Re: Software suggestions

cmj917 wrote:
There's another freely available solver written in Java - jOpt.

Check it out at http://jopt.sourceforge.net/.

I'm one of the project admins and would be happy to answer any set-up
or configuration questions. It might be just what you're looking
for. Regardless, don't hesitate to contact me via e-mail if you have
any questions.

Thanks, and good luck.

Chris


Chris,

jOpt is in fact the first option I've tried (and the only one I've downloaded so far). I was attracted to it both because it's a Java solver and because I've got licenses for OPL Studio and CPLEX (and actually one for an older version of ILOG's Solver ... but I would need to return to C++ to use it, and I'm growing allergic to C/C++ in my old age). I ran the examples, but I was a bit concerned by the one that finds sets x, y, z in {0,1,2} such that x and y are disjoint and z is their union. It reported 109 ms to find all 27 solutions. I don't have any experience whatsoever with CP so far, but that seemed to be more time than I would have expected, and I don't know how much was "fixed cost" and how much was "variable cost". When I get a moment, I'm going to play around with boosting the domains and see how the run time grows with problem size. Meanwhile, if you have any performance benchmarks posted somewhere, I'd love to see them.

Incidentally, while I was on the project web site, I noticed that you're also planning a math programming module. I was curious whether your intent was to build one from scratch or create the plumbing to link to an existing one (perhaps lp_solve, which is released under the LPGL)?

Cheers,
Paul
.