Suppose I am looking at an optimization problem with a large number of interconnected constraints, but the solution is - in some regions - extremely volatile (With volatile I mean: small mutations might drastically change the evaluation of the proposed solution. Also note that not all changes are volatile in this way, changing other parts of the solution might result in much smaller changes in the value of the solution).

In this case, is it still practical to try and find an optimal solution by using a genetic algorithm? I am worried that the potential volatility would more or less change the problem into almost a brute-force search, or are there ways around this?

  • 1
    $\begingroup$ Welcome to CS.SE! This sounds pretty vague. Can you be more precise about what type of constraints you have, or what you're trying to optimize? And what has you focused on a genetic algorithm as the way to solve it? I think you'd do better if you described what kind of constraints you have and then asked what techniques might be appropriate for solving them. Genetic algorithms might be one approach, but there might be others. As for your specific answer, the only answer we can give you is likely to be "you have to try it and see". $\endgroup$ – D.W. Dec 9 '16 at 19:03
  • $\begingroup$ Are you using / can you use real-coded genetic algorithms? $\endgroup$ – manlio Dec 10 '16 at 21:24

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.