Research On The Aggregation Pattern Of Highway Traffic Network Based On Complex Networ

The road transport network is an important infrastructure serving the economy,society and the public.Research on the structure of road traffic networks and the identification of the aggregation function blocks of the networks can provide a basis for improving the operational efficiency of road networks and enhancing the overall service level of road networks.At present,many scholars have studied the aggregation pattern of road networks based on complex network theory,but the theoretical analysis model of road networks constructed in the current study does not add dynamic weight influence factors,and in the analysis of the aggregation pattern of road networks,the determination of the clustering centre is mostly selected by human experience or randomly determined by the algorithm,in addition,the k-means algorithm used in the clustering of road networks In addition,the k-means algorithm used for the clustering of the road network has the limitation of tending to cluster the results into spherical clusters,which makes the results of the clustering blocks of the road network somewhat different from the actual situation.The main work of this paper is as follows.Firstly,this paper addresses the problem of not adding dynamic weighting factors to the road network theoretical analysis model constructed in the current study,and constructs a road network theoretical analysis model with dynamic traffic congestion degree weights.This makes up for the shortcomings of the current model.Then,for the problem of selecting clustering centres,this paper compares and analyses various critical node ranking algorithms.Based on this,an improved Page Rank algorithm based on the road network weights is proposed to determine the critical node ranking of the road network.The clustering centres are then determined scientifically in combination with the shortest path distance.In view of the limitations of the current algorithms for road network aggregation block classification,this paper compares and analyses various clustering algorithms.Based on this,an improved spectral clustering algorithm based on the actual characteristics of the road network is proposed,which makes the results of road network aggregation block classification more consistent with the actual situation of the road network.Then,the experimental validation of the algorithm is carried out by combining two scales: city and district.After experimental verification,the results of road network aggregation pattern analysis using the algorithm constructed in this paper are better than the traditional k-means algorithm.This is because it avoids the limitation of the k-means algorithm which tends to cluster the results into spherical clusters.And it takes advantage of the adaptability of the spectral clustering algorithm to the data distribution.And the clustering centres are scientifically determined,making the analysis results more in line with the actual situation of the road network.At the same time,based on the results of the road network aggregation block classification,the critical paths of the two example analysis areas were identified.This provides a decision reference for the traffic planning and maintenance of the road network in the region.