Research On Road Traffic Chunk Detection Method Based On Complex Network Theory

Road traffic network is an important infrastructure to serve the economy,society and the public,and is the backbone of the comprehensive transportation system.China is a large country in road traffic,with a total length of 4.8465 million kilometers at all levels,which constitutes a complex network system.It is urgent to evaluate whether the traffic network planning is reasonable,whether the road network structure design is reasonable,and whether the traffic network maintenance and maintenance conditions are theoretical and practical problems to be solved.Based on complex network theory,to analyzing the structure and function of the road traffic network research,access to effective road traffic network structure characteristics,road traffic network chunk detection method is put forward,the traffic network sharing traffic characteristic chunk identification across administrative boundaries imprison method of traffic network structure is a new perspective to analyze the problem.In order to explore the complexity of the complex road network structure,the complex road network structure was analyzed from different levels,and the node mapping method was used to construct the road traffic network model to analyze and identify the vulnerability,robustness and key nodes distribution of the road network.Then,the improved K-Means method(FWLPA-K-Means)is adopted to divide the chunk of the road traffic network and divide the chunk locations of the shared characteristics of the road traffic network.Finally,taking Liaoning province as the research object,three levels(province,city and county)were divided to analyze the chunk characteristics of the complex road network.The results show that the road traffic network has the characteristics of chunk aggregation across administrative divisions at different levels,which provides a new scientific reference for traffic planning,design and maintenance.The main research contents are as follows:(1)By using complex network theory and graph theory,the paper studies the structure and characteristics of road traffic network,and applies the connection diversity,dynamic complexity,node diversity and multi-complexity fusion of the complex network to the road traffic network.Define the road nodes,road sides and weights of the road traffic network,and add influence factors on the weight of the road traffic network,including the adjacency matrix of the calculated network of Location,Distance and Road-level.The model of road traffic network is constructed by node mapping method,and the features of scale-free,small-world and collectitization of road traffic network are analyzed.(2)Based on the complex network theory,the structure of road traffic network is analyzed topologically.Qualitative and quantitative analysis was conducted from the perspectives of the whole province,the whole city and the whole county to analyze the characteristics of average path length,clustering coefficient and degree distribution.A variety of factors affecting the characteristics of road traffic network are proposed,the vulnerability analysis model and robustness analysis model of road traffic network are constructed,and the vulnerability and robustness of road traffic network are analyzed by using random attack mode and deliberate attack mode.(3)Considering that the interaction between link nodes in road traffic network is related to public infrastructure attributes and related to road network grade and topological distance,the concept of key node in road traffic network is proposed.Analysis of existing classic critical node analysis method,contrast to summarize the advantages and disadvantages,using Page Rank Algorithm(PA)method to sort the network key nodes of ideas,make use of the location attribute,topological distance and road level impact factor introduced algorithms,building improvement Page Rank Algorithm(FWLPA)sorting method road traffic network key nodes.(4)A traffic chunk detection method based on complex network theory is proposed.Define the concept of road traffic network chunk division.According to the results of the ranking of key nodes in the road traffic network and the concept of the road traffic network chunk detection,the shared characteristics of the road traffic network chunk detection is carried out.The improved K-Means Algorithm(FWLPA-K-Means)is applied to achieve the division result of special common chunk scope of road traffic network by crossing the limit of administrative divisions through the two indexes of key node ranking and shortest path distance.By using the road traffic network system of Liaoning province,this paper analyzes the structure characteristics and chunk structure distribution results of the network,and analyzes the general aggregation of the road traffic network from the perspective of geographic space.Through the above research contents,the special general aggregation situation of road traffic network is analyzed from the perspective of new geographic space,and the spatial structure characteristics of road traffic network and the regional aggregation scope generated from traffic operation are analyzed in depth beyond the limits of existing administrative divisions.For traffic managers and decision makers to provide decision-making reference for the rationality of traffic planning to provide a new data support.