Environments + agents for Open AI Gym

• Solving partial state/multi-agent control problems with RL
• Implementations of continuous control PAAC and A3C
• Finance/trading environments

Swarm environment

Learning to stop a swarm of locusts from "rolling"

Random	After 24 hours

To train

python scripts/train_paac_conv.py -d /gpu:0 --height=84 --clip_norm=1

Acknowledgements

Bibs and bobs borrowed from https://github.com/Alfredvc/paac and https://github.com/dennybritz/reinforcement-learning/