[Scip] "Hidden" cut generators

Timo Berthold berthold at zib.de
Thu Nov 22 18:25:47 MET 2012


Hi Giacomo,

so, I guess what you want to do is
SCIP> set sepa emphasis off
or
SCIP_CALL( SCIPsetSeparating(scip, SCIP_PARAMSETTING_OFF, TRUE) );
(also see docu for this and related methods in scip.h or at
http://scip.zib.de/doc/html/scip_8h.shtml)

Cheers
Timo

PS: The idea is that general cuts (like Gomory or clique cuts) that use
global structure (the LP tableau or the clique table) are implemented as
separators,
whereas constraint specific cuts (knapsack cover or gradient cuts for
convex MINLP) are implemented within the corresponding constraint handler.

Also see http://scip.zib.de/doc/html/FAQ.shtml#Q4.3


Am 22.11.2012 16:45, schrieb Giacomo Nannicini:
> Hi all,
> I noticed a strange behavior in SCIP that makes me think there are
> some "hidden" cut generators.
>
> I manually disabled *all* separators from scipdefplugin.c (I commented
> out all lines with SCIPincludeSepa). No cut generators are listed in
> SCIP when I use "set separating" or in the final "display statistics".
> However, some cuts *are* generated. In some logs, I get a nonzero
> number of cuts indicated, even though "display statistics" says that
> there are no separators. Cut pool size stays at zero. Also, the
> solution path is different from what I get if I set "set separating
> maxroundsroot 0". With the latter setting, no cuts appear.
>
> Can somebody enlighten me?
>
> Thanks a lot!
>
> Giacomo
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip



More information about the Scip mailing list