[Scip] how to get non optimal solutions

Harzallah, Yosri h1rv3 at unb.ca
Tue Dec 2 14:23:30 MET 2008


Hi Timo,

Thanks for the response. 
Can the values for the time limit and the remaining primal-dual gap be set in a ".set"
file?

Yosri

Quoting berthold at zib.de:

> Hi Yosri.
> 
> In the binary shell of SCIP zou can set a variety of limits, e.g.
> SCIP> set limits time 3600
> for a time limit of one hour.
> You can also set limits on the solving nodes, the remaining primal-dual
> gap (!) and even the number of found solutions.
> Furthermore, when pressing CTRL-C SCIP will stop and return to the SCIP
> shell, being able to continue the search.
> In all cases, typing
> SCIP> display solution
> will give you the objective function value and the variable setting of the
> current best solution.
> 
> By now, however, it is not possible to see suboptimal solutions in the
> SCIP shell.
> 
> Best regards, Timo
> 
> > Dear All,
> >
> > I'm trying to solve an ILP with the binary distribution of SCIP. In some
> > cases, when it
> > takes hours to get the optimal solution, a solution with a gap of even 10%
> > is ok.
> > So I was wondering if there is a way to configure SCIP to output all the
> > solutions it
> > finds and not only the best one. Also is it possible to stop solving a
> > problem when a
> > solution is found with a gap less or equal to a certain %.
> > Thanks,
> >
> > Yosri Harzallah
> > _______________________________________________
> > Scip mailing list
> > Scip at zib.de
> > http://listserv.zib.de/mailman/listinfo/scip
> >
> 
> 


-- 
Yosri Harzallah

CS Master's Student - University of New Brunswick 
email: yosri.harzallah at unb.ca

Part-Time Student Employee - NRC-IIT Fredericton
email: yosri.harzallah at nrc.ca 


More information about the Scip mailing list