All workBackend
Reinforcement learning · personal projectOngoing
Pokémon GO battle simulator
Training battle agents with evolutionary reinforcement learning.
Case studyOngoingTraining battle agents with evolutionary reinforcement learning.Pokémon GO battle simulator · Reinforcement learning · personal project
A personal research project exploring how evolutionary reinforcement learning performs inside the Pokémon GO battle system.
The simulator runs battles and evolves agents over generations, steering them toward strategies that hold up across different matchups.
- Battle simulation environment for repeatable experiments.
- Evolutionary loop that mutates and selects agents over generations.
- Agent comparison across matchups and strategies.
Still to confirm
- Not public yet — results and training approach can be shared on request.