A Research On The Complex Network Characteristics Of Urban Railway Traffic System And The Passenger Traffic Features

The city resident trip is more and more dependent on rail transit and the rail transit characteristics analysis has become a hot topic. At the same time, human dynamics is becoming increasingly active in the field of traffic study. However, comparing with the study on highway, bus lines, and railway transportation, the research on the network characteristic of urban railway traffic system is obviously insufficient. Based on above understanding, using the data collected from Urban Metro, the features of urban subway traffic system and passenger travel behavior are studied in this paper based on the theory and methods of statistical and human dynamics. The main works are as following:1. Analyze and explore the network properties of rail traffic, propose a new method of subway station hierarchical structure partition based on OD matrix, and divide Tianjin Metro stations into two levels.2. Empirical study on the characteristics of urban subway passenger travel behavior in a P space, including paroxysmal and memory, fluctuation and periodicity. The paper constructs 2D SOM subway passenger travel time model(i.e. classifier), and vivificates of the correctness of the method by using the cross validation of models, are very state detection using the validated model on the travel behavior of urban subway passengers.3. Reveals the spatial characteristics of subway passengers in a P space, based on the subway passengers travel volume and travel distance distribution analysis. The paper established phenomenological probability transition model simulation model based on single step displacement and single trip time on the results of the analysis on the characteristics of subway passenger space.The paper found that the subway passenger trip generation and trip time exhibited strong paroxysmal and weak memory, in line with the "human behavior of strong paroxysmal weak memory" of the law. The distribution of passenger travel volume on the site with power-law distribution, trip distance distribution accords with the two negative binomial distribution. The maximum probability of urban subway passengers travel distance for 8 stations. The city subway passengers travel amount was very state detection using 2D SOM model, the correct rate is 90%.