[SCIP] Slow computation time for SCIP compared to fmincon

[Marc Pfetsch]

Hi Greg,

as you write: fmincon searches for a local optimium, but SCIP for a
global optimum. The latter is much harder.

If SCIP takes a long time to find a good primal solution you might want
to supply it with the solution of fmincon. Then SCIP will either prove
it to be globally optimal or find a better one.

Best

Marc


On 18/03/2021 22.12, Gregory R Hendrickson wrote:
> Hello,
> 
> I am using SCIP to solve a nonlinear program and compare the computation
> time to fmincon. SCIP is able to find a solution with a gap of 10% in
> about 30 minutes whereas fmincon finds a local minimum in a few seconds.
> Is there any reason why there would be such a significant difference?
> 
> I attached my script in a zip file if you want to take a look, it's
> written in YALMIP if that's okay.
> 
> Regards,
> 
> *Greg Hendrickson*
> 
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip
>