README.md

Job-Shop Scheduling Environment

An optimized OpenAi gym's environment to simulate the Job-Shop Scheduling problem.

If you've found our work useful for your research, you can cite the paper as follows:

@misc{tassel2021reinforcement,
      title={A Reinforcement Learning Environment For Job-Shop Scheduling}, 
      author={Pierre Tassel and Martin Gebser and Konstantin Schekotihin},
      year={2021},
      eprint={2104.03760},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Getting Started

This repository is available as a pip package:

pip install JSSEnv==1.0.0

Once installed, the environment will be available in your OpenAi's gym environment and can be used to train a reinforcement learning agent:

import gym
env = gym.make('JSSEnv:jss-v1', env_config={'instance_path': 'INSTANCE_PATH'})

Important: Your instance must follow Taillard's specification.

Project Organization

├── README.md             <- The top-level README for developers using this project.
├── JSSEnv
│   └── envs              <- Contains the environment.
│       └── instances     <- Contains some intances from the litterature.
│
└── tests                 
    │
    ├── test_state.py     <- Unit tests focus on testing the state produced by
    │                        the environment.
    │
    ├── test_rendering.py <- Unit tests for the rendering, mainly used as an example
    |                        how to render the environment.
    │
    └── test_solutions.py <- Unit tests to ensure that our environment is correct checking
                             known solution in the litterature leads to the intended make-
                             span. We also check if all actions provided by the solution are
                             legal in our environment.

---

License

MIT License