A spatiotemporal graph generative adversarial networks for short-term passenger flow prediction in urban rail transit systems

ABSTRACT

Most short-term passenger flow prediction methods only consider absolute errors as the optimization objective, which fails to account for spatial and temporal constraints on the predictions. To overcome these limitations, we propose a deep learning-based spatiotemporal graph generative adversarial network (STG-GAN) to accurately predict network-wide short-term passenger flows of the urban rail transit with higher efficiency and lower memory occupancy. Our model is optimized in an adversarial learning manner and includes (1) a generator network including gated temporal conventional networks (TCN) and weight sharing graph convolution networks (GCN) to capture structural spatiotemporal dependencies and generate predictions with a small computational burden; (2) a discriminator network including a spatial discriminator and a temporal discriminator to enhance spatial and temporal constraints of the predictions. The STG-GAN is evaluated on two datasets from Beijing Subway. Results illustrate its superiority and robustness.

KEYWORDS:

• Deep learning
• short-term passenger flow prediction
• GCN
• GAN
• urban rail transit

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (Nos. 72201029, 72288101), China Postdoctoral Science Foundation (No. 2022M720392), and the Double-First Class” Major Research Programs, Educational Department of Gansu Province (No. GSSYLXM-04).

Notes on contributors

Jinlei Zhang

Jinlei Zhang was born in Hebei Province, China. He received a Ph.D. degree from Beijing Jiaotong University, China. He is currently an assistant professor at Beijing Jiaotong University. His research interests include machine learning, deep learning, traffic datamining and applications, and dynamic traffic modeling and management.

Hua Li

Hua Li was born in Hebei Province, China. She is currently working toward a master's degree with the State Key Laboratory of Rail Traffic Control and Safety of Beijing Jiaotong University. Her research interests include machine learning, deep learning, traffic datamining and applications, and traffic flow prediction.

Shuxin Zhang

Shuxin Zhang was born in GuangDong Province, China. He is currently working toward a PHD's degree with the State Key Laboratory of Rail Traffic Control and Safety of Beijing Jiaotong University. His research interests include machine learning, deep learning, traffic data mining and applications, and traffic flow prediction.

Lixing Yang

Lixing Yang received the B.S. and M.S. degrees from the Department of Mathematics, Hebei University, Baoding, China, in 1999 and 2002, respectively, and the Ph.D. degree from the Department of Mathematical Sciences, Tsinghua University, Beijing, China, in 2005. Since 2005, he has been with the State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, where he is currently a Professor. He is the author or co-author of more than 80 papers published in national conferences, international conferences, and premier journals. His current research interests include stochastic programming, fuzzy programming, intelligent systems, and applications in transportation problems and rail traffic control systems.

Guangyin Jin

Guangyin Jin is a Ph.D. candidate at the College of Systems Engineering of the National University of Defense Technology. His research interest falls in the area of spatiotemporal data mining, urban computing, and intelligent transportation.

Jianguo Qi

Jianguo Qi received a B.S. degree from Shandong Jianzhu University, Jinan, China, in 2012, and a Ph.D. degree from Beijing Jiaotong University, Beijing, China, in 2019. Since then, he has been a Lecturer with the State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University. His current research interests include the optimization method applied in railway operations, including train timetabling, stop planning, and passenger distributions.