Sometime in May, Jens proposed the following:
|
| Hi, all
|
| Sorry I can not make myself understood.
|
| I programed an optimization problem in gams, and it took 6 hours to
| handle 48 data. And after checking, my tutor said I could try to
| program the same optimization problem in matlab, which he thought can
| run faster and more data. Because I am not skilled at Matlab, so I
| would like to have a shortcut that if I can exchange .gms file to .m
| file, like word file coverting to PDF file ect.
|
| Any suggestion?
Jens,
What kind of a model are you trying to implement in matlab? If you have
integer variables in your model and you intend to do some sort of
optimization, I don’t believe that the optimization toolbox in matlab will
do this. It didn’t a couple years ago at least, so it’s possible that it
can solve IP and MIP types of problems now. To the best of my knowledge the
only way to do this in matlab is to purchase something like Tomslab —
this is a matlab interface to different solvers (e.g., CPLEX).
Problem formulation can impact solution speed. If you could describe the
problem you are trying to solve, some of the people here may be able to
help. It’s not terribly clear to me what you mean by 48 data. Is that the
number of variables?
There may be a posting FAQ somewhere, but I think some of the important
things to mention would include:
What kind of problem are you trying to solve (MIP, IP, MINLP, NLP, etc.).
If the system is nonlinear what is the nature of the nonlinearities? Is
it bilinear (a product of two variable), exponential, etc.
How many equations and variables are created at execution?
Does gams give you any interesting output that might indicate problems
(like numerical issues).
\