Author: | Aswin Shriram Thiagarajan |
---|
- Created an environment simulator as a python package similar to the style of OpenAI gym's environment class
- Contains a start and goal (can be randomized or passed as input), has defined reward structure and transition probability (can be set to 0 for deterministic envs)
- Can be used as a test bed to train your RL/DRL agents
Planned updates:
- Dynamically changing obstacles
- Obstacle map can be provided as input file
- scale up map to different sizes (dynamic)