Sumo-gym

- Using Petting-Zoo style MARL environment 52, 8193aa6e724dcc32afeebfe888c2061deea75620.
- Shrink the state space by setting battery level and location area.
- Set volume for charging stations 6655e84cc5fd89062a7f383c7c82b612d49725a3.

- Onboard one-agent Sumo-Gym with StableBaselines3 50

Trying to make the superagent be compatible with the reinforcement learning interfaces.

- Charging includes current and target 45
- Add "Takes_action" in observation 895d3c9e78ab1f68e105662a296c6fcfb16d7c6a, 48
- Seperate render and mode 4f9ed59a2a356bbf4f0306168b17640578933669, bf0c3101635dd3ee1d97285a1682d27e653e9db8
- Reward and battery setting for rl 0eba7b7cb70e3f57211f9aecd3cebc59662667c1
- Add COSMOS map 0258177dc3351d21a0f26c486051412798529842, 46

- Allow customized demand and cs.add.xml file ccbdf99e897d14cd585d64034057ded65a440e63
- Use the jumbo scenario and remove redundant files ccbdf99e897d14cd585d64034057ded65a440e63
- Initialize the rl training env 42
- Better the user run 43

- Register a Fleet Management Problem (FMP) environment on OpenAI-Gym 3
- Construct FMP Markov Decision Process (MDP) 6 16 20 26
- Support for sumo real-time rendering 23