Assume that I want to find the global minimum of a non-linear, non-convex, multidimensional function subject to several restrictions.

Could you recommend me any deterministic strategy which can generally find an acceptable solution in a reasonable time-frame (say less than 1 hour)? Or should I go with the heuristic / stochastic route?

Ok, I will slightly rephrase the question:

If you had to find the global optimum of a multidimensional non-convex function, are there any type of problem/structure where deterministic strategies will be preferred to other stochastic/heuristic methods like simulated annealing or a genetic algorithm.

  • 2
    $\begingroup$ There's no method that can solve general non-convex large optimization problems efficiently. It's necessary to know more about the structure of your problem. $\endgroup$ – littleO Jul 16 '14 at 13:50
  • $\begingroup$ How many variables and restrictions? Are the objective function and the restrictions given by explicit expressions? $\endgroup$ – lhf Jul 16 '14 at 14:00
  • $\begingroup$ @littleO, my question is of general nature. I would like to know if there are any deterministic methods which perform well in a broad class of non-convex functions. My impression is that most people use heuristic / stochastic algorithms for large non-convex problems, but I am not an expert in the subject $\endgroup$ – user152210 Jul 16 '14 at 14:03
  • $\begingroup$ There are deterministic approaches to global optimization. See en.wikipedia.org/wiki/Global_optimization#Deterministic_methods. $\endgroup$ – lhf Jul 16 '14 at 14:05
  • 1
    $\begingroup$ This question is probably too broad. $\endgroup$ – lhf Jul 16 '14 at 14:06

Your Answer

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

Browse other questions tagged or ask your own question.