[Couenne] Problems with couenne

Stefan Vigerske stefan at math.hu-berlin.de
Wed Nov 6 06:52:28 EST 2013


Hi,

Couenne does not report a dual solution, but just a feasible (and maybe 
optimal) point, as it assumes a nonconvex problem, where a globally 
valid dual solution couldn't be expressed by simply associating scalar 
values with the constraints.

If you need marginal values, you should just add a second solve into 
your GAMS model that uses a local NLP solver (e.g., Ipopt or CONOPT). It 
should pick up the solution found by Couenne as starting point and 
report locally valid dual values.
If you have a MINLP, you would have to fix all discrete variables to the 
solution reported by Couenne before calling the NLP solver (using 
modeltype RMINLP).

Best,
Stefan

On 11/06/2013 12:30 PM, Taher kahil wrote:
> Hello:
>
> I'm a Ph.D student in economics. I'm using the couenne solver through GAMS
> software to solve a large scale river basin model, that includes a
> piece-wise linear function.
>
> My question is: I find an optimal solution to my problem, but I can't see
> the marginal values of the resources. These values are very important for
> my study. Is couenne solver not able to find marginal values due to its
> solution procedure or it could be a problem in my model structure?
>
> I appreciate a lot your help.
>
> Looking forward to hearing from you.
>
> Taher
>
>
>
> _______________________________________________
> Couenne mailing list
> Couenne at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/couenne
>



More information about the Couenne mailing list