Subdivision method of urban road traffic based on spectral clustering

Abstract

The invention provides a method for dividing urban road traffic sub-regions based on spectral clustering, establishes an undirected road network graph, and calculates the road network density Laplacian matrix L; according to the eigenvector v corresponding to k eigenvalues 1 ,...,v k Construct the clustering basic matrix, and implement the K-means clustering algorithm to obtain k clusters; form the initial sub-area; randomly merge isolated points into adjacent sub-areas to obtain the current total number of sub-areas; identify outliers through deviation detection , and solve the optimal outlier attribution problem through boundary adjustment; through the similarity evaluation inside and outside the sub-area and the similarity change analysis before and after sub-area splitting and merging, iteratively optimize the sub-area scheme, and finally obtain the nodes in the sub-area The road traffic network division scheme with the largest correlation similarity and the largest sub-interval difference. This method efficiently and stably obtains the division scheme of the traffic road network according to the similar characteristics of the node correlation degree, and guarantees the characteristic difference of the sub-interval and the similarity of the characteristics within the sub-region to the greatest extent.

Description

technical field [0001] The invention relates to a method for dividing urban road traffic sub-regions based on spectrum clustering. Background technique [0002] The urban road traffic network is a complex network with random characteristics. A collection of closely related intersections and road sections is composed of sub-areas, and the control, induction, management, and planning are implemented according to the characteristics of each sub-area, which can reduce the complexity of the road traffic system. Even from the macro level to achieve the effect of coordination and optimization. In the field of traffic signal control, dynamic control sub-area division can improve the stability and reliability of signal cooperative control; in the field of traffic guidance, the division of sub-areas provides support for the synergy induction effect; In terms of domains, static sub-divisions are mainly used, taking macroscopic characteristics such as land use, population, and economy ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method for dividing urban road traffic sub-regions based on spectral clustering, through multiple constraint conditions and iterative evaluation of similarity, to obtain the optimal solution with the largest similarity in the sub-region and the largest difference between sub-intervals, The algorithm for solving the bipartite clustering existing in the prior art does not make full use of other eigenvectors containing useful segmentation information when the graph segmentation results are obtained based on Laplacian non-zero small eigenvalues ​​and their eigenvector clustering, and in When multiple subgraphs need to be divided, it must be used iteratively, and the problems of large amount of information loss, low calculation efficiency and unstable defects appear

Images