[Scip] Accessing primal variables in pricer

[Gerald Gamrath]

Hi Daniel,

probably the objective limit was reached during the LP solve with the 
dual simplex. Then, the LP is not solved to optimality and only the dual 
LP solution is feasible, which is sufficient for the pricing routine, 
because you only need the dual solution values.
You can check the LP solution status by calling SCIPgetLPSolstat(scip), 
the solution does only need to be primal feasible if it is 
SCIP_LPSOLSTAT_OPTIMAL (or SCIP_LPSOLSTAT_UNBOUNDEDRAY).

Best,
Gerald

Am 27.08.2012 16:40, schrieb Daniel Karch:
> Hi,
>
> I need to access the values of primal variables in a pricer, but a lot 
> of them are violating their bounds.
> When I print them with SCIPprintSol(scip, NULL, 0, 0) in the 
> scip_redcost callback, the output looks
> something like this:
>
> ...
> y_6#(2,0)_21#(2,3)                  -0.16988416988417 (obj:0)
> y_6#(2,0)_22#(2,4)                                  1 (obj:0)
> y_6#(2,0)_24#(2,6)                  -0.16988416988417 (obj:0)
> y_6#(2,0)_26#(2,8)                                  1 (obj:0)
> y_6#(2,0)_28#(2,10)                                 1 (obj:0)
> y_6#(2,0)_29#(2,11)                                 1 (obj:0)
> y_6#(2,0)_56#(7,8)                                  1 (obj:0)
> x_7#(2,1)_1                                         1 (obj:0)
> x_7#(2,1)_3                          1.33976833976834 (obj:0)
> x_7#(2,1)_6                          1.16988416988417 (obj:0)
> ...
>
> All of those variables should be between 0 and 1.
> Shouldn't the LP be valid in the scip_redcost callback?
>
> Best regards,
>
>   Daniel
>
>
> _______________________________________________
> 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/20120827/8960d54b/attachment.html