Page 175 - 《软件学报》2021年第12期

软件学报 ISSN 1000-9825, CODEN RUXUEW E-mail: jos@iscas.ac.cn
Journal of Software,2021,32(12):3839−3851 [doi: 10.13328/j.cnki.jos.006115] http://www.jos.org.cn
©中国科学院软件研究所版权所有. Tel: +86-10-62562563

∗
基于可变形卷积时空网络的乘车需求预测模型

1
1
1
于瑞云 , 林福郁 , 高宁蔚 , 李 婕 2
1
(东北大学 软件学院,辽宁 沈阳 110169)
2 (东北大学 计算机科学与工程学院,辽宁 沈阳 110169)
通讯作者: 林福郁, E-mail: 1871097@stu.neu.edu.cn

摘 要: 随着滴滴、Uber 等出租车服务的日益普及,用户的乘车需求预测逐渐成为智慧城市、智慧交通的重要组
成部分.准确的预测模型既可以满足用户的出行需求,也可以降低道路车辆空载率,有效地避免资源浪费,并缓解交
通压力.车辆服务商可以收集到大量 GPS 数据及用户需求数据,然而,如何合理运用数据进行需求预测,是关键且实
用的问题.提出一种结合城市 POI 的可变形卷积时空网络(DCSN)模型来预测区域乘车需求,模型包括两部分——可
变形卷积时空模型与 POI 需求关联模型:前者即通过 DCN 与 LSTM 建模未来需求与时空之间的相关性,后者则通
过区域 POI 差异化指数与需求差异化指数捕捉区域间的相似关系.最后使用全连接网络将两个模型整合起来,进而
得出预测结果.使用滴滴出行的大型真实乘车需求数据进行实验,最终实验结果表明,所提出的方法在预测精度上优
于现有的预测方法.
关键词: 城市计算;时空相关性;可变型卷积网络
中图法分类号: TP18

中文引用格式: 于瑞云,林福郁,高宁蔚,李婕.基于可变形卷积时空网络的乘车需求预测模型.软件学报,2021,32(12):
3839−3851. http://www.jos.org.cn/1000-9825/6115.htm
英文引用格式: Yu RY, Lin FY, Gao NW, Li J. Passenger demand forecast model based on deformable convolution spatial-
temporal network. Ruan Jian Xue Bao/Journal of Software, 2021,32(12):3839−3851 (in Chinese). http://www.jos.org.cn/1000-
9825/6115.htm

Passenger Demand Forecast Model Based on Deformable Convolution Spatial-temporal Network

1
1
1
YU Rui-Yun , LIN Fu-Yu , GAO Ning-Wei , LI Jie 2
1 (Software College, Northeastern University, Shenyang 110169, China)
2 (School of Computer Science and Engineering, Northeastern University, Shenyang 110169, China)
Abstract: With the increasing popularity of taxi services such as Didi and Uber, passengers’ demand has gradually become an important
part of smart cities and smart transportation. The accurate prediction model can not only meet the travel needs of users, but also reduce the
no-load rate of road vehicles, which can effectively avoid waste of resources and relieve traffic pressure. Vehicle service providers can
collect a large amount of GPS data and passenger demand data, but how to use this big data to forecast demand is a key and practical
problem. This study proposes a deformable convolution spatial-temporal network (DCSN) model that combines urban POI to predict
regional ride demand. Specifically, the model proposed in this study consists of two parts: the deformable convolution spatial-temporal
model and the POI requirement correlation model. The former models the correlation between future demand and time and space through
DCN and LSTM, while the latter captures the similar relationship among regions through the regional POI differentiation index and the

∗ 基金项目: 国家自然科学基金(62072094); 辽宁省兴辽英才计划(XLYC2005001); 辽宁省重点研发计划(2020JH2/10100046);
中央高校基本科研业务费专项资金(N182608004)
Foundation item: National Natural Science Foundation of China (62072094); LiaoNing Revitalization Talents Prograrn (XLYC
2005001); Key Research and Development Project of Liaoning Province (2020JH2/10100046); Fundamental Research Funds for the
Central Universities (N182608004)
收稿时间: 2019-12-20; 修改时间: 2020-03-08, 2020-04-29, 2020-06-15; 采用时间: 2020-07-03