conjugate-gradient
Conjugate gradient solver
npm install conjugate-gradient
Want to see pretty graphs? Log in now!
1 | downloads in the last day |
1 | downloads in the last week |
17 | downloads in the last month |
Last Published By | |
---|---|
Version | 0.0.2 last updated 20 days ago |
License | MIT |
Keywords | linear, solve, algebra, math, numerical, method, conjugate, gradient, preconditioner, jacobi, diagonal, iterative, sparse, matrix, vector, physics |
Repository | git://github.com/mikolalysenko/conjugate-gradient.git (git) |
Homepage | https://github.com/mikolalysenko/conjugate-gradient |
Bugs | https://github.com/mikolalysenko/conjugate-gradient/issues |
Dependencies | almost-equal, bit-twiddle |
Dependents | mikolalysenko-hoarders |
conjugate-gradient
Solves sparse symmetric positive definite linear systems. These problems arise in many physical applications, like linear elasticity, heat transfer and other diffusion based transport phenomena.
This code implements the conjugate gradient method using a Jacobi preconditioner.
Install
npm install conjugate-gradient
Example
var pcg = require("conjugate-gradient")
, CSRMatrix = require("csr-matrix")
//Create a matrix
var A = CSRMatrix.fromDense([[-2, 1, 0],
[ 1,-2, 1],
[ 0, 1,-2]])
//Create input vector
var B = new Float64Array([1, 0, 0])
//Solve equation:
//
// A x = B
//
console.log(pcg(A, b))
require("conjugate-gradient")(A, b[, x0, tolerance, max_iter])
Solves the equation Ax = b by conjugate gradient
A
is a symmetric positive definite matrix represented as a CSRMatrixb
is an array of length nx0
is an optional initial guess for the solution to the equation. If specified, the result of the solution will also get stored in this arraytolerance
is a cutoff tolerance for the solution. (Default is 1e-5)max_iter
is the maximum number of iterations to run the solver. (Default is min(n, 20))
Returns An array encoding the solution to the equation Ax = b
Credits
(c) 2013 Mikola Lysenko. MIT License