 |
 |
Hi all, I'm having some problems running GP with coevolution with more than one subpopulation. When I run my problem with only 1 subpop, the run goes fine. I am able to run every individual in the subpop against N random individuals in the same subpop. Problem #1: What I would like to do is run every individual in every subpop against N random individuals from different subpops. Any idea on which parameters control this? Or do I have to write custom code, and if so, what do I need to modify? Problem #2: When I run more than one subpop, it seems that the individuals generated for the second subpop are all only one terminal of the return type, as if the initializer is not generating individuals properly. Can someone shed light on this? I'm attaching my params file for your perusal. Thanks for any help! --Rob