Is It always better increase the mutation probability if the problem is more complex?

I'm doing different experiments trying to check if a higher mutation probability helps to find a solution when the problem is more complex. In this case i have a sphere función with 5 dimensions as "simple problem" and a Schwefel función with 10 dimensions as "complex problem". I'm trying to solve both problems using a genetic algorithm. Each variable is representes by 5 bits encoding real values in the [-10,10] range in the sphere function and 10 bits encoding values in the range [-500,500] for the Schwefel function. The point is that trying different mutation probabilities I get the best result for mutation probability between 0.01 and 0.03 for both problems. Increase the mutation probability make both problems work worst. Is this counter intuitive? Should I get better results for the complex problem for higher mutation probabilities?



Your Answer

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

Browse other questions tagged or ask your own question.