search

schrum2 / MM-NEAT

Modular Multiobjective (Hyper) Neuro-Evolution of Augmenting Topologies + MAP-Elites: Java code for evolving intelligent agents in Ms. Pac-Man, Tetris, and more, as well as code for Procedural Content Generation in Mario, Zelda, Minecraft, and more!
http://people.southwestern.edu/~schrum2/re/mm-neat.php
Other
50 stars 20 forks source link

MOME pop cap should use crowding distance instead #937

Closed schrum2 closed 1 year ago

schrum2 commented 1 year ago

When the MOME sub-pop cap is reached, use NSGA-II crowding distance to discard instead of a random discard

schrum2 commented 1 year ago

This is running and not crashing, and is now the default approach used by MOME.

Now it should really be impossible for max scores in an objective to ever drop, so that is at least one thing to check in order to have some confidence that this is working correctly. So, wait a bit longer on current test runs to take a look at this.

schrum2 commented 1 year ago

I haven't seen any contradictory evidence yet, so I'm going to close this issue