A multi agent re-enforcement learning environment for many on many bot fight between space ships.
Random Agent on 2 vs 2 gameplay
Tasks | File names | Status |
---|---|---|
Environment | battle_env.py | Done |
Self play bot: CE | self_play_cross_entropy.py | Done |
Genetic Algorithm bot: ES | es_multi_processor_aws.py | In progress |
RL bot: PPO | space_battle_ppo.py | In progress |
Resources: RL: Re-enforcement learning John Schulman 1: Deep Reinforcement Learning
- https://www.youtube.com/watch?v=aUrX-rP_ss4
- sample code ( Cross Entropy and Actor Critic Method) : http://rl-gym-doc.s3-website-us-west-2.amazonaws.com/mlss/index.html
ES: Evolutionary strategies:
- https://blog.openai.com/evolution-strategies/ https://arxiv.org/abs/1703.03864 CE: Cross Entropy
- Learning Tetris using the noisy cross-entropy method http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.81.6579&rep=rep1&type=pdf
PPO: Proximal Policy Optimization Algorithms
email: [email protected]