[Couenne] incorrect solution

[Pietro Belotti]

Hi Francois,

I can replicate the problem, which seems to be related x_73 and x_74  
being redundant variables. Each of them is equal to another variable  
(x_58 and x_44 respectively), so Couenne ignores them during  
optimization and sets them to zero. I'll fix this in the next few  
days. The value -13 and 1 of x_58 and x_44 are indeed correct for the  
bounds on x_73 and x_74.

Best,
Pietro
_________________________________________
Pietro Belotti, Lehigh University
Dept. of Industrial & Systems Engineering
200 W Packer Ave, Bethlehem PA 18015.
phone: 610-758-3865   fax: 610-758-4886
email: belotti at lehigh.edu
web:   http://www.lehigh.edu/~pib208


On 04/06/2009, François Dionne <brainstorm at videotron.ca> wrote:

> Dear mailing list,
>
> I downloaded the stable branch of Couenne by SVN on 31 march 2009,
> compiled it with g++ 4.2.4 and ran it on Ubuntu Linux Hardy Heron.
> I wrote my model in Ampl and generated some test data.
>
> Here's the problem: when I solve the problem, Couenne reports an optimal
> solution but when I read the value of the variables, one of the integer
> variables is given a value below its lower bound, in violation of the
> problem's constraints.
>
> You will find a stub named mymodel.nl attached to this message.
>
> Goodbye, have a nice day.
>
>



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.