RWHermSomeEigServer< TypeT > Class Template Reference
[Symmetric Eigenvalue Decomposition]

The Hermitian eigenvalue server class, allowing the computation of a subset of the eigenvalues and (optionally) their corresponding eigenvectors. More...

#include <rw/lapack/heigsrv.h>

Inheritance diagram for RWHermSomeEigServer< TypeT >:

List of all members.

Public Member Functions
	RWHermSomeEigServer (bool computeVecs=true)
virtual RWSymEigDecomp< double >	decompose (const RWTriDiagDecomp< TypeT > &A)
double	setTolerance (double tolerance)
RWSlice	setRange (const RWSlice &range)
virtual bool	computeEigenVectors () const
void	computeEigenVectors (bool computeVecs)

---

Detailed Description

template<class TypeT>
class RWHermSomeEigServer< TypeT >

The Hermitian eigenvalue server class, RWHermSomeEigServer<T> allows the computation of a subset of the eigenvalues and (optionally) their corresponding eigenvectors. The computation uses the bisection method to find the eigenvalues, followed by inverse iteration to find the eigenvectors. The subset of eigenvalues to be computed is specified using the RWSlice class, or one of its subclasses, RWRange or RWToEnd. This provides the flexibility to specify either the smallest eigenvalues, the largest, or a selection in between. The eigenvalue ordering is smallest to largest.

Synopsis

 #include <rw/lapack/heigsrv.h>
 
 RWHermSomeEigServer<DComplex> server;
 RWHermEigDecomp<DComplex> eig = server(A);  // A is a
                                             // RWHermBandMat<DComplex>

Examples

 #include <iostream>
 #include <rw/lapack/heigsrv.h>
 
 int main()
 {
     RWHermMat<DComplex> A;                  // input a matrix
     std::cin >> A;
 
     RWHermSomeEigServer<DComplex> server;   // configure a server
     server.setRange(RWSlice(0,5));          // the 5 smallest eigenvalues
     RWHermEigDecomp<DComplex> eig = server(A);

     return 0;
 }

---

Constructor & Destructor Documentation

template<class TypeT>

RWHermSomeEigServer< TypeT >::RWHermSomeEigServer

(

bool

computeVecs = true

)

Constructs a server. The parameter indicates whether this server should be configured to compute eigenvectors as well as eigenvalues.

---

Member Function Documentation

template<class TypeT>

void RWHermSomeEigServer< TypeT >::computeEigenVectors

(

bool

computeVecs

)

Sets whether or not the server should compute eigenvectors as well as eigenvalues.

template<class TypeT>

virtual bool RWHermSomeEigServer< TypeT >::computeEigenVectors

(

)

const [virtual]

Returns true if this server is configured to compute eigenvectors as well as eigenvalues.

Implements RWHermEigServer< TypeT >.

template<class TypeT>

virtual RWSymEigDecomp<double> RWHermSomeEigServer< TypeT >::decompose

(

const RWTriDiagDecomp< TypeT > &

A

)

[virtual]

Computes the eigenvalue decomposition of the tridiagonal matrix inside the tridiagonal decomposition.

Implements RWHermEigServer< TypeT >.

template<class TypeT>

RWSlice RWHermSomeEigServer< TypeT >::setRange

(

const RWSlice &

range

)

Sets the range of eigenvalues to be computed. Returns the previous range.

template<class TypeT>

double RWHermSomeEigServer< TypeT >::setTolerance

(

double

tolerance

)

Sets the tolerance to which we have to compute the eigenvalues. Returns the previous tolerance.