[Scip] Why is SCIP not applying my cuts?

[Stefan Heinz]

Hi Daniel,

it is most like that there is a "technical" issue. Meaning that your 
cuts are applied but not counted. Therefore, we first need to check this.

Which method are you using to create a row. You should use one of the 
following:

SCIPcreateRowCons()
http://scip.zib.de/doc/html/scip_8h.shtml#a8f71263b6e58cb5182ff009720342651

SCIPcreateEmptyRowCons()
http://scip.zib.de/doc/html/scip_8h.shtml#a41ae66e0ec0e8b636a3185825092f39f

Note that SCIPcreateEmptyRow() and SCIPcreateRow() are deprecated since 
the 3.0.0 release.
See documentation:
http://scip.zib.de/doc/html/scip_8h.shtml#a0f3f5d2982142da67c59fb8de0ccbd9c
http://scip.zib.de/doc/html/scip_8h.shtml#ad250ffa5f36870fc0eb2c10e1e2eaa67

Using one of the "new" methods and passing your constraint handler as 
third argument allows for counting the applied cuts. This actually is 
the reason why we changed the interface for the row creation. With the 
new methods we are now able to print the statistics which type of cuts 
are applied in the end (at least once). Therefore, we (the SCIP core) 
needs to know to which separator or constraint handler a cut belongs.

Are you using the "new" methods?

Best Stefan



On 02/01/2013 04:54 PM, Daniel Karch wrote:
> Hi,
>
> I have written a constraint handler for precedence constrained knapsack
> constraints.
> When a fractional solution is available, the constraint handler looks for
> violated induced cover
> inequalities, and separates them as cuts.
> However, the cuts are not applied. When the problem is solved, the output
> shows that
> e.g. 60 cuts have been found, but 0 applied.
> Why is that? I am making sure that the inequalities are in fact violated by
> the LP solution,
> but the cuts are not applied ...
>
> Best regards,
>
>    Daniel
>
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip