[Bonmin] Bonmin (1.3.2) bug

[Stefan Vigerske]

Hi,

try Bonmin 1.4.
I think there were fixes in this part of the code a while ago.

Stefan

> Hi,
>
> I would like to report a bug encountered in Bonmin (stable release 1.3.2) and was
> wondering if others are aware of this, and if it has in fact been fixed in later versions?
>
>
> For a simple 2v problem:
>
> F(X) = (x1-8)^2 + (x2-2)^2
>
> with 0<xi<10 (i=1,2)   and   Xstart = [5 5]
>
> If X is assumed continuous, the solution [8,2] with F = 0 is obtained trivially.
>
> However, if X is integer, Bonmin crashes at the following call:
>
> CoinIotaN(&sortedIndex[0], nnz_jac_g, 0);
>
> in BonminHeuristicDiveMIP with a 'vector subscript out of range' error.
>
>
> Thus, the problem concerns the use of integer variables and a zero value for nnz_jac_g (the
> number of nonzeros in the jacobian of the constraints) when no constraints are applied. If any
> one constraint is added, this problem does not appear.
>
> Is there a known fix to this problem?
>
> Thanks and Regards,
> Kashif
>
>
>
>
> _______________________________________________
> Bonmin mailing list
> Bonmin at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/bonmin


-- 
Stefan Vigerske
Humboldt University Berlin, Numerical Mathematics
http://www.math.hu-berlin.de/~stefan