[Scip] add constraints while solver is running?

Sandy Ryza sandyryza at gmail.com
Fri Mar 29 22:24:40 MET 2013


Thanks for the help Ambros,

Are you saying that I can use SCIPsetRealParam(scip, "limits/objective",
value) from a separate thread after I have started the solving?  Or does
your second email mean that to accomplish this I need to use an event
handler?

-Sandy

On Fri, Mar 29, 2013 at 1:28 PM, Ambros Gleixner <gleixner at zib.de> wrote:

> I just realized: Maybe your question was more about how to interrupt the
> solving process in order to communicate the new upper bound.
>
> For this, using the event handler system of SCIP is probably best, see
>
>     http://scip.zib.de/doc/html/EVENT.shtml
>
> Maybe check out how it's done in the code of our parallelization
> framework UG, which is included in the SCIP Optimization Suite.
>
> - Ambros
>
>
>
> Am 29.03.2013 21:23, schrieb Ambros Gleixner:
> > Hi Sandy,
> >
> > both local and global constraints can be created and added during the
> > solution process as usual via SCIPcreateCons...() and SCIPaddCons().
> >
> > However, you do not have to/should not add the upper bound as a
> > constraint, but can use the parameter "limits/objective"; this can be
> > changed at any time during the solution process via
> >
> >     SCIP_CALL( SCIPsetRealParam(scip, "limits/objective", value) );
> >
> > Also, you might consider transferring the solutions themselves via
> > SCIPtrySol() or SCIPaddSol() because this can improve the performance of
> > some primal heuristics; but then again the solution from one subproblem
> > may not be feasible for the others and hence be rejected.
> >
> > Hope this helps,
> >
> > Ambros
> >
> >
> >
> >
> > Am 29.03.2013 19:25, schrieb Sandy Ryza:
> >> Hi All,
> >>
> >> Is there an easy way in SCIP to add constraints while the solver is
> >> running?  My use case is that I want to try running multiple solvers in
> >> parallel on different subproblems, and if one of the solvers comes
> >> across a good solution I want to be able to set the upper bound on the
> >> other ones for better pruning.
> >>
> >> thanks for any guidance,
> >> Sandy
> >>
> >>
> >> _______________________________________________
> >> Scip mailing list
> >> Scip at zib.de
> >> http://listserv.zib.de/mailman/listinfo/scip
> >>
> >
>
> --
> ____________________________________________________________
> Ambros M. Gleixner
> Zuse Institute Berlin - Matheon - Berlin Mathematical School
> http://www.zib.de/gleixner
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listserv.zib.de/mailman/private/scip/attachments/20130329/744434a0/attachment.html


More information about the Scip mailing list