RoadnetSZ: Road networks and traffic flows in Shenzhen

In this repository, we release the Shenzhen dataset and code for multi-agent traffic signal control.

• Dataset. We provide two versions that can run on both SUMO and CityFlow platforms.
• Code. We provide three RL-based methods as baselines.

Table of contents

• RoadnetSZ: Road networks and traffic flows in Shenzhen
  • Table of contents
  • Open Datasets
  • Code list
  • Cite

Open Datasets

#	Name	Platform	Figure	Dataset
1	Fuhua (hilight)	CityFlow		Roadnet: roadnet_1_33.json Flow: anon_1_33_fuhua_24hto1w_2490.json anon_1_33_fuhua_4_27_24hto1w_4089.json
2	Fuhua (metavim)	CityFlow		Roadnet: fuhua_cityflow.json Flow: fuhua_real_1775.json fuhua_2570.json fuhua_4770.json
3	FuTian	SUMO		FuTian.net.xml FuTian.edg.xml FuTian.nod.xml FuTian.tll.xml FuTian.typ.xml FuTian.con.xml
4	BaoAn	SUMO		BaoAn.net.xml BaoAn.edg.xml BaoAn.nod.xml BaoAn.tll.xml BaoAn.typ.xml BaoAn.con.xml
5	PCL	SUMO		pcl.net.xml pcl.edg.xml pcl.nod.xml pcl.tll.xml pcl.typ.xml pcl.con.xml pcl.trips.xml pcl.sumocfg

Code list

#	Title	Figure	Code	Tutorial
1	Hierarchically and Cooperatively Learning Traffic Signal Control		code	tutorial
2	MetaVIM: Meta Variationally Intrinsic Motivated Reinforcement Learning for Decentralized Traffic Signal Control		code	tutorial
3	Learning To Share In Multi-Agent Reinforcement Learning		code	tutorial

Cite

If you use Shenzhen Dataset in your work, please cite it as follows:

@misc{RoadnetSZ,
	title = {RoadnetSZ},
	author = {Bingyu, Xu and Liwen, Zhu and Yuxuan, Yi and Zongqing, Lu and other contributors},
	year = {2022},
	howpublished = {\url{https://github.com/zhuliwen/PKU_Traffic_Lights}},
	note = {Accessed: 2022-05-01},
}

@inproceedings{xu2021hierarchically,
  title={Hierarchically and cooperatively learning traffic signal control},
  author={Xu, Bingyu and Wang, Yaowei and Wang, Zhaozhi and Jia, Huizhu and Lu, Zongqing},
  booktitle={AAAI Conference on Artificial Intelligence},
  year={2021}
}

@inproceedings{yi2022learning,
  title={Learning to Share in Multi-Agent Reinforcement Learning},
  author={Yi, Yuxuan and Li, Ge and Wang, Yaowei and Lu, Zongqing},
  booktitle={Advances in Neural Information Processing Systems},
  year={2022}
}

@article{zhu2023variationally,
  title={MetaVIM: Meta Variationally Intrinsic Motivated Reinforcement Learning for Decentralized Traffic Signal Control},
  author={Zhu, Liwen and Peng, Peixi and Lu, Zongqing and Tian, Yonghong},
  journal={IEEE Transactions on Knowledge and Data Engineering, online},
  year={2023}
}