[SCIP] Question

Ambros Gleixner gleixner at zib.de
Tue Sep 5 00:26:10 CEST 2017


Dear Walelign,


Am 04.09.2017 um 16:52 schrieb Walelign Nikshi:
>  Dear SCIP developers,
>
> Sorry for bothering you guys.
> This is one of my multiple trial to get response for my question. I
> would really aprecitae if you can get back to me.
>
> I have a Mixed integer optimal control problem and I would like to solve
> this problem as a MINLP. I have tried to solve this using genetic
> algorithm and I failed to get a solution to it. I wanted to try this in
> SCIP and I think it is written in C language. All I have done is in
> Matlab and I was wondering if there is a way to run the Matlab code in
> SCIP package

Currently this is not possible.


> or how hard is it formulate the problem in C for this software?

If you ave no C experience, it is difficult.  You can look at the 
example code in

http://scip.zib.de/doc/html/CALLABLELIBRARY_MAIN.php

but an easier possibility might be to build a ZIMPL model 
(http://zimpl.zib.de).


> CAn the solver handle
>     -Nonlinear,

Yes.

>     -Discontinuous
>     -Non smooth , and

In principle yes, but we have no "piecewise" constraints so far.  But 
you can always use binary variables to model piecewise constraint function.

>     - Non-convex problems in objective function, constraint, and dynamics?

Yes.

> And will upto how many degrees of freedom system can be solved with
> SCIP? I have a 6 degree of freedom optimal control system (6 states)

We cannot answer this question.  SCIP is not an OC solver and how it 
performs for a MINLP discretization will very much depend on your 
formulation.

This is not the right mailing list for general questions on how to solve 
an MIOCP using an MINLP solver.  But if you make a serious effort 
yourself and get stuck on how to use SCIP, then we can try to help.

Best,
Ambros



> Could you please help me on this one?
>
> Thanks,
> Wale
>
> --
> */
> /*
> */Best regards,/*
> */
> /*
> */~~~~~~/*
> */Walelign M. Nikshi /*
> */PhD student/*
> */Dept. Mechanical Engineering/*
> */South Dakota School of Mines & Technology/*
> */501 E. Saint Joseph St/*
> */Rapid City, SD, 57701/*
> */(605) 321 3805 (mobile)/*
> */~~~~~~~~~~~~~~~~~~~~~/*
> */
> /*
>
>
> _______________________________________________
> Scip mailing list
> Scip at zib.de
> https://listserv.zib.de/mailman/listinfo/scip
>

-- 
Ambros Gleixner, Zuse Institute Berlin, http://www.zib.de/gleixner


More information about the Scip mailing list