jsat.linear

Class LUPDecomposition

java.lang.Object

jsat.linear.LUPDecomposition

All Implemented Interfaces:

Serializable, Cloneable

public class LUPDecomposition
extends Object
implements Cloneable, Serializable

This class uses the LUP decomposition of a matrix to provide efficient methods for solving A x = b, as well as computing the determinant of A.

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
LUPDecomposition(Matrix A)
LUPDecomposition(Matrix A, ExecutorService threadpool)
LUPDecomposition(Matrix L, Matrix U, Matrix P)

Method Summary

Modifier and Type	Method and Description
static Matrix	backSub(Matrix U, Matrix y) Solves for the matrix x such that U x = y
static Matrix	backSub(Matrix U, Matrix y, ExecutorService threadpool) Solves for the matrix x such that U x = y
static Vec	backSub(Matrix U, Vec y) Solves for the vector x such that U x = y
LUPDecomposition	clone()
double	det()
static Matrix	forwardSub(Matrix L, Matrix b) Solves for the matrix x such that L x = b
static Matrix	forwardSub(Matrix L, Matrix b, ExecutorService threadpool) Solves for the matrix x such that L x = b
static Vec	forwardSub(Matrix L, Vec b) Solves for the vector x such that L x = b
boolean	isSquare()
Matrix	solve(Matrix B)
Matrix	solve(Matrix B, ExecutorService threadpool)
Vec	solve(Vec b)

Methods inherited from class java.lang.Object

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

Constructor Detail

LUPDecomposition

public LUPDecomposition(Matrix L,
                        Matrix U,
                        Matrix P)

LUPDecomposition

public LUPDecomposition(Matrix A)

LUPDecomposition

public LUPDecomposition(Matrix A,
                        ExecutorService threadpool)

Method Detail

isSquare

public boolean isSquare()

Returns:

true if the original matrix A, from which this factorization is from, is a square matrix

det

public double det()

Returns:

the determinant of the original Matrix A, |A|

solve

public Vec solve(Vec b)

solve

public Matrix solve(Matrix B)

solve

public Matrix solve(Matrix B,
                    ExecutorService threadpool)

clone

public LUPDecomposition clone()

Overrides:

clone in class Object

forwardSub

public static Vec forwardSub(Matrix L,
                             Vec b)

Solves for the vector x such that L x = b

Parameters:

L - a lower triangular matrix

b - a vector whos length is equal to the rows in L

Returns:

x such that L x = b

forwardSub

public static Matrix forwardSub(Matrix L,
                                Matrix b)

Solves for the matrix x such that L x = b

Parameters:

L - a lower triangular matrix

b - a matrix with the same number of rows as L

Returns:

x such that L x = b

forwardSub

public static Matrix forwardSub(Matrix L,
                                Matrix b,
                                ExecutorService threadpool)

Solves for the matrix x such that L x = b

Parameters:

L - a lower triangular matrix

b - a matrix with the same number of rows as L

threadpool - source of threads for the parallel computation

Returns:

x such that L x = b

backSub

public static Vec backSub(Matrix U,
                          Vec y)

Solves for the vector x such that U x = y

Parameters:

U - an upper triangular matrix

y - a vector whos length is equal to the rows in U

Returns:

x such that U x = y

backSub

public static Matrix backSub(Matrix U,
                             Matrix y)

Solves for the matrix x such that U x = y

Parameters:

U - an upper triangular matrix

y - a matrix with the same number of rows as U

Returns:

x such that U x = y

backSub

public static Matrix backSub(Matrix U,
                             Matrix y,
                             ExecutorService threadpool)

Solves for the matrix x such that U x = y

Parameters:

U - an upper triangular matrix

y - a matrix with the same number of rows as U

threadpool - source of threads for the parallel computation

Returns:

x such that U x = y