cern.colt.matrix.tdouble.algo.decomposition

Class SparseDoubleKLUDecomposition

java.lang.Object

cern.colt.matrix.tdouble.algo.decomposition.SparseDoubleKLUDecomposition

All Implemented Interfaces:

SparseDoubleLUDecomposition

public class SparseDoubleKLUDecomposition
extends Object
implements SparseDoubleLUDecomposition

LU decomposition implemented using JKLU.

Author:

Piotr Wendykier (piotr.wendykier@gmail.com), Richard Lincoln (r.w.lincoln@gmail.com)

Constructor Summary

Constructors

Constructor and Description
SparseDoubleKLUDecomposition(DoubleMatrix2D A, int order, boolean checkIfSingular)
SparseDoubleKLUDecomposition(DoubleMatrix2D A, int order, boolean checkIfSingular, boolean preOrder) Constructs and returns a new LU Decomposition object; The decomposed matrices can be retrieved via instance methods of the returned decomposition object.

Method Summary

Methods

Modifier and Type	Method and Description
double	det() Returns the determinant, det(A).
DoubleMatrix2D	getL() Returns the lower triangular factor, L.
int[]	getPivot() Returns a copy of the pivot permutation vector.
edu.ufl.cise.klu.common.KLU_symbolic	getSymbolicAnalysis() Returns a copy of the symbolic LU analysis object
DoubleMatrix2D	getU() Returns the upper triangular factor, U.
boolean	isNonsingular() Returns whether the matrix is nonsingular (has an inverse).
void	solve(DoubleMatrix1D b) Solves A*x = b(in-place).

Methods inherited from class java.lang.Object

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

Constructor Detail

SparseDoubleKLUDecomposition

public SparseDoubleKLUDecomposition(DoubleMatrix2D A,
                            int order,
                            boolean checkIfSingular,
                            boolean preOrder)

Constructs and returns a new LU Decomposition object; The decomposed matrices can be retrieved via instance methods of the returned decomposition object.

Parameters:

A - Square matrix

order - ordering option (0 to 1); 0: AMD, 1: COLAMD

checkIfSingular - if true, then the singularity test (based on BTFJ) is performed.

preOrder - use BTF pre-ordering, or not

Throws:

IllegalArgumentException - if A is not square or is not sparse.

IllegalArgumentException - if order is not in [0,1]

SparseDoubleKLUDecomposition

public SparseDoubleKLUDecomposition(DoubleMatrix2D A,
                            int order,
                            boolean checkIfSingular)

Method Detail

det

public double det()

Description copied from interface: SparseDoubleLUDecomposition

Returns the determinant, det(A).

Specified by:

det in interface SparseDoubleLUDecomposition

getL

public DoubleMatrix2D getL()

Description copied from interface: SparseDoubleLUDecomposition

Returns the lower triangular factor, L.

Specified by:

getL in interface SparseDoubleLUDecomposition

Returns:

L

getPivot

public int[] getPivot()

Description copied from interface: SparseDoubleLUDecomposition

Returns a copy of the pivot permutation vector.

Specified by:

getPivot in interface SparseDoubleLUDecomposition

Returns:

piv

getU

public DoubleMatrix2D getU()

Description copied from interface: SparseDoubleLUDecomposition

Returns the upper triangular factor, U.

Specified by:

getU in interface SparseDoubleLUDecomposition

Returns:

U

getSymbolicAnalysis

public edu.ufl.cise.klu.common.KLU_symbolic getSymbolicAnalysis()

Description copied from interface: SparseDoubleLUDecomposition

Returns a copy of the symbolic LU analysis object

Specified by:

getSymbolicAnalysis in interface SparseDoubleLUDecomposition

Returns:

symbolic LU analysis

isNonsingular

public boolean isNonsingular()

Description copied from interface: SparseDoubleLUDecomposition

Returns whether the matrix is nonsingular (has an inverse).

Specified by:

isNonsingular in interface SparseDoubleLUDecomposition

Returns:

true if U, and hence A, is nonsingular; false otherwise.

solve

public void solve(DoubleMatrix1D b)

Description copied from interface: SparseDoubleLUDecomposition

Solves A*x = b(in-place). Upon return b is overridden with the result x.

Specified by:

solve in interface SparseDoubleLUDecomposition

Parameters:

b - A vector with of size A.rows();