Flow is a computational framework for deep RL and control experiments for traffic microsimulation.
See our website for more information on the application of Flow to several mixed-autonomy traffic scenarios. Other results and videos are available as well.
Please direct your technical questions to Stack Overflow using the flow-project tag.
We welcome your contributions.
- Please report bugs and improvements by submitting GitHub issue.
- Submit your contributions using pull requests. Please use this template for your pull requests.
- We are actually on SUMO 1.6.0, so after installing everything visit https://github.com/eclipse/sumo and install version 1.6.0 by looking at the versions for the right commit. Then follow the instructions here if you have a mac https://sumo.dlr.de/docs/Installing/MacOS_Build.html. If they haven't upgraded SUMO yet, you can also follow the Brew instructions to install it via Brew.
- The run scripts are located in examples/rllib/multiagent_exps.
- The relevant environments that define step, reset, etc. are in flow/envs/multiagent/bayesian_0_no_grid_env.py
- To replay a policy, look at flow/visualize/visualizer_rllib.py
- For an example that uses future state prediction run examples/sumo/bayesian_1_predict.py
- The rule based controller is called RuleBasedController
- The pretrained controller which you can use by giving a path to a policy is called PreTrainedController.
If you use Flow for academic research, you are highly encouraged to cite our paper:
C. Wu, A. Kreidieh, K. Parvate, E. Vinitsky, A. Bayen, "Flow: Architecture and Benchmarking for Reinforcement Learning in Traffic Control," CoRR, vol. abs/1710.05465, 2017. [Online]. Available: https://arxiv.org/abs/1710.05465
If you use the benchmarks, you are highly encouraged to cite our paper:
Vinitsky, E., Kreidieh, A., Le Flem, L., Kheterpal, N., Jang, K., Wu, F., ... & Bayen, A. M, Benchmarks for reinforcement learning in mixed-autonomy traffic. In Conference on Robot Learning (pp. 399-409). Available: http://proceedings.mlr.press/v87/vinitsky18a.html
Flow is supported by the Mobile Sensing Lab at UC Berkeley and Amazon AWS Machine Learning research grants. The contributors are listed in Flow Team Page.