| 676 | 0 | 806 |
| 下载次数 | 被引频次 | 阅读次数 |
【背景】随着共享单车在城市交通中的普及，供需失衡问题日益突出，精准预测其骑行时空需求是解决该问题的关键，而现有研究在捕捉共享单车需求的时空关联方面仍存在不足。【目标】构建一种融合时空特征关联的图卷积神经网络（Graph Convolutional Neural Networks,GCN）模型，以提升共享单车骑行需求的预测精度，为运营商提供科学的调度依据。【方法】以上海市2018年摩拜单车订单数据为基础，引入骑行时序特征矩阵，并结合地理空间、土地利用及需求序列相似性构建邻接矩阵，刻画研究区域内不同位置间的关联。【结果】融合时空特征关联的GCN模型在R2、RMSE、MAE和WMAPE等指标上均显著优于长短期记忆神经网络模型，其中基于土地利用的GCN模型预测效果最优，借车与还车需求的R2分别达到0.86和0.85，证实土地利用是影响需求分布的核心因素。【应用】本研究可为共享单车的精准需求预测提供技术支持，助力运营商实现高效调度与运维，提升共享单车系统的服务效率与用户满意度。
Abstract：[Background] Owing the widespread use of shared bikes in urban transportation, the issue of supply-demand imbalance has become prominent. An accurate prediction of their spatiotemporal demand is essential to address this issue. Existing studies do not adequately capture the spatiotemporal correlations of bike-sharing demand. [Objective] This study proposes a graph convolutional neural network(GCN) model that captures both spatial and temporal correlations to improve the prediction accuracy of bike-sharing demands and provide a scientific basis for operational scheduling.[Method] Utilizing the 2018 Mobike bike-sharing data from Shanghai, the model incorporates a time-series feature matrix and adjacency matrices spanning three dimensions: geospatial proximity,land use, and demand-sequence similarity. These matrices enable the model to represent the complex relationships among various locations within the study area. [Result] The GCN models, which incorporate spatial features, significantly outperforms the long short-term memory model, which only considers temporal dynamics across all evaluated metrics, including R2, RMSE, MAE, and WMAPE.Among the three types of adjacency matrices, the GCN model that incorporates land use yields the most accurate predictions, with R2 values of 0.86 and 0.85 for pick-up and drop-off demands, respectively. These results underscore the importance of land use in bike-sharing demands. [Application]This study offers valuable technical support for precise demand prediction in bike-sharing ridership,thereby assisting operators in implementing more efficient scheduling and operational strategies, as well as enhancing the service efficiency and user satisfaction of the bike-sharing system.
[1]上海市城乡建设和交通发展研究院.数读上海交通——2024上海综合交通发展分析报告[EB/OL].(2025-04-10)[2025-06-07]. https://mp.weixin.qq.com/s/6pGE5Vrikn5_PEjKuZbzWg.
[2]SATHISHKUMAR V E, CHO Y. A rule-based model for Seoul bike sharing demand prediction using weather data[J]. European Journal of Remote Sensing, 2020, 53(sup1):166-183.
[3]孙启鹏,曾开邦,张锴琦,等.北京市共享单车出行的时空规律与需求预测研究[J].交通运输系统工程与信息,2022, 22(1):332-338.SUN Qipeng, ZENG Kaibang, ZHANG Kaiqi, et al. Spatiotemporal travel patterns and demand prediction of shared bikes in Beijing[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(1):332-338.
[4]XU C, JI J, LIU P. The station-free sharing bike demand forecasting with a deep learning approach and large-scale datasets[J]. Transportation Research Part C:Emerging Technologies, 2018, 95:47-60.
[5]周瑜,张梦蝶.基于长短期记忆网络的共享单车真实需求预测方法[J].科学技术与工程, 2025, 25(1):394-403.ZHOU Yu, ZHANG Mengdie. Real demand prediction method of shared bike based on LSTM[J]. Science Technology and Engineering, 2025, 25(1):394-403.
[6]曹旦旦,范书瑞,张艳,等.基于长短期记忆神经网络模型的共享单车短时需求量预测[J].科学技术与工程,2020, 20(20):8344-8349.CAO Dandan, FAN Shurui, ZHANG Yan, et al. Shortterm demand forecasting of shared bicycles based on long short-term memory neural network model[J]. Science Technology and Engineering, 2020, 20(20):8344-8349.
[7]MA C, LIU T. Demand forecasting of shared bicycles based on combined deep learning models[J]. Physica A:Statistical Mechanics and Its Applications, 2024, 635:129492.
[8]LI X, XU Y, ZHANG X, et al. Improving short-term bike sharing demand forecast through an irregular convolutional neural network[J]. Transportation Research Part C:Emerging Technologies, 2023, 147:103984.
[9]马云鹤,王玉玫,赵宇帆.基于时空融合图的共享单车需求预测系统[J].计算机测量与控制, 2023, 31(2):97-103.MA Yunhe, WANG Yumei, ZHAO Yufan. Spatial-temporal fusion graph based bike-sharing demand prediction systems[J]. Computer Measurement&Control, 2023, 31(2):97-103.
[10]SARDINHA C, FINAMORE A C, HENRIQUES R.Context-aware demand prediction in bike sharing systems:incorporating spatial, meteorological and calendrical context[EB/OL].(2021-05-03)[2025-06-09]. https://arxiv.org/abs/2105.01125.
[11]QIAN H, WANG J, CHEN Y, et al. CGA-STNet:a dockless shared bicycle demand prediction model considering multiple spatial features and time periodicity[J]. Expert Systems with Applications, 2025, 265:126100.
[12]杜姿晨,郑长江,郑树康,等.基于ResNet-GCN-Transformer的多时间粒度地铁短时客流预测[J/OL].交通运输工程与信息学报:1-24.(2025-03-24)[2025-06-09].https://doi.org/10.19961/j.cnki.1672-4747.2025.02.018.DU Zichen, ZHEN Changjiang, ZHEN Shukang, et al.Short-time passenger flow prediction of multi-time granularity subway based on ResNet-GCN-Transformer[J/OL]. Journal of Transportation Systems Engineering and Information Technology:1-24.(2025-03-24)[2025-06-09]. https://doi.org/10.19961/j.cnki.1672-4747.2025.02.018.
[13]ZHOU C, HU J, ZHANG X, et al. A bike-sharing demand prediction model based on spatio-temporal graph convolutional networks[J]. PeerJ Computer Science,2024, 10:e2391.
[14]YANG D, WU R, WANG P, et al. FF-STGCN:a usage pattern similarity based dual-network for bike-sharing demand prediction[J]. PLoS One, 2024, 19(3):e0298684.
[15]冯紫嫣,李想,昌锡铭,等.基于时空需求预测的共享单车动态搬运与回收优化研究[J].系统工程理论与实践, 2025, 45(8):2753-2772.FENG Ziyan, LI Xiang, CHANG Ximing, et al. Research on optimization of dynamic bike-sharing repositioning and collection based on spatio-temporal demand prediction[J]. Systems Engineering-Theory&Practice,2025, 45(8):2753-2772.
[16]SATHISHKUMAR V E, PARK J, CHO Y. Using data mining techniques for bike sharing demand prediction in metropolitan city[J]. Computer Communications, 2020,153:353-366.
[17]廖洋,罗霞,王红杰.考虑时空自相关性的共享电动汽车出行选择影响因素分析[J].交通运输工程与信息学报, 2024, 22(4):67-84.LIAO Yang, LUO Xia, WANG Hongjie. Analysis of factors influencing shared electric vehicle travel choice considering spatiotemporal autocorrelation[J]. Journal of Transportation Engineering and Information, 2024, 22(4):67-84.
[18]吴静娴,唐桂孔,李文翔.建成环境对共享单车分时骑行量的非线性作用研究:以上海市为例[J].交通运输系统工程与信息, 2024, 24(1):290-298, 310.WU Jingxian, TANG Guikong, LI Wenxiang. Nonlinear effect of built environment on bike-sharing ridership at different time periods:a case study from Shanghai[J].Journal of Transportation Systems Engineering and Information Technology, 2024, 24(1):290-298, 310.
[19]尹超英,周悦,徐震宇,等.建成环境与主观感知对共享单车地铁换乘出行的非线性影响[J].交通运输系统工程与信息, 2025, 25(5):169-178.YIN Chaoying, ZHOU Yue, XU Zhenyu, et al. Nonlinear effects of built environment and subjective perception on bike-sharing and metro transfer trips[J]. Journal of Transportation Systems Engineering and Information Technology, 2025, 25(5):169-178.
[20]HUA M, CHEN J, CHEN X, et al. Forecasting usage and bike distribution of dockless bike-sharing using journey data[J]. IET Intelligent Transport Systems, 2020, 14(12):1647-1656.
[21]柯日宏,吴升,柯玮文.一种识别共享单车潮汐点的时空模型和基于KNN-LightGBM的租还需求预测方法[J].地球信息科学学报, 2023, 25(4):741-753.KE Rihong, WU Sheng, KE Weiwen. A spatial-temporal model for identifying tidal shared-bicycle stops and bicycle sharing demand prediction based on KNN-LightGBM[J]. Journal of Geo-Information Science, 2023, 25(4):741-753.
[22]陈红,陈恒瑞,史转转,等.公共自行车使用时空特性挖掘及租还需求预测[J].交通运输系统工程与信息,2021, 21(2):238-244, 250.CHEN Hong, CHEN Hengrui, SHI Zhuanzhuan, et al.Spatiotemporal characteristics mining and demand forecasting of shared bicycle borrow and return[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(2):238-244, 250.
[23]刘恒孜,贺玉龙,宋太龙,等.共享单车需求预测及调度优化[J].科学技术与工程, 2021, 21(35):15247-15254.LIU Hengzi, HE Yulong, SONG Tailong, et al. Demand forecast and scheduling optimization of shared bicycles[J]. Science Technology and Engineering, 2021, 21(35):15247-15254.
[24]ZHANG S, TONG H, XU J, et al. Graph convolutional networks:a comprehensive review[J]. Computational Social Networks, 2019, 6(1):11.
[25]TOBLER W R. A computer movie simulating urban growth in the Detroit region[J]. Economic Geography,1970, 46(sup1):234-240.
[26]HERRMANN M, WEBB G I. Amercing:an intuitive and effective constraint for dynamic time warping[J].Pattern Recognition, 2023, 137:109333.
基本信息:
DOI：10.19961/j.cnki.1672-4747.2025.07.035
中图分类号:U491.225
引用信息:
[1]吴静娴,董汉宁,唐桂孔.融合需求时空特征关联的共享单车骑行量预测[J].交通运输工程与信息学报,2026,24(01):80-89.DOI:10.19961/j.cnki.1672-4747.2025.07.035.
基金信息:
上海市哲学社会科学规划课题项目（2022ZGL008）
2025-07-30
2025
2025-12-11
2025
1
2025-08-15
2025-08-15
2025-08-15