APES is a grid world environment with ability for partial observability. It could be used for multi-agent experiments. Its uniqueness is coming from the wide options for vision for every agent.
To read more information about the environment please visit : Arxiv link coming soon
To install the environment we suggest to use conda. In Linux you can see the following steps:
git clone https://github.com/aqeel13932/APES.git
conda create -n env_name python=3 matplotlib numpy jupyter scikit-image
source activate env_name
If you want to check video recording you can see an example in Example.ipynb but you need to install additional library
pip install scikit-video
You can check that everything is alright by running the following commands, it'll also print out some information about the performance.
cd APES
python Performance.py
If you want to see an example of the environment please check Example.ipynb, it contains a simple and a complex example.
Sample video slowed down and generated by Example.ipynb is below
If you have any question please feel free to email me at: aqeel.labash@gmail.com
Any suggestion or pull request is welcome. If our environment was helpful for you and you want to cite our paper:
- Parallelize FOV_Calculator for better performance (pull requests are welcome :) ).
- Improve the documentation.