[Scip] about LPI

[Gerald Gamrath]

Hi Ahmad,

you should *not* ask for the LPI pointer like this and then mess around
in the LP. You should get the current LP, but you don't know which
variable in the LP solver corresponds to which SCIP variable, this is
all handled by SCIP and you should not try to do it yourself.

What do you want to do? If you want to solve some auxiliary LPs based on
the current LP, but with slight modifications, you should use the diving
or probing mode of SCIP, see the documentation starting at
http://scip.zib.de/doc/html/scip_8h.shtml#a61e0d3177ad999bf713b05bfa53c65ae
and
http://scip.zib.de/doc/html/scip_8h.shtml#a71a4ba1e820f5e875b940a801bcc1e8e

The diving mode just works on the current LP, you can change bounds and
the objective function. The probing mode creates temporary B&B nodes and
thus includes some overhead, but is more powerful as it allows to
perform domain propagation and pricing of new variables. However, it
does not allow to change the objective function.

Best,
Gerald


On 09.04.2013 15:48, Ahmad Moradi wrote:
> Dear all,
>
> Consider the following piece of code,
>
> SCIP_LPI* mylpi = NULL;
> SCIP_CALL( SCIPgetLPI(scip, &mylpi) );
>
> and suppose that the underlying lp solver is cplex, I just want to
> know, is the lp pointer given by mylpi->cpxlp is the lp of the current
> node of BB tree?
>
> Bests, Ahmad
>
>
>
> _______________________________________________
> 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/20130409/f8afa80b/attachment.html