ndarray-crout-decomposition

LU decomposition using the crout algorithm

example

var ndarray = require('ndarray');
var zeros = require('zeros');
var show = require('ndarray-show');
var crout = require('ndarray-crout-decomposition');

var A = ndarray([ 4, 3, 6, 3 ], [ 2, 2 ]);
var L = zeros([ 2, 2 ]);
var U = zeros([ 2, 2 ]);

crout(A, L, U);
console.log('L=\n' + show(L));
console.log('U=\n' + show(U));

output:

L=
   4.000    0.000
   3.000   -1.500
U=
   1.000    1.500
   0.000    1.000

or to save space, you can use a single matrix to store the L and U values:

var ndarray = require('ndarray');
var zeros = require('zeros');
var show = require('ndarray-show');
var crout = require('ndarray-crout-decomposition');

var A = ndarray([ 4, 3, 6, 3 ], [ 2, 2 ]);
var LU = zeros([ 2, 2 ]);

crout(A, LU);
console.log('LU=\n' + show(LU));

output:

LU=
   4.000    1.500
   3.000   -1.500

methods

var crout = require('ndarray-crout-decomposition')

var ok = crout(A, L, U)

Decompose the matrix A into L and U, mutating L and U in-place.

A is not modified.

If you don't pass in a U matrix, L will be used to store both the L and U values, omitting the diagonal of ones from U to make room.

If A was non-square or the algorithm could not find a solution, ok is false. Otherwise ok is true.

install

With npm do:

npm install ndarray-crout-decomposition

license

MIT