GridWorld is an example simulation created for PyGridWorld.
It simulates a simple 2D "grid world", where AI agents can move around to avoid dangerous obstacles. It features a small and simple neural network system that evolves over time, by duplicating and mutating agents that perform the best in a given time period. To this end, it provides some tweakable parameters in its state, so that a user can modify how the simulation behaves, and examine differences in performance based on the tweaks.
- Eigen
- EnTT 3.0
- RapidJSON