Data Analysis And Optimization Of Tianjin Public Transport Network Based On Complex Network Theory

With the continuous expansion of urban scale,the travel demand of residents is increasingly intensified.Bus travel residents become the preferred efficient and fast way to travel,but in the actual bus system,there are sites and line distribution is not reasonable,lack of theoretical support line planning,passenger travel problems such as insufficient data by using,for Tianjin bus network topology characteristics,analysis of passenger travel law,optimize the bus network,the main research content is as follows:Firstly,this thesis takes Tianjin public transport network as the research object.Based on the existing 612 routes and 5435 stations in Tianjin,the basic theory of complex network is applied to construct Tianjin public transport station and route network,analyze the static topology characteristics of both,and make robust analysis of public transport station and route network.Random attack and deliberate attack destroy the connectivity of the network,and the public transport network shows different characteristics to random attack and deliberate attack,showing robustness to the former and vulnerability to the latter.Through the analysis of the static public transport network,it is concluded that the public transport network in Tianjin is both scale-free and small-world,and some public transport transfer stations,namely some key stations,are vulnerable to some extent,which may reduce the stability of the public transport network.Secondly,in order to calculate the destination station based on the passenger travel data,a bus destination station prediction algorithm based on the bus IC card swipe data and bus scheduling data based on the decision tree C4.5 algorithm was designed.Based on the supervised learning of passenger travel data,it can distinguish the categories according to the attributes of the data,thereby predicting the passenger's drop-off station.Taking Tianjin public transport system as an example,using this method to predict the bus to get off station success rate was 75.36%,which is more efficient than other methods.The predicted results can obtain the distribution of OD(Origin and Destination)stations of passengers and provide a strong basis for the study of largescale passenger travel rules.Finally,this thesis takes Tianjin 686 bus line as an example to optimize the line and station,improve the line operation efficiency,reduce unnecessary waste of resources,and provide reference for bus network optimization.