[Scip] Slow LP re-optimization after branching

[Xavier Schepler]

Please find attached to this mail 3 logs.
Two of them are complete logs for 2 instances.
The third one is a log without SoPlex output, where the message "LP solver
hit time limit" appears several times.
CG iteration #1 indicates a new node.

Best regards,

Xavier Schepler
PhD student,
University of Le Havre


2014-02-21 10:25 GMT+01:00 Xavier Schepler <xavier.schepler at gmail.com>:

> Thanks Dr. Gerald for your quick response.
>
> I will send you a complete log today.
>
>
> 2014-02-21 10:09 GMT+01:00 Gerald Gamrath <gamrath at zib.de>:
>
>  Dear Xavier,
>>
>> if you don't set a timelimit, then the LP solver should normally not have
>> a time limit either which it can hit. Could you send me the complete log?
>>
>> A few minutes for LP solving might happen, especially if there are some
>> numerical troubles. Anyway, it is definitely not what one would hope for,
>> especially not for a branch-and-price approach.
>>
>> You can set "display/lpinfo" to TRUE in order to see the SoPlex output
>> and "display/verblevel" to 5 to get a bit more output, in particular about
>> numerical problems.
>>
>> We are currently finishing the next SCIP release and hope to get it done
>> today. I would suggest you switch to that when it is available, because
>> there has been some improvement for branch-and-price in there, in
>> particular for the master problem reoptimization.
>>
>> Best,
>> Gerald
>>
>> Am 21.02.2014 09:55, schrieb Xavier Schepler:
>>
>>  Hi professor Lübbecke,
>>
>> The B&P is working fine on small instances.
>>  For some bigger instances, most of the computing time is spent in
>> solving linear relaxations of restricted master problems.
>>
>>   Can you plot the objective function values (over iterations) of the
>> restricted master problem (to check whether you have strong tailing off)?
>>
>>   Did you check dual variable development (you can plot this as well) to
>> check whether you have unstable duals?
>>
>>  Tailing off is a real problem, since an optimal value to the linear
>> relaxation of a master problem (v_mp) is attained in a few column
>> generation iterations,
>> but, the lagrangian bound is very far away, and needs many more
>> iterations to meet v_mp .
>>
>>    However, you could tell us about your branching rule. Branch on
>> master vars (uah...) or on other information (like "original" vars)?
>> Ryan-Foster?
>>
>> Branching is done on original variables, and branching constraints are
>> added to the structural constraints,
>> as it is described in Gerald Gamrath thesis.
>>
>>   Give us a little more detail. [BTW, you fix the *local* UBs of the
>> vars, right?]
>>
>>   Well, I fix local upper bounds.
>>
>>   Out of curiosity: what is "time limit"
>>
>> I don't know exactly what the time limit is, but SCIP outputs :
>> LP solver hit time limit.
>>
>>   and "a very long time?"
>>
>> Maybe several minutes, which seems quite long for LP re-optimization.
>>
>>  I would like SoPlex to display information ...
>>
>>  Best regards,
>>
>> Xavier
>>
>>
>> _______________________________________________
>> Scip mailing listScip at zib.dehttp://listserv.zib.de/mailman/listinfo/scip
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20140221/c15f0708/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bp logs.tar.gz
Type: application/x-gzip
Size: 173789 bytes
Desc: not available
URL: <http://listserv.zib.de/pipermail/scip/attachments/20140221/c15f0708/attachment.bin>