Gym4ReaL is a comprehensive suite of realistic environments designed to support the development and evaluation of RL algorithms that can operate in real-world scenarios.
The suite includes a diverse set of tasks exposing RL algorithms to a variety of practical challenges.

Please refer to our Website for further information about environments and reproducibility.

docs/                   # website and documentation
examples/               # example code for running each environment
gym4real/               # main Python package
    algorithms/
        {env}/          # per-env algorithms
    data/
        {env}/          # per-env data files
    envs/
        {env}/          # per-env modules

If you would like to add a new environment, propose bug fixes, or otherwise contribute to Gym4ReaL, please see the Contributing Guide.

Gym4ReaL is released under a Apache-2.0. See the LICENSE file for the full terms.

Please cite Gym4ReaL as

@misc{salaorni2025gym4realsuitebenchmarkingrealworld,
      title={Gym4ReaL: A Suite for Benchmarking Real-World Reinforcement Learning}, 
      author={Davide Salaorni and Vincenzo De Paola and Samuele Delpero and Giovanni Dispoto and Paolo Bonetti and Alessio Russo and Giuseppe Calcagno and Francesco Trovò and Matteo Papini and Alberto Maria Metelli and Marco Mussi and Marcello Restelli},
      year={2025},
      eprint={2507.00257},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2507.00257}, 
}