13

Genetic algorithms evolve in fewer generations with a larger population, but also take longer to compute a generation. Are there some guide lines for balancing those two factors, in order to arrive at a viable solution as soon as possible?

Also, is this the best place for the question?

Suresh Venkat
  • 32,071
  • 4
  • 95
  • 271
Matt Munson
  • 503
  • 1
  • 5
  • 10
  • The question probably belongs on Cross Validated. – chrisaycock Feb 26 '11 at 06:45
  • 5
    I think this question is fine here, and I would like to see it answered here. As far as I understand genetic algorithms do not have as rigorous analysis as many of the common topics on cstheory, but this question is still a cstheory question and I would argue at the research level. We have answered genetic algorithm questions before: http://cstheory.stackexchange.com/questions/844/provable-statements-about-genetic-algorithms There's also a meta discussion on these sort of questions: http://meta.cstheory.stackexchange.com/questions/73/which-metaheuristic-methods-are-admissible-for-discussion – Artem Kaznatcheev Feb 26 '11 at 07:56

1 Answers1

15

There are some works on this issue. Since you don't specify for what problems you want guidelines I will list some major and some last papers of general interest. They are:

Oleksandr Bondarenko
  • 4,215
  • 1
  • 25
  • 46