Cross entropy (CE) method is a probabilistic model based method. How to show that it always converges to an exact global optimal solution? How to set the number of samples $N$ and the number of elite samples $N_{\text{elite}}$?
Asked
Active
Viewed 56 times
1
-
Did you manage to solve the problem? Actually, I have the proof from Lieber(1998) but I can not understand some details of the proof. Maybe we could collaborate on this? – entropy Aug 24 '23 at 22:37
-
I have posted a question here. You can check and if necessary we can chat a bit about the problem. – entropy Aug 24 '23 at 22:48