[Scip] Pricing generates an already existing column
Martin Bergner
bergner at or.rwth-aachen.de
Mon May 19 15:35:58 CEST 2014
Hi Cristina, (cc Marco, since that was the same question)
> I'm working on a B&Price algorithm for a minimization MIP and I realized
> that sometimes the algorithm generates already existing columns.
whenever I read that, my experience is that one of the following hold
(in order of appearance from my experience.
1. Your reduced cost computation is wrong. In this case, check it
again. And then again, because usually, that's really the case.
2. The check for variables with negative reduced costs is wrong:
Do you use an floating point check for the reduced costs, i.e.,
if( SCIPisNegative(scip, redcost) ) instead of if( redcost < 0? )
Additionally, it might be the following:
3. You use CPLEX as an LP solver
With CPLEX 12.2 and 12.2, I noticed this behaviour. The callable
library returned different reduced cost than reading in the master
problem in the CPLEX command line version. The dual values from
CPLEX as standalone product would be correct, that is no variable
would price out favorably whereas the dual values from the API lead
to new priced variables. You might want to try a different LP solver
in this case. For me, CPLEX stopped returning these useless dual
variables after some iterations, and since I had a set partitioning
master, I didn't care so much about this issue anymore since in an
optimal solution, the problem would go away anyway ... .
I hope I could give you another hint.
Regards,
Martin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4514 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://listserv.zib.de/pipermail/scip/attachments/20140519/0393050b/attachment.p7s>
More information about the Scip
mailing list