This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: Powell's method?
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: hersh at ri dot cmu dot edu
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Sun, 7 Apr 2002 15:49:00 +0100 (BST)
- Subject: Re: Powell's method?
- References: <15534.2866.343761.812104@gs220.sp.cs.cmu.edu>
hersh@ri.cmu.edu writes:
> I'm in need of a multidimensional minimizer which does not require
> derivatives. The GSL does not seem to include any such, but the
> Numerical Recipes in C book does have some, Amoeba (simplex) and
> Powell's method. Is anyone currently working on including such
> algorithms in GSL, or do you know of some other free place to get
> them?
I don't know of anybody working on that, or of any existing free
implementation (although I personally haven't spent much time looking
for one).
> Otherwise, I'd like to write at least Powell's from scratch, and maybe
> it can be included in the GSL. I realize I can't use the NRC source
> code of course, I would be re-writing it from scratch.
Sounds good. See the GSL design document on
http://sources.redhat.com/gsl for tips.
Brian