( ) const

inline

Returns a vector of all computed eigenvalues.

eigenVector()

template<class TypeT >

const RWMathVec< TypeT > RWSymEigDecomp< TypeT >::eigenVector

(

int

i

)

const

Returns the i ^th eigenvector.

eigenVectors()

template<class TypeT >

const RWGenMat< TypeT > RWSymEigDecomp< TypeT >::eigenVectors ( ) const

inline

Returns a matrix whose columns are the eigenvectors.

factor() [1/2]

template<class TypeT >

void RWSymEigDecomp< TypeT >::factor	(	const RWSymBandMat< TypeT > &	A,
		bool	computeVecs = true )

Constructs a representation of the eigenvalues and eigenvectors of the matrix A. The boolean parameter controls whether eigenvectors are computed. The current contents of the decomposition are lost.

factor() [2/2]

template<class TypeT >

void RWSymEigDecomp< TypeT >::factor	(	const RWSymMat< TypeT > &	A,
		bool	computeVecs = true )

Constructs a representation of the eigenvalues and eigenvectors of the matrix A. The boolean parameter controls whether eigenvectors are computed. The current contents of the decomposition are lost.

fail()

template<class TypeT >

bool RWSymEigDecomp< TypeT >::fail

(

)

const

Returns true if an eigenvalue or eigenvector that is supposed to be computed fails to be computed.

good()

template<class TypeT >

bool RWSymEigDecomp< TypeT >::good

(

)

const

Returns true if all desired eigenvalues and eigenvectors are successfully computed to full desired accuracy.

inaccurate()

template<class TypeT >

bool RWSymEigDecomp< TypeT >::inaccurate

(

)

const

Returns true if either an eigenvalue or eigenvector that is supposed to be computed fails to be computed, or some of the computed quantities are not computed to full desired accuracy.

numEigenValues()

template<class TypeT >

unsigned RWSymEigDecomp< TypeT >::numEigenValues ( ) const

inline

Returns the number of eigenvalues in this object.

numEigenVectors()

template<class TypeT >

unsigned RWSymEigDecomp< TypeT >::numEigenVectors ( ) const

inline

Returns the number of eigenvectors in this object.

operator=()

template<class TypeT >

RWSymEigDecomp< TypeT > & RWSymEigDecomp< TypeT >::operator=

(

const RWSymEigDecomp< TypeT > &

A

)

Assigns this decomposition the passed value. The current contents of the decomposition are lost.

rows()

template<class TypeT >

unsigned RWSymEigDecomp< TypeT >::rows ( ) const

inline

Returns the number of rows in the decomposed matrix.