I'm looking for a solution approach for my MIP model, but I couldn't find any specific books about this issue whose solution approach is suitable for my model. There are a lot of exact, heuristic or hybrid approaches. Besides, there are a lot of acceleration techniques that can reduce the solution time and I don't know which of them is appropriate for mine. Could anyone help me on what should I do?

  • 4
    $\begingroup$ It may help to give a formulation of your model. $\endgroup$ – TheSimpliFire Mar 2 at 10:33
  • $\begingroup$ @maryam Welcome to OR.SE, providing your model or at least explain which kind of constraints you have in your model increases the possibility of finding a good answer. $\endgroup$ – Oguz Toragay Mar 2 at 15:58

I guess that depends on the application and the constraints and expectations that the end users have.

For example, do users work with the model interactively, trying different parameter choices and then look at the results immediately? This would mean that the solution time should be in seconds or few minutes. But if it's a planning problem that can be run over night, you can invest hours into the solving time.

Is an optimal solution (or bounds) required? Then a model-based approach would be required, otherwise a heuristic might be better.

Did you try solving the model with a MIP solver? If it's fast enough, then I'd recommend sticking with that, because it's often easier to modify a model with new requirements, compared to changing a handcrafted heuristic algorithm.

  • $\begingroup$ thanks for your answer. an optimal solution is required and I'm looking for a model-based approach. I've solved it with a MIP solver but it took a long time to be solved. actually, I'm so interested to use exact approaches like branch and price or row generation,.... but, I don't know that either they are suitable for my model or not. $\endgroup$ – maryam pourbeirami Mar 2 at 8:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.