I have a working - and long - R code to generate a mps file to be read by cplex. Because I work with large instances I am finding troubles (too much time) in producing a large mps file.

Is there a way to take advantage on the mps file generator code to add directly (without need of file generation) the problem into Cplex through Rcplex, for example?

This is because I would like to avoid a reprogramming of the problem to callback Cplex using other input styles.

  • $\begingroup$ If you don't need to check your model by produced file, you can try .sav file (binary structure), I haven't use Cplex with R. I used Cplex with C# before. But I think they can produce same output format. $\endgroup$
    – ytsao
    Dec 20, 2023 at 13:56
  • 1
    $\begingroup$ Your question is a bit unclear, specifically the part about reprogramming. Are you asking whether you can pass the MPS text directly to CPLEX without saving it as a file and reading it back, or are you asking whether the code that generates the MPS text could be modified to create the model directly (which would involve some amount of reprogramming)? $\endgroup$
    – prubin
    Dec 20, 2023 at 16:40
  • $\begingroup$ [probin] - Yes. I would like to pass a MPS-style information to cplex through R without saving it as a file. $\endgroup$
    – Di Al
    Dec 21, 2023 at 17:21


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.