1

I'm working on an Evolutionary Algorithms framework, but I've hit a bit of confusion when it comes to the random selection method used within Ranked Selection; should I be using the same method used in Roulette Wheel Selection (i.e. selecting each solution independently), or should I use the same method used in Stochastic Universal Sampling (i.e. selecting all solutions simultaneously). I know the latter would probably be better, but I don't want to implement it this way if it's not standard practice.

Thanks in advance!

Lee.J.Baxter
  • 495
  • 4
  • 12

1 Answers1

0

Since every time people address Rank Selection they talk about Roulette Wheel, and SUS works with probabilities itself, and with a reference to this answer it seems you should select like Roulette Wheel.

Community
  • 1
  • 1
Mehraban
  • 3,164
  • 4
  • 37
  • 60