[Cbc] Fwd: Gomory or not Gomory, that is the question.

[Giacomo Nannicini]

John,
as Yves mentioned, twomir also seems to give some trouble. In these
situations it's hard to put the blame on a single generator: most likely,
it is a combination of things. So I'm not too surprised that swapping GMI
for Gomory is not enough.

But I'm glad to hear that using CglGMI only seems to work on this problem
and does not give the same numerical troubles!

Giacomo


On Wed, Mar 4, 2015 at 9:47 PM, John Forrest <john.forrest at fastercoin.com>
wrote:

>  Giacomo,
>
>
> I tried swapping GMI for Gomory and the same thing happened.
>
> It just seems to be errors building up, especially with general integer
> variables.
>
> John
>
> On 04/03/15 13:30, Giacomo Nannicini wrote:
>
> Hi Yves,
> did you try the CglGMI generator? You may want to see if using CglGMI
> instead of CglGomory helps. In my experience, the cuts from CglGMI are less
> prone to numerical errors: in fact, that's the reason the generator was
> developed. See
> http://link.springer.com/article/10.1007%2Fs12532-013-0057-4
>
>  Giacomo
>
> On Wed, Mar 4, 2015 at 7:06 PM, Yves Touchard <ytouch at dxo.com> wrote:
>
>>  Hello,
>>
>> Switching off the twomir cuts makes also the problem feasible.
>>
>> Cheers
>>
>> Yves
>>
>>
>> -------- Message transféré --------  Sujet : Gomory or not Gomory, that
>> is the question.  Date : Tue, 03 Mar 2015 20:03:03 +0100  De : Yves
>> Touchard <yves.touchard at sfr.fr> <yves.touchard at sfr.fr>  Pour :
>> cbc at list.coin-or.org
>>
>>  Hello,
>>
>> You will find here (
>> https://www.dropbox.com/s/ayo93hi4togpc9m/noGom.lp?dl=0) a constraint
>> file for which cbc founds a solution when Gomory cuts are removed or for
>> specific cutlength values.
>>
>> In other words, the following command leads to an infeasible problem:
>> */soft/cbc-trunk-2149/bin/cbc noGom.lp branch printingOptions normal
>> solution noGom.sol*
>>
>> and with the following commands, cbc claims that there is a solution:
>> */soft/cbc-trunk-2149/bin/cbc noGom.lp gomory off branch printingOptions
>> normal solution noGom.sol*
>> */soft/cbc-trunk-2149/bin/cbc noGom.lp cutlength xx branch
>> printingOptions normal solution noGom.sol* (with xx equal to -1, 0, 10,
>> 20... and different from 30...)
>>
>> Since our problems are dynamically generated, it is very painful to try
>> to solve them with different options, hoping that the right set is in the
>> list.
>>
>> So, is there a way to find this set of options.
>> Or could you explain what are the rows that are difficult for cbc.
>> And, does the change of the continuous variables into integer variables
>> help? (for our problems, it makes sense).
>>
>> Thanks and regards
>>
>> Yves
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> This message and any attachments (the "message") are confidential and
>> intended solely for the addressee(s). Any unauthorized use or dissemination
>> is prohibited. E-mails are susceptible to alteration. Neither DxO Labs nor
>> any of its subsidiaries or affiliates shall be liable for the message if
>> altered, changed or falsified.
>> Ce message et toutes les pièces jointes (ci-après le "message") sont
>> confidentiels et établis à l'intention exclusive de ses destinataires.
>> Toute utilisation ou diffusion non autorisé est interdite. Tout message
>> électronique est susceptible d'altération. DxO Labs et ses filiales
>> déclinent toute responsabilité au titre de ce message s'il a été altéré,
>> modifié ou falsifié.
>>
>> _______________________________________________
>> Cbc mailing list
>> Cbc at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/cbc
>>
>>
>
>
> _______________________________________________
> Cbc mailing listCbc at list.coin-or.orghttp://list.coin-or.org/mailman/listinfo/cbc
>
>
>
> _______________________________________________
> Cbc mailing list
> Cbc at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cbc
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20150305/0367a6d9/attachment.html>