[SCIP] Reoptimization: ensure that the same solution is not used twice?

[Jakob Witzig]

Hi Natalia,

usually calling SCIPfreeTransform() is absolutely fine whenever you want 
to change and resolve your problem. But when you use the reoptimization 
feature you need to call SCIPfreeReoptSolve() and afterwards 
SCIPchgReoptObjective() to install the new objective function. This 
preserves some needed internal data and should fix you assertion failure.

Cheers,
Jakob

Am 02.05.2017 um 19:04 schrieb Natalia Perina:
> Hi Jakob,
>
> Thanks for your response. When I enable reoptimization - I am now 
> getting a fail when I try to solve that says:
>
> Assertion failed: (SCIPhashmapExists(reopt->glblb, (void*)vars[i])), 
> function SCIPreoptInstallBounds, file src/scip/reopt.c, line 8175.
>
>
> I am enabling reoptimization, setting objectives and constraints, 
> solving, then applying the FreeTransform function, then changing the 
> objective value of some constraints, then solving again. Is this 
> correct?/Any idea why the assert would fail?
>
>
> Thanks,
>
> Natalia
>
>
>
> On Tue, May 2, 2017 at 5:28 AM, Jakob Witzig <witzig at zib.de 
> <mailto:witzig at zib.de>> wrote:
>
>     Hi Natalia,
>
>     when reoptimization is enabled SCIP keeps track of all solutions
>     found so far over all reoptimization-runs. SCIP will also check
>     all these solutions w.r.t to the new objective function. Thus,
>     there is no need that you keep track of the solutions found so far
>     by your self. Moreover, I would not recommend to forbid these
>     solutions explicitly. Adding constraints/cuts that separate these
>     solutions will lead to a slow down.
>
>     Cheers,
>     Jakob
>
>
>     Am 30.04.2017 um 06:51 schrieb Natalia Perina:
>>     Hi there,
>>
>>     I am trying to use reoptimization on a subscip in my pricer. Is
>>     there a way to find one solution, change the objective function,
>>     and store/negate the solution to ensure that the same solution is
>>     not found and used in subsequent rounds of pricing/reoptimization?
>>
>>     Thanks so much - all help appreciated greatly!
>>     Natali
>>
>>
>>     -- 
>>     Natalia Perina
>>     Princeton University l 2017
>>     nperina at princeton.edu <mailto:nperina at princeton.edu>
>>     202 360 9536 <tel:%28202%29%20360-9536>
>>
>>
>>     _______________________________________________
>>     Scip mailing list
>>     Scip at zib.de <mailto:Scip at zib.de>
>>     https://listserv.zib.de/mailman/listinfo/scip
>>     <https://listserv.zib.de/mailman/listinfo/scip>
>
>     -- 
>     Jakob Witzig
>
>     Zuse Institute Berlin (ZIB)
>
>     Division Mathematical Optimization and Scientific Information
>     Research Group Mathematical Optimization Methods
>
>     Takustrasse 7
>     14195 Berlin
>
>     Tel. :+49 (0)30 84185-416 <tel:+49%2030%2084185416>
>     Fax  :+49 (0)30 84185-269 <tel:+49%2030%2084185269>
>     email:witzig at zib.de <mailto:witzig at zib.de>
>
> -- 
> Natalia Perina
> Princeton University l 2017
> nperina at princeton.edu <mailto:nperina at princeton.edu>
> 202 360 9536
-- 
Jakob Witzig

Zuse Institute Berlin (ZIB)

Division Mathematical Optimization and Scientific Information
Research Group Mathematical Optimization Methods

Takustrasse 7
14195 Berlin

Tel. : +49 (0)30 84185-416
Fax  : +49 (0)30 84185-269
email: witzig at zib.de
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.zib.de/pipermail/scip/attachments/20170503/3e98adb6/attachment.html>