Traffic Light Control Baselines

Install

Requirements

Before you begin, ensure you have met the following requirements:

• numpy==1.20.2
• keras==v2.3.1
• python==3.7.10
• tensorflow==1.14.0
• pandas==1.2.3
• scipy==1.6.2
• seaborn==0.11.1
• CityFlow

Newer versions of the above items may not be fully compatible with our code.

Conda Env

To make reproducibility easier, using a conda environment it is possible to load all dependencies.

To create an environment from an environment file:

$conda env create -f conda_environment.yaml

Usage

Just run any of the run_*.py scripts and pass the path of config file.

Example:

python run_dqn.py ./envs/jinan_3_4/config.json

How to cite this research

For citing this work, please use the following entries:

@InProceedings{Schreiber+2022ijcnn,
	author = {Schreiber, Lincoln and Alegre, Lucas N. and Bazzan, Ana L. C. and Ramos, Gabriel {\relax de} O.},
	title = {On the Explainability and Expressiveness of Function Approximation Methods in RL-Based Traffic Signal Control},
	booktitle = {2022 International Joint Conference on Neural Networks (IJCNN)},
	OPTpages = {},
	year = {2022},
	address = {Padova, Italy},
	month = {July},
	publisher = {IEEE},
	OPTdoi = {},
	OPTurl = {https://doi.org/},
	note = {Forthcoming}
}

Publications

1. L. Schreiber, L. N. Alegre, A. L. C. Bazzan, and G. O. Ramos, “On the Explainability and Expressiveness of Function Approximation Methods in RL-Based Traffic Signal Control,” in 2022 International Joint Conference on Neural Networks (IJCNN), Padova, Italy, 2022. [LINK IN PROGRESS]

2. Schreiber, L. V., Ramos, G. de O. & Bazzan, A. L. C. (2021). Towards Explainable Deep Reinforcement Learning for Traffic Signal Control [Oral Presentation]. International Conference on Machine Learning Conference: LatinX in AI (LXAI) Research Workshop 2021, Virtual. LINK

3. Alegre, L. N., Ziemke, T. & Bazzan, A. L. C. (2021). Using reinforcement learning to control traffic signals in a real-world scenario: an approach based on linear function approximation. IEEE Transactions on Intelligent Transportation Systems. LINK

License

This project uses the following license: MIT.

Created from a repo that provides a OpenAI Gym compatible environments for traffic light control scenario - tlc-baselines