fr.cnes.sirius.patrius.math.optim.joptimizer.algebra

Class Matrix1NornRescaler

java.lang.Object

fr.cnes.sirius.patrius.math.optim.joptimizer.algebra.Matrix1NornRescaler

All Implemented Interfaces:

MatrixRescaler

public final class Matrix1NornRescaler
extends Object
implements MatrixRescaler

Calculates the matrix rescaling factors so that the 1-norm of each row and each column of the scaled matrix asymptotically converges to one.

Since:

4.6

Author:

alberto trivellato (alberto.trivellato@gmail.com) See Daniel Ruiz, "A scaling algorithm to equilibrate both rows and columns norms in matrices" See Philip A. Knight, Daniel Ruiz, Bora Ucar "A Symmetry Preserving Algorithm for Matrix Scaling"

Constructor Summary

Constructors

Constructor and Description
Matrix1NornRescaler() Constructor
Matrix1NornRescaler(double epsilon) Constructor

Method Summary

Modifier and Type	Method and Description
boolean	checkScaling(RealMatrix aOriginal, RealVector u, RealVector v) Check if the scaling algorithm returned proper results.
RealVector[]	getMatrixScalingFactors(RealMatrix a) Scaling factors for not singular matrices.
RealVector	getMatrixScalingFactorsSymm(RealMatrix a) Scaling factors for symmetric (not singular) matrices.
static double	getRowInfinityNorm(RealMatrix aSymm, int r)

Methods inherited from class java.lang.Object

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

Constructor Detail

Matrix1NornRescaler

public Matrix1NornRescaler()

Constructor

Matrix1NornRescaler

public Matrix1NornRescaler(double epsilon)

Constructor

Parameters:

epsilon - epsilon value

Method Detail

getMatrixScalingFactors

public RealVector[] getMatrixScalingFactors(RealMatrix a)

Scaling factors for not singular matrices. See Daniel Ruiz, "A scaling algorithm to equilibrate both rows and columns norms in matrices" See Philip A. Knight, Daniel Ruiz, Bora Ucar "A Symmetry Preserving Algorithm for Matrix Scaling"

Specified by:

getMatrixScalingFactors in interface MatrixRescaler

Parameters:

a - matrix A

Returns:

array with R,T

getMatrixScalingFactorsSymm

public RealVector getMatrixScalingFactorsSymm(RealMatrix a)

Scaling factors for symmetric (not singular) matrices. Just the subdiagonal elements of the matrix are required. See Daniel Ruiz, "A scaling algorithm to equilibrate both rows and columns norms in matrices" See Philip A. Knight, Daniel Ruiz, Bora Ucar "A Symmetry Preserving Algorithm for Matrix Scaling"

Specified by:

getMatrixScalingFactorsSymm in interface MatrixRescaler

Parameters:

a - matrix A

Returns:

array with R,T

checkScaling

public boolean checkScaling(RealMatrix aOriginal,
                            RealVector u,
                            RealVector v)

Check if the scaling algorithm returned proper results. Note that AOriginal cannot be only subdiagonal filled, because this check is for both symm and bath notsymm matrices.

Specified by:

checkScaling in interface MatrixRescaler

Parameters:

aOriginal - the ORIGINAL (before scaling) matrix

u - the return of the scaling algorithm

v - the return of the scaling algorithm

Returns:

true if scaled

getRowInfinityNorm

public static double getRowInfinityNorm(RealMatrix aSymm,
                                        int r)

Parameters:

aSymm - symm matrix filled in its subdiagonal elements

r - the index of the row

Returns:

infinity norm