org.ojalgo.matrix.decomposition
Interface MatrixDecomposition<N extends Number>

All Superinterfaces:

Serializable

All Known Subinterfaces:

Cholesky<N>, Eigenvalue<N>, LU<N>, QR<N>, SingularValue<N>

All Known Implementing Classes:

CholeskyDecomposition, EigenvalueDecomposition, JamaCholesky, JamaEigenvalue, JamaLU, JamaQR, JamaSingularValue, LUDecomposition, QRDecomposition, SingularValueDecomposition

public interface MatrixDecomposition<N extends Number>
extends Serializable

Some standard matrix names:

• [A] could be any matrix.
• [I] is an identity matrix.
• [P] is a permutation matrix - an identity matrix with interchanged rows (or columns).
• [L] is a lower (left) triangular matrix.
• [D] is a diagonal matrix. Possibly bi-, tri- or block-diagonal.
• [U] is an upper (right) triangular matrix. It is equivalent to [R].
• [Q] is an orthogonal, or unitary, matrix. [Q][Q]^T=[I].
• [R] is a right (upper) tringular matrix. It is equivalent to [U].
• [V] is an eigenvector matrix.

With the above definitions we can express various decompositions as:

• LU: [A] = [P][L][D][U]
• Cholesky: [A] = [R]^T[R]
• QR: [A] = [Q][R]
• Singular Value: [A] = [Q1][D][Q2]^T
• Eigenvalue: [A] = [V][D][V]^-1

Author:

apete

Method Summary

boolean	compute(MatrixStore<N> aStore)
boolean	equals(BasicMatrix aMtrx, NumberContext aCntxt)
boolean	equals(MatrixStore<N> aStore, NumberContext aCntxt)
MatrixStore<N>	getInverse() The output must be a "right inverse" and a "generalised inverse".
MatrixStore<N>	invert(MatrixStore<N> aStore) A convenience method that produces exactly the same result as if you first call compute(MatrixStore) and then getInverse().
boolean	isComputed()
boolean	isFullSize()
boolean	isSolvable()
void	reset()
MatrixStore<N>	solve(MatrixStore<N> aRHS)
Future<DecomposeAndSolve<N>>	solve(MatrixStore<N> aBody, MatrixStore<N> aRHS) Will solve [aBody][X]=[aRHS] concurrently by first calling compute(MatrixStore) using [aBody], and then solve(MatrixStore) using [aRHS].

Method Detail

compute

boolean compute(MatrixStore<N> aStore)

Parameters:

aStore - A matrix to decompose

Returns:

true if the computation suceeded; false if not

equals

boolean equals(BasicMatrix aMtrx,
               NumberContext aCntxt)

equals

boolean equals(MatrixStore<N> aStore,
               NumberContext aCntxt)

getInverse

MatrixStore<N> getInverse()

The output must be a "right inverse" and a "generalised inverse".

See Also:

BasicMatrix.invert()

invert

MatrixStore<N> invert(MatrixStore<N> aStore)

A convenience method that produces exactly the same result as if you first call compute(MatrixStore) and then getInverse().

isComputed

boolean isComputed()

Returns:

true if computation has been attemped; false if not.

See Also:

compute(MatrixStore), isSolvable()

isFullSize

boolean isFullSize()

Returns:

True if the implementation generates a full sized decomposition.

isSolvable

boolean isSolvable()

Returns:

true if it is ok to call solve(MatrixStore) (computation was successful); false if not

See Also:

solve(MatrixStore), isComputed()

reset

void reset()

solve

MatrixStore<N> solve(MatrixStore<N> aRHS)

solve

Future<DecomposeAndSolve<N>> solve(MatrixStore<N> aBody,
                                   MatrixStore<N> aRHS)

Will solve [aBody][X]=[aRHS] concurrently by first calling compute(MatrixStore) using [aBody], and then solve(MatrixStore) using [aRHS]. If either of the input [aBody] or [aRHS] is set to null the corresponing calculation is skipped.

Parameters:

aBody - The equation system body

aRHS - The equation system right hand side

Returns:

The matrix decomposition and the equation system solution, [X]