- TODO
- clustering
- Fix the distance metric x Maybe employ clustering on the genome difference data
- Figure out a way to have cluster continuity across generations (ie what clusters are the same from one generation to the next)
- Implement crossover
- Make a better selection rule
- Integrate random mutation
- Allow parameterization for mutation, crossover, species difference thresholds, at the top level in run_neuroevolution
- Make a progress bar
helblazer811 / neat-polearm Goto Github PK
View Code? Open in Web Editor NEWReproducing the original NEAT paper for RL using Open AI gym