[SCIP] checking gaplimit before branching

[Jakob Witzig]

Hi Marko,

can you explain the meaning of the 6th, 7th, and 8th column? I think the 
last column is the gap?! Are you sure, that you have an feasible 
solution, because you have an infinite gap which means (i) no primal 
solution exists or (ii) primal or dual bound is equal to 0.

Just as a remark, the gap set with the parameter "limits/gap" is 
relative to 1, e.g., 1 corresponds to 100%.

Best,
Jakob

On 25.09.2015 12:21, Horváth Markó wrote:
> Dear all,
>
> I'm using SCIP-3.2.0 and I'm dealing with a branch-and-cut procedure, 
> and I set gaplimit to 0.005 (that is, "limits/gap" was set to 0.005). 
> Here is my output:
>
> /> (...)/
> /> 35.0    1  38r    663    663    6913.000      -- 6924.000    Inf//
> //> (pricing) no new column(s)//
> //> (general event handler) [lpsolved] LP of node 1 was completely 
> solved with cut & price//
> //> (branching) branching at node 1//
> //> (branching) branching node: 1 same node: 2 diff node: 3//
> //> (general event handler) [nodebranched] the focus node 1 has been 
> solved by branching//
> //> problem has been solved with gap 0.0016/
> /> (...)
> /
> In this output you can see that after some iterations my pricer could 
> not find new columns (hence my pricer sets result to RESULT_SUCCESS), 
> and SCIP says that the LP of node 1 was completely solved with cut & 
> price/.
> /So the root lowerbound is equal to 6913.000 and we also have a primal 
> solution with value 6924.000. So we have reached the gaplimit at the 
> root node, but as you can see SCIP didn't terminate but performed a 
> branching (actually using my branching rule). It's not a big deal, but 
> I wonder if we could avoid this unnecessary step (i.e., we could check 
> gap limit after a node is solved with cut & price).
>
> Best regards,
> Marko
>
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip
-- 
Jakob Witzig

Konrad-Zuse-Zentrum für
Informationstechnik Berlin (ZIB)

Division Mathematical Optimization and Scientific Information
Research Group Mathematical Optimization Methods

Takustrasse 7
14195 Berlin

Tel. : +49 (0)30 84185-416
Fax  : +49 (0)30 84185-269
email: witzig at zib.de
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20150925/e82ae6b4/attachment.html>