no.uib.cipr.matrix

Class SymmBandEVD

java.lang.Object

no.uib.cipr.matrix.SymmBandEVD

public class SymmBandEVD
extends Object

Computes eigenvalues of symmetrical, banded matrices

Constructor Summary

Constructors

Constructor and Description
SymmBandEVD(int n, boolean upper) Sets up an eigenvalue decomposition for symmetrical, banded matrices.
SymmBandEVD(int n, boolean upper, boolean vectors) Sets up an eigenvalue decomposition for symmetrical, banded matrices

Method Summary

Methods

Modifier and Type	Method and Description
SymmBandEVD	factor(LowerSymmBandMatrix A) Computes the eigenvalue decomposition of the given matrix
SymmBandEVD	factor(UpperSymmBandMatrix A) Computes the eigenvalue decomposition of the given matrix
static SymmBandEVD	factorize(Matrix A, int kd) Convenience method for computing the full eigenvalue decomposition of the given matrix
double[]	getEigenvalues() Gets the eigenvalues (stored in ascending order)
DenseMatrix	getEigenvectors() Gets the eigenvectors, if available
boolean	hasEigenvectors() True if the eigenvectors have been computed

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

SymmBandEVD

public SymmBandEVD(int n,
           boolean upper)

Sets up an eigenvalue decomposition for symmetrical, banded matrices. Computes all eigenvalues and eigenvectors

Parameters:

n - Size of the matrix

upper - True if the upper part of the matrix is stored, and false if the lower part of the matrix is stored instead

SymmBandEVD

public SymmBandEVD(int n,
           boolean upper,
           boolean vectors)

Sets up an eigenvalue decomposition for symmetrical, banded matrices

Parameters:

n - Size of the matrix

upper - True if the upper part of the matrix is stored, and false if the lower part of the matrix is stored instead

vectors - True to compute the eigenvectors, false for just the eigenvalues

Method Detail

factorize

public static SymmBandEVD factorize(Matrix A,
                    int kd)
                             throws NotConvergedException

Convenience method for computing the full eigenvalue decomposition of the given matrix

Parameters:

A - Matrix to factorize. The upper triangular part is extracted, and the matrix is not modified

kd - Number of diagonals to extract

Returns:

Newly allocated decomposition

Throws:

NotConvergedException

factor

public SymmBandEVD factor(LowerSymmBandMatrix A)
                   throws NotConvergedException

Computes the eigenvalue decomposition of the given matrix

Parameters:

A - Matrix to factorize. Overwritten on return

Returns:

The current eigenvalue decomposition

Throws:

NotConvergedException

factor

public SymmBandEVD factor(UpperSymmBandMatrix A)
                   throws NotConvergedException

Computes the eigenvalue decomposition of the given matrix

Parameters:

A - Matrix to factorize. Overwritten on return

Returns:

The current eigenvalue decomposition

Throws:

NotConvergedException

getEigenvalues

public double[] getEigenvalues()

Gets the eigenvalues (stored in ascending order)

getEigenvectors

public DenseMatrix getEigenvectors()

Gets the eigenvectors, if available

hasEigenvectors

public boolean hasEigenvectors()

True if the eigenvectors have been computed