考虑交叉口时空相似性的路网动态分区方法  

A Dynamic Zoning Method for Road Networks Considering the Temporal and Spatial Similarity of Intersections

作　　者：张士杰 张蕊 侯先磊 马尚 ZHANG Shi-jie;ZHANG Rui;HOU Xian-lei;MA Shang(School of Civil and Traffic Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Beijing Shunyi District Municipal Commission of Urban Management,Beijing 101300,China)

机构地区：[1]北京建筑大学土木与交通工程学院,北京100044 [2]北京市顺义区城市管理委员会,北京101300

出　　处：《计算机仿真》2024年第7期189-194,221,共7页Computer Simulation

基　　金：国家自然科学青年基金(5170080357);北京未来城市设计高精尖创新中心项目(UDC2019032924)。

摘　　要：对路网中具备同质性和关联性的交叉口进行分区,是实施路网分区协调控制策略的前提。提出了一种考虑流量时序相似度和交叉口空间邻接关系的路网动态分区方法。首先针对交叉口不同方向进口道的车流量时序数据,提出了时序相似度算法,考虑最长公共子串、豪斯多夫距离计算流量时序相似度矩阵;其次考虑交叉口空间邻接关系,利用谱聚类方法结合相似度矩阵构建了路网动态分区模型;最后以北京国贸周边路网为例,结合模块度为评价准则进行算法性能分析。算例结果表明:提出的两种考虑时空相似性算法的划分方法,其模块度均在0.3至0.7之间,表明空间划分效果好,其中基于最长公共子串的划分方法最优,其模块度为0.49,而基于豪斯多夫距离的划分方法较差为0.33。Zoning intersections with homogene ity and correlation in the road network is a prerequisite for implementing coordinated control strategies for road network zoning.This paper proposes a dynamic partitioning method of a road network that considers the temporal sim ilarity of traffic flow and the spatial adjac ency of intersections.Firstly,the paper focuses on different directions of the intersection.Based on the time-series data of traffic flow at the entrance,a time-series similarity algorithm is proposed,which considers the longest common substring and Hausdorff distance to calculate the traffic-series similarity mat rix;Secondly,considering the spatial adjac ency relationship of the intersection,the spectral clustering method is used to combine the similarity The matrix constru cts the dynamic partition model of the road network.Finally,taking th e surrounding road network of Beijing Intern ational Trade Center as an example,combined with the modularity as the evaluation criterion,the performance analys is of the partition effect of the two algorithms considering the average s aturation and the spatial and temporal simil arity of the flow is carried out respectively.The calculation example results show that the modularity of the two proposed partitioning methods considering the spatiotemporal similarity is be tween 0.3 and 0.7,indicating that the road ne twork partitioning effect is good.The longest common substring-based par titioning method is better than The partitio ning effect of the algorithm considering the average saturation is better,and its modularity is 0.49,while the partit ioning method based on Hausdorff distance is worse at 0.33.

关 键 词：智能交通 路网空间划分 时序相似度算法 谱聚类 模块度 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]