You can also try Prof. Ye's Matlab software, it runs unmodified on
Octave (+ forge).
Visit http://www.stanford.edu/~yyye/ and follow the links.
It dates from his PhD back in the late 1980s, but I'm using it and it
works for me.
It uses numerical derivatives (forward differences).
Cheers,
Paul.
Eric S Fraga wrote:
I am looking for an algorithm that does constraint optimization. For
the time being I only require bounds and linear constraints, but it
must
be able to work with optimization for multivariate problems (e.g.
fmincon).
Maybe have a look at the Matlab functions described by Kelley:
@Book{kelley-1999,
author = {C. T. Kelley},
title = {Iterative methods for optimization},
publisher = {{SIAM}},
year = 1999
}
I believe he has made a number of functions available so do a web
search.
Alternatively, if your constraints are linear and if your objective
function
is well-defined outside the feasible region, incorporated the
constraints
directly into a new objective function as part of a penalty function
and use
a non-constrainted optimization method.
cheers,
eric
-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.
Octave's home on the web: http://www.octave.org
How to fund new projects: http://www.octave.org/funding.html
Subscription information: http://www.octave.org/archive.html
-------------------------------------------------------------