[Dip] Dippy says problem is infeasible

[Ted Ralphs]

Hi Romain,

I'm looking into this now. I noticed you are using a very old version of
DIP. There have been many improvements and some bug fixes since that
version, so the first thing to do is update to the latest version of Dip
(0.92.3) and DipPy (1.92.3). With the latest version, the instance you sent
is no longer declared infeasible, but I do get the wrong solution when
using Cbc as the MILP solver, so there's some kind of problem there. I
guess it's a bug in Cbc. That might not be easy to find.

Using SYMPHONY as the solver, I get the same correct solution in both cases
(option "DecompIPSolver SYMPHONY" in latest version), so you could try that
for now, although the performance might not be as good. You can also use
CPLEX and Gurobi, although I noticed just now that I'm getting a seg fault
with those for some reason. I guess that will be easy to fix. I'll keep
poking at it.

Cheers,

Ted

On Thu, Feb 16, 2017 at 11:24 PM, Romain Montagné <
romain.montagne at polymtl.ca> wrote:

> I am using DipPy version 1.9.8. I will gladly share my scripts with you.
> All you have to do is adjust your path at the top of the RR_dippy.py so
> that it locates the file of Instances.
>
> Also, in the script I use the NetworkX module, you might have to download
> it.
>
> Let me know if I forgot something.
>
> Thanks!
>
>
>
>
> Le 16 févr. 2017 à 16:59, Ted Ralphs <ted at lehigh.edu> a écrit :
>
> Hi Romain,
>
> What's the version of DipPy you are using? How are you reading the problem
> into DipPy? Are you able to share the Python script you're using?
>
> Cheers,
>
> Ted
>
>
>
> On Thu, Feb 16, 2017 at 2:04 PM, Romain Montagné <romain.montagne@
> polymtl.ca> wrote:
>
>> Hello Ted,
>>
>> I am using DIP as a generic solver. I have simply created a
>> dippy.DipProblem that I solve with and without the doPriceCut option, like
>> in the examples that are provided.
>>
>> I have attached 4 files:
>>
>> - LP.lp is the .lp file
>> - LP_blocks.txt is the block file
>> - DoPriceCut0_terminal_output.txt is what I see on my screen when I
>> solve the problem without the doPriceCut option (and it works fine)
>> - DoPriceCut1_terminal_output.txt is what I see on my screen when I
>> solve it with the doPriceCut option (and it tells me problem infeasible)
>>
>> The version of DIP that I am using is 1.4.8. I work on a linux 64 bit
>> machine, and my compiler is gcc, version 4.4.7.
>>
>> If you need other information please let me know.
>>
>> Thank you very much for your help,
>>
>> Romain
>>
>>
>>
>>
>>
>>
>>
>>
>> Le 16 févr. 2017 à 11:38, Ted Ralphs <ted at lehigh.edu> a écrit :
>>
>> Hi Romain,
>>
>> Are you using DIP as a generic solver or have you built an application?
>> Can you send the .lp file and the block file, as well as other information,
>> such as version number, platform, compiler, etc.? I would like to try and
>> replicate this.
>>
>> Cheers,
>>
>> Ted
>>
>> On Thu, Feb 9, 2017 at 6:36 PM, Romain Montagné <romain.montagne@
>> polymtl.ca> wrote:
>>
>>> PS: After saying the problem is infeasible, I get the following error:  "Segmentation
>>> fault (core dumped)"
>>>
>>>
>>> Le 9 févr. 2017 à 17:45, Romain Montagné <romain.montagne at polymtl.ca> a
>>> écrit :
>>>
>>> Hello DIP community,
>>>
>>> I am using Dippy to solve a MILP. I want to compare  branch-and-bound
>>> and branch-and-price performances.
>>>
>>> When I solve the problem with branch-and-bound (i.e., with the option
>>> ‘doPriceCut’ set to 0), it works great. But when I activate the
>>> ‘doPriceCut’ option for branch-and-price, it tells me the problem is
>>> infeasible.
>>>
>>> Does anyone know what is happening?
>>>
>>> I have looked at the .lp file as well as the block file and they look ok.
>>>
>>> Any help is greatly appreciated.
>>>
>>> Cheers,
>>>
>>> Romain Montagné, PhD
>>>
>>>
>>>
>>> _______________________________________________
>>> Dip mailing list
>>> Dip at list.coin-or.org
>>> https://urldefense.proofpoint.com/v2/url?u=http-3A__list.coi
>>> n-2Dor.org_mailman_listinfo_dip&d=CwICAg&c=Ngd-ta5yRYsqeUsED
>>> gxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=rnUB298ByR5BGgo6vv2b6g&m=SQ6Ey
>>> INjMS0JVGwJJandNupwBLGNKe61En7rTWKugOY&s=-XIQIHlDQ8zaq0_-XbJ
>>> 7Y9UA_fut5QgegRRy2uMeO1k&e=
>>>
>>
>>
>>
>> --
>> Dr. Ted Ralphs
>> Visiting Researcher, Zuse Institute Berlin
>> Professor, Lehigh University
>> (610) 628-1280
>> ted 'at' lehigh 'dot' edu
>> coral.ie.lehigh.edu/~ted
>>
>>
>>
>>
>
>
> --
> Dr. Ted Ralphs
> Visiting Researcher, Zuse Institute Berlin
> Professor, Lehigh University
> (610) 628-1280
> ted 'at' lehigh 'dot' edu
> coral.ie.lehigh.edu/~ted
>
>
>
>


-- 
Dr. Ted Ralphs
Visiting Researcher, Zuse Institute Berlin
Professor, Lehigh University
(610) 628-1280
ted 'at' lehigh 'dot' edu
coral.ie.lehigh.edu/~ted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/dip/attachments/20170227/1ce93207/attachment.html>