Mercurial > hg > octave-nkf
view scripts/control/util/zgfslv.m @ 5967:d542d9197839 ss-2-9-8
[project @ 2006-08-24 21:24:53 by jwe]
author | jwe |
---|---|
date | Thu, 24 Aug 2006 21:27:41 +0000 |
parents | 4c8a2e4e0717 |
children | 93c65f2a5668 |
line wrap: on
line source
## Copyright (C) 1996, 1998 Auburn University. All rights reserved. ## ## This file is part of Octave. ## ## Octave is free software; you can redistribute it and/or modify it ## under the terms of the GNU General Public License as published by the ## Free Software Foundation; either version 2, or (at your option) any ## later version. ## ## Octave is distributed in the hope that it will be useful, but WITHOUT ## ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or ## FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License ## for more details. ## ## You should have received a copy of the GNU General Public License ## along with Octave; see the file COPYING. If not, write to the Free ## Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA ## 02110-1301 USA. ## -*- texinfo -*- ## @deftypefn {Function File} {} zgfslv (@var{n}, @var{m}, @var{p}, @var{b}) ## Solve system of equations for dense zgep problem. ## @end deftypefn ## Author: A. S. Hodel <a.s.hodel@eng.auburn.edu> ## Converted to Octave by R Bruce Tenison, July 3, 1994 function x = zgfslv (n, m, p, b) nmp = n+m+p; gam1 = (2*n)+m+p; gam2 = n+p; gam3 = n+m; G1 = givens(sqrt(m),-sqrt(p))'; G2 = givens(m+p,sqrt(n*(m+p)))'; x = b; ## 1) U1 e^n = sqrt(n)e_1^n ## 2) U2 e^m = sqrt(m)e_1^m ## 3) U3 e^p = sqrt(p)e_1^p xdx1 = 1:n; xdx2 = n+(1:m); xdx3 = n+m+(1:p); x(xdx1,1) = zgshsr(x(xdx1,1)); x(xdx2,1) = zgshsr(x(xdx2,1)); x(xdx3,1) = zgshsr(x(xdx3,1)); ## 4) Givens rotations to reduce stray non-zero elements idx1 = [n+1,n+m+1]; idx2 = [1,n+1]; x(idx1) = G1'*x(idx1); x(idx2) = G2'*x(idx2); ## 6) Scale x, then back-transform to get x en = ones(n,1); em = ones(m,1); ep = ones(p,1); lam = [gam1*en;gam2*em;gam3*ep]; lam(1) = n+m+p; lam(n+1) = 1; # dummy value to avoid divide by zero lam(n+m+1)=n+m+p; x = x ./ lam; x(n+1) = 0; # minimum norm solution ## back transform now. x(idx2) = G2*x(idx2); x(idx1) = G1*x(idx1); x(xdx3,1) = zgshsr(x(xdx3,1)); x(xdx2,1) = zgshsr(x(xdx2,1)); x(xdx1,1) = zgshsr(x(xdx1,1)); endfunction