[Couenne] Couenne is declaring infeasibility of a feasible model

Pietro Belotti petr.7b6 at gmail.com
Sat Jan 4 08:33:32 EST 2014


Wendel,

it was a bug in bound tightening for auxiliary expressions defined as
y >= sqrt(x). It should be fixed now both in trunk and stable/0.4.

Regards,
Pietro

On Sat, Jan 4, 2014 at 10:44 AM, Wendel Melo <wendelalexandre at gmail.com> wrote:
> Hi everybody
>
>
> I have an AMPL model to solve (in attachment). I have submitted to Couenne,
> but it is declaring infeasiblity:
>
>
>> ampl model_mmx_.mod inst3_.dat    couenne
> Couenne 0.4.7 --  an Open-Source solver for Mixed Integer Nonlinear
> Optimization
> Mailing list: couenne at list.coin-or.org
> Instructions: http://www.coin-or.org/Couenne
> couenne:
> ANALYSIS TEST:  Couenne: Warning, tightened NLP is infeasible
> problem infeasible
> Error at _cmdno 2 executing "solve" command
> (file couenne.cmd, line 4, offset 28):
> can't open /tmp/at2590.sol
>
>
>
>
> However, I can get a feasible solution if I submit the the model to other
> solvers like Bonmin, for example. Is there some parameter or tolerance that
> I should set to remediate this? I have tried parameter "feas_tolerance", but
> I have not had success.
>
>
> Thanks in advanced
>
>
> --
> Wendel Melo
>
>
>
>
> _______________________________________________
> Couenne mailing list
> Couenne at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/couenne


More information about the Couenne mailing list