Linear Algebra

Methods

LU Decomposition

GSL_Matrix#LU_decomp

This method factorizes the square matrix into the LU decomposition. This method returns an array of three elements. The first is the matrix of the LU decomposition, the second is a GSL_permutation object. The third is the sign of the permutation.

GSL_Matrix#LU_solve(perm, b)

This method solves the system A x = b using the LU decomposition of A given by LU_decomp and the permutation perm. The returned value is a GSL_vector object which contains the solution vector x. The following is an example to solve a linear system

A x = b, b = (1, 2, 3, 4)

with LU_decomp and LU_solve,

ex.)
 (require 'narray')         <--- if Ruby/GSL is compiled with --with-narray flag
  require 'gsl'

  m = GSL_Matrix.new([0.18, 0.60, 0.57, 0.96], [0.41, 0.24, 0.99, 0.58],
                     [0.14, 0.30, 0.97, 0.66], [0.51, 0.13, 0.19, 0.85])
  lu, p = m.LU_decomp                <- m is decomposed into lu
  b = [1, 2, 3, 4]
  x = lu.LU_solve(p, b.to_v)         <- "b" must be given as a GSL_Vector
  p x.to_a                           <- The solution is returned by a vector.

GSL_Matrix#LU_invert(perm)

This method computes and returns the inverse of a matrix from its LU decomposition. This requires the permutation perm given by LU_decomp.

ex.)
  lu, p, signum = m.LU_decomp
  inv = lu.LU_invert(p)
  inv.print

GSL_Matrix#LU_det(sugnum)

This method computes and returns the determinant of a matrix from its LU decomposition. This requires the value of signum given by LU_decomp.

ex.)
  lu, p, signum = m.LU_decomp
  p lu.LU_det(signum)

Singular Value Decomposition

GSL_Matrix#SV_decomp

This function factorizes the M-by-N matrix into the singular value decomposition A = U S V^T (A: reciever). This method returns an array of three elements. The first is a GSL_Matrix object of a matrix U. The second is the matrix V, and the last is a GSL_vector which corresponds to S.

SV_solve

This method solves a linear system using the SV decomposition.

ex.)
  m = GSL_Matrix.new([1, 2, 3], [6, 5, 4], [7, 8, 1])
  U, V, S = m.SV_decomp
  x = gsl_linalg_SV_solve(U, V, S, [2, 3, 4].to_v)
  x.print

GSL_Matrix#SVD_solve(b)

This method computes the SV decomposition of the reciever matrix, and solve the linear system with a given vector b.

ex.)
  m = GSL_Matrix([1, 2, 3], [6, 5, 4], [7, 8, 1])
  b = [2, 3, 4].to_v
  x = m.SVD_solve(b)
  x.print