Academic Journal of Engineering and Technology Science, 2024, 7(6); doi: 10.25236/AJETS.2024.070610.
Huan Dai
School of Transportation and Logistics, Southwest Jiaotong University, Chengdu, 610031, China
Accurate railway freight volume prediction can effectively support the dynamic adjustment of freight station organization, enhance the service level and competitiveness of railway transportation. Considering the impact of railway network relationships on freight volume, this paper proposes a railway freight volume prediction method based on Spatiotemporal Graph Convolutional Neural Network (STGNN). The spatial convolution module adopts relational graph convolution to explore and integrate spatial characteristics of the railway physical network, inter-station relationships based on freight volume, and service relationships based on operational plans. In the time series module, a multi-layer Gated Recurrent Unit (GRU) is used for multi-step freight volume prediction of freight station groups. Using freight volume data along the Xiang-Yue section as the research object, the prediction results under different step sizes are compared, showing that STGNN significantly outperforms the baseline models.
Railway freight volume prediction; Relational graph convolution; Spatiotemporal graph convolutional neural network
Huan Dai. Railway Freight Volume Prediction Based on Spatiotemporal Graph Convolutional Neural Network. Academic Journal of Engineering and Technology Science (2024) Vol. 7, Issue 6: 69-74. https://doi.org/10.25236/AJETS.2024.070610.
[1] Wu Wei, Fu Zhuo, Wang Xiao. Passenger Volume Prediction Method for Transport Corridors [J]. Journal of Railway Science and Engineering, 2012, 9(05): 96-102. DOI: 10.19713/j.cnki.43-1423/u. 2012. 05.018.
[2] Li Jianguo, Xiang Wanli, Wang Jiugeng. Railway Freight Volume Prediction Based on Grey Genetic BP Neural Network [J]. Science & Technology and Industry, 2022, 22(01): 119-124.
[3] Guo Dongdong, Wang Xifu. Railway Freight Volume Prediction Based on BP Network [J]. Railway Freight, 2006(02): 21-23.
[4] Zhang Guandong, Yang Chen. Freight Volume Prediction Based on Multi-Dimensional Long Short-Term Memory Network [J]. Statistics and Decision, 2022, 38(12): 180-183. DOI: 10.13546/j.cnki.tjyjc. 2022. 12.036.
[5] Liu Qidong, Liu Chaoyue, Qiu Zixin et al. Time-Aware Transformer-Based Traffic Flow Prediction Method [J]. Computer Science, 2023, 50(11): 88-96.
[6] Xu Li, Fu Xiangyuan, Li Haoran. Spatio-Temporal Traffic Flow Prediction Model Based on Gated Convolution [J]. Computer Applications, 2023, 43(09): 2760-2765.
[7] Li, Yaguang, Yu, Rose, Shahabi, Cyrus, Liu, Yan. Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting [J]. arXiv, 2017, 20(12): 38.
[8] He Bisheng. Research on Coordination Optimization Theory and Methods of High-Speed Railway Train Operation Plans and Timetables [D]. Beijing: Beijing Jiaotong University, 2014: 101-110.
[9] Chen Rong, Liang Changyong, Lu Wenxing et al. Forecasting Tourism Flow Based on Seasonal PSO-SVR Model [J]. Systems Engineering-Theory & Practice, 2014, 34(5): 1290-1296.
[10] Chen Zhonghui, Ling Xianyao, Feng Xinxin et al. Short-Term Traffic State Prediction Approach Based on FCM and Random Forest [J]. Journal of Electronics & Information Technology, 2018, 40(8): 1879-1886.