[Scip] Error when displaying output

[Stefan Heinz]

Hi Steve,

there is an error in the heuristic dins in case of column generation.
Could please try the following:

Add the line

SCIP_CALL(SCIPsetHeuristics(scip, SCIP_PARAMSETTING_AGGRESSIVE, TRUE));

back to your code and turn off the heuristic dins

SCIP_CALL( SCIPsetIntParam(scip, "heuristics/dins/freq", -1) );

The order of the two code lines have to be as shown above. Rerun you
program. Is the error still there?

Thanks Stefan

> Hi Michael,
>
> I have discovered a cause of the problem for me. I am using the
> following call for the heuristic settings.
>
> SCIP_CALL(SCIPsetHeuristics(scip, SCIP_PARAMSETTING_AGGRESSIVE, TRUE));
>
> When I remove this line from my code everything works correctly. So I am
> not sure whether there is an underlying problem that using the above
> command was bring out, or whether there is just a problem for me using
> that command.
>
> In addition to my problem that I sent through earlier, I was also
> getting a memory issue when adding coefficients to linear constraints in
> my pricer. This has also cleared up when I remove the above command.
>
> Thanks for your help,
>
> Steve
>
> Stephen J Maher wrote:
>> Hi Michael,
>>
>> The LP solver that I am using is Soplex and I have compiled SCIP using
>> gcc version 4.1.2. I am using SCIP as a callable library in my own
>> code written in C++. The branching rule that I am using is Ryan/Foster
>> branching that I have written also in C++.
>>
>> I am also using a variable pricer with my model to perform column
>> generation.
>>
>> I am sorry, but I couldn't get valgrind to work on my PC. I will have
>> to talk to IT about that one.
>>
>> Thanks for your help,
>>
>> Steve
>>
>> michael.winkler at zib.de wrote:
>>> Hi Steve,
>>>
>>> I haven't seen your problem till now. But maybe you can provide us/me
>>> with
>>> some more details about your crash. Which LP solver did you use and
>>> which
>>> compiler version under what system did you compile SCIP? Is it SCIP
>>> itself
>>> or have you written your own code? What kind of branching rule have you
>>> applied?
>>>
>>> It looks like you already did a gdb run, to determine the error. Maybe
>>> you
>>> can try a valgrind run as well, to see whether you can trust the gdb
>>> output.
>>> (e.g. valgrind -v scip -f filename)
>>>
>>> Best, Michael
>>>
>>>
>>>> Hi everyone,
>>>>
>>>> I am having some trouble at the moment with SCIP producing a
>>>> segmentation fault at line 419 of disp.c. When it tries to access
>>>> set->disps[i] is seems to not be pointing to any memory address.
>>>>
>>>> This only happens directly after the model applies a branching rule.
>>>> Also it only occurs when I compile the code with OPT=opt-gccold, it
>>>> works fine when I use OPT=dbg.
>>>>
>>>> I am using SCIP 2.0.
>>>>
>>>> Has anyone seen any problems like this before?
>>>>
>>>> Cheers,
>>>>
>>>> Steve
>>>> _______________________________________________
>>>> Scip mailing list
>>>> Scip at zib.de
>>>> http://listserv.zib.de/mailman/listinfo/scip
>>>>
>>>>
>>>
>>>
>>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Scip mailing list
>> Scip at zib.de
>> http://listserv.zib.de/mailman/listinfo/scip
>>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> http://listserv.zib.de/mailman/listinfo/scip
>