[SCIP] Separators statistic

[Benjamin Müller]

Dear Abbas, dear Tony,

probably this is just a bug in SCIP that is related to the statistics of 
a separator. Abbas, could you send me the instance so that I can 
reproduce this behavior? Thank you!

Best,
Benjamin

On 8/10/19 1:56 AM, Wang, Tony Congqian wrote:
> 
> Dear abbas omidi,
> 
> You might want to take a look at 
> https://scip.zib.de/doc/html/scip__solvingstats_8c_source.php#l02821.
> 
> The total number of applied can be found in general statistics or in the 
> solving output.
> 
> Regarding the Seprators statistic table: Calls keeps track of how often 
> a Separator has been called, Cutoffs keeps track of how often a node has 
> been proved to be infeasible or to be having a lower dual bound than the 
> current primal bound by a Separator, DomReds are essentially the number 
> of singleton cuts found.
> 
> Now what is weird though, is that Cuts is supposed to be the number of 
> Cuts found by a Separator and Applied keeps track of how many Cuts, 
> found by a Separator have been applied. Now as it can be seen in this 
> picture, sometimes Applied for a given Separator is positive, although 
> Cuts is zero.
> 
> This is something I was also wondering about and I would be interested 
> in knowing the reason for this behavior. Is there simply some bug, or is 
> there some intention, that I missed?
> 
> Regards,
> Tony
> 
> ------------------------------------------------------------------------
> *Von:* Scip <scip-bounces at zib.de> im Auftrag von abbas omidi 
> <abb.omidi at gmail.com>
> *Gesendet:* Samstag, 10. August 2019 00:19
> *An:* scip at zib.de
> *Betreff:* [SCIP] Separators statistic
> Dear community team,
> 
> I'm trying to solve a MIP just using separating options. I turn off 
> Emphasis settings for presolving and heuristics. When I solve the model, 
> I get some results in the following:
> 
> Separators.jpg
>   I was wondering if,
> a) what do "separating columns" mean? (calls, cutoffs, ...)
> b) how can I find the number of Cutting Planes which be applied in the 
> model?
> c) what is the difference between calls, cuts and applied columns?
> 
> Regards
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip
> 

-- 
______________________________
Benjamin Müller
Zuse Institute Berlin
Takustr. 7, 14195 Berlin
benjamin.mueller at zib.de
+49 30 841 85-195