A Dynamic Zoning Method for Road Networks Considering the Temporal and Spatial Similarity of Intersections
Zoning intersections with homogeneity and correlation in the road network is a prerequisite for imple-menting coordinated control strategies for road network zoning.This paper proposes a dynamic partitioning method of a road network that considers the temporal similarity of traffic flow and the spatial adjacency 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 matrix;Secondly,considering the spatial adjacency relationship of the intersec-tion,the spectral clustering method is used to combine the similarity The matrix constructs the dynamic partition model of the road network.Finally,taking the surrounding road network of Beijing International Trade Center as an example,combined with the modularity as the evaluation criterion,the performance analysis of the partition effect of the two algorithms considering the average saturation and the spatial and temporal similarity of the flow is carried out respectively.The calculation example results show that the modularity of the two proposed partitioning methods con-sidering the spatiotemporal similarity is between 0.3 and 0.7,indicating that the road network partitioning effect is good.The longest common substring-based partitioning method is better than The partitioning effect of the algorithm considering the average saturation is better,and its modularity is 0.49,while the partitioning method based on Haus-dorff distance is worse at 0.33.
Intelligent transportationRoad network space divisionTime sequence similarity algorithmSpectral clusteringModularity
NETL
NSTL
万方数据
