Hello everybody,
I have a general question. Is it possible to code a hybrid metaheuristic in GAMS? I believe it wont be efficient in computational time. Consider that we have a simple location allocation problem in which location variables are set by simulated annealing, once they are fixed, the allocation variables will be defined by solving an optimization problem. I used MATLAB as it is more efficient for meta heuristic coding but the optimization part in matlab is much slower than GAMS. I wonder if I should switch to another software (ie. using C for the metaheuristic and link it to CPLEX for the allocation part) or I can handle both in GAMS.
Thanks for your help
–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.
Hello Fateme: Two things:
- I’m sure you would be able to put it together in GAMS. I don’t know about the computing time, my guess is it wouldn’t be significantly worse than MatLab, at least for not-very-large structures.
Although a usually under-represented side of GAMS, I have learned over time and use to think of GAMS as a very powerful vector computing engine, besides the optimization part. I can help you if you have any concrete doubt about a particular part of the process. i.e. generating random variables, storing, looping and such.
- If you already have a MatLab implementation, another option should be to use gdx and call GAMS from MatLab for the optimization part. Even if it appears as a complicated option, it should be much easier than writing custom C code. Computing time could be an issue here, I don’t have experience with GAMS calls from MatLab but iirc some discussion happened on this list about the subject.
Best regards and good luck !
Claudio
On Fri, Feb 20, 2015 at 10:10 AM, Fateme Fotuhi wrote:
Hello everybody,
I have a general question. Is it possible to code a hybrid metaheuristic in GAMS? I believe it wont be efficient in computational time. Consider that we have a simple location allocation problem in which location variables are set by simulated annealing, once they are fixed, the allocation variables will be defined by solving an optimization problem. I used MATLAB as it is more efficient for meta heuristic coding but the optimization part in matlab is much slower than GAMS. I wonder if I should switch to another software (ie. using C for the metaheuristic and link it to CPLEX for the allocation part) or I can handle both in GAMS.
Thanks for your help
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.