Wednesday, April 1, 2009

Deceptive genomes

Having worked out the kinks in my BinInt problem, I got my Deception code working also. As with BinInt, I've now scaled the output fitness to adapt to the number and size of subproblems. This isn't important in any one run, or averaging runs with constant parameters, but it does make it easier to compare runs with different parameter choices. I'm happy!

No comments: