scikit-opt
scikit-opt copied to clipboard
guofei9987
Suggestion to add 3 stop criteria for the GA algorithm
one is when the fitness reaches a specific value, it stops the loop
another is: when there is no progress on the best fitness for K generation, it stops the loop.
third: when the GA has run x minutes, it will stop.
Thank you for those suggestion!
Second and third suggestion is reasonable, I will accept it.
But I do not understand why stop loop when fitness reaches a specific value is useful.
Could you please explain more or give me a link?
in many studies, we can design the fitness function such that the optimal score is 1, and when the best individual fitness reaches 0.9, we are satisfied and can stop the running and report the results.
in many studies, we can design the fitness function such that the optimal score is 1, and when the best individual fitness reaches 0.9, we are satisfied and can stop the running and report the results.
This is reasonable too!
