[SCIP] Method used by the solver for solving nonlinear models

[Stefan Vigerske]

Hi,

it is a spatial branch-and-bound algorithm.
An LP relaxation is constructed and used for bounding. Branching may 
happen on variables in nonconvex terms. NLPs are solved by a NLP solver 
to find feasible solutions.
For a summary, see https://doi.org/10.1080/10556788.2017.1335312
(preprint: http://nbn-resolving.de/urn:nbn:de:0297-zib-59377)

If the NLP is recognized as convex by SCIP, then the solution from the 
NLP solver should be globally optimal. SCIP will confirm this by 
updating the LP relaxation until it has the same optimal value (i.e., 
gap is closed).

Stefan

On 11/30/21 11:52, Marcela Tais Bregadioli wrote:
> Good Morning! My name is Marcela.
> 
> I would like to know what methodology is used by SCIP solver for solving
> continuous non-linear mathematical models.
> 
> I have been looking at the solver page, but I have not found this specific
> information for continuous non-linear programming.
> 
> Thanks in advance.
> 
> Sincerely,
> 
> Marcela
> 
> Enviado do Email <https://go.microsoft.com/fwlink/?LinkId=550986> para Windows
> 
> 
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>
> 	Livre de vírus. www.avast.com
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>.
> 
> 
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> 
> 
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip
>