[SCIP] Sub-optimal solution produced after upgrade to SCIP version 8.0.3

[Lisa-Marie Manke]

Hi, I was also running into a similar problem after upgrading to scip 8.0.3.

I am using benders decomposition to solver rather large instances. After 
I updated scip 7 to scip 8.0.3 some infeasible instances are getting 
solved to "optimality" but when checking the proposed solution it was 
obviously not even feasible for the solved instance.

Attached you can find an example instance on which this problem 
occurred. The orig.lp is the non decomposed formulation with whom the 
infeasibility is correctly detected, and the master.lp and sub__.lp 
files contain the respective parts of the decomposed formulation. Also 
attached are some log files from the solution process.

This could either be due to vast numerical issues or a bug within the 
scip code, but this didn't happened when using scip 7.0.3.

Regards,

Lisa-Marie Manke

On 24.04.23 15:06, Barry Hurley wrote:
> Hi, we were looking at upgrading from SCIP version 7 to version 8.0.3 
> but one of our unit-tests is catching a regression where SCIP appears 
> to be producing a sub-optimal solution and claiming optimality. We are 
> using SCIP via the LP file interface, the files to reproduce are 
> attached. problem.lp is the original problem, and 
> problem-forced-solution.lp extends this with an additional constraint 
> called `FORCED_SOLUTION` to force the optimal solution. Also including 
> the parameter settings used.
>
> The solution should have an optimal objective value of 115, but SCIP 
> version 8.0.3 is incorrectly producing a solution with objective value 
> 120, even with gap zero. The solution is valid, just sub-optimal. 
> There is one difference in the assignment, binary variable X_2__1_ has 
> an objective coefficient of 15, but SCIP 8 is incorrectly assigning 
> binary variable X_3__1_ to 1 instead even though its objective 
> coefficient is higher at 20.
>
> Would you have any advice if there is something we are using SCIP 
> incorrectly, or if this is a legitimate bug? Other than the attached 
> parameter settings, we are using the pre-built binaries on ubuntu x86. 
> Many thanks in advance.
>
> Regards,
> Barry Hurley
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip

-- 
Lisa-Marie Manke, M.Sc.

Technische Universität Carolo-Wilhelmina zu Braunschweig

Institut für Mathematische Optimierung   |         AIP
        Universitätsplatz 2               |   Mühlenpfortstr. 23
                                 38106 Braunschweig

Cluster of Excellence Sustainable and Energy-Efficient Aviation (SE²A)
Hermann-Blenk-Str. 42
38108 Braunschweig
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20230425/6f563197/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: problem_formulations.zip
Type: application/zip
Size: 4681768 bytes
Desc: not available
URL: <http://listserv.zib.de/pipermail/scip/attachments/20230425/6f563197/attachment.zip>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: log.zip
Type: application/zip
Size: 4292 bytes
Desc: not available
URL: <http://listserv.zib.de/pipermail/scip/attachments/20230425/6f563197/attachment-0001.zip>