On Tue, Jun 8, 2010 at 3:16 PM, Olaf Till <[email protected]> wrote:
> Hi,
>
> a function 'optim_problems' has been submitted to the optim package,
> returning test problems for optimization. As yet it's a skeleton with
> a few examples for curve fitting.
>
> I think there is a need for such a repository of tests for assessing
> and comparing optimizers.
>
> TODO:
>
> 1. Really "difficult" problems, preferably with literature references.
>
> 2. General minimization and zero finding.
>
> 3. Good tests for constraints, linear and non-linear, equality and
>   inequality, multiple constraints.
>
> Feel free to add tests. Also, I would be grateful for pointers to
> known test problems.
>
> Olaf
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Octave-dev mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/octave-dev
>

I think that having a test suite that can be run using different
solvers would be a very good idea. When I first developed bfgsmin and
samin, I used the Moré-Garbow-Hillstrom test suite. A link that has
the materials is http://pareto.uab.es/mcreel/MINTOOLKIT/ There is code
to run the entire test suite, using old versions of bfgsmin and samin.
I haven't used this in a long time, and I don't know if it will work
with current Octave. It could certainly be adapted pretty easily,
though.
Michael

------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate 
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
lucky parental unit.  See the prize list and enter to win: 
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev

Reply via email to