This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: High-dimensional Minimization without analytical derivatives
- From: Joakim Hove <hove at ift dot uib dot no>
- To: "Anatoliy Belaygorod" <belaygorod at wustl dot edu>
- Cc: <gsl-discuss at sources dot redhat dot com>
- Date: Sat, 04 Sep 2004 10:33:18 +0200
- Subject: Re: High-dimensional Minimization without analytical derivatives
- References: <EB6AEAD6EC432548BCCE3453DCC2818F151996@wub-mail.olin.wustl.edu>
"Anatoliy Belaygorod" <belaygorod@wustl.edu> writes:
> My understanding is that 'in general' in high-dimensional cases with
> rough surface, the Simulated Annealing (SA) method is better tuned
> for finding a GLOBAL maximum , than Gradient-based methods, because
> the latter are better tuned for 'zeroing in' the local maximums. In
> that regard, is Simplex Method closer to SA, or Gradient-based
> methods?
Well, excuse me if I am completely off base, but as far as I am aware
the simplex method is restricted to *linear* problems - where it is
'guaranteed' to find the optimial solution. Gradient based methods and
SA can be used for more general (nonlinear) problems, but can really
not be compared to the two.
JH
--
Joakim Hove
hove AT ift uib no
+47 (55 5)8 27 90
http://www.ift.uib.no/~hove/