Urban Traffic Flow Simulation Based On Complex Network

With the development of our social economy and improvement of people’s living standard, the number of motor vehicles is increasing; traffic congestion it brings has become a big problem that each big city must face in China. Urban road traffic network is a complex giant system, complex networks theory as an important tool for studying complex systems received wide attention of scholars at home and abroad. A great deal of research in urban traffic networks have been conducted, but most of these research based on static equilibrium assignment theory, rarely consider the dynamic evolution of traffic flow on the traffic network, almost no consideration travelers’dynamic route choice behavior in travel, which are inconsistent with reality. Therefore, in this thesis, a dynamic traffic assignment simulation model (Green shields model included in the cell transmission model, GS-CTM) which based on dynamic user equilibrium theory can better simulate the dynamic characteristics of the traffic flow and route choice behavior, furthermore, it can also simulate the characteristics of traffic flow including the source of traffic congestion, congestion propagation and congestion dissipation, through which we can analyze the impact of network structure on the performance of traffic flow.In this thesis, it conducts a simulation of traffic flow under different network structures based on the complex network theory and dynamic traffic assignment simulation model GS-CTM, the main research work includes the following aspects:(1) This thesis establishes a GS-CTM dynamic traffic assignment simulation model by introducing the relationship between speed and density of Green shields model into CTM, the key issues such as road resistance and path selection have been studied, and GS-CTM’s algorithm have been proposed.(2) This thesis simulates the performance characteristics of urban traffic flow under different network structures. Firstly, a MATLAB program is written based on the complex network theory and generate randomly several network structures:regular network, small-world network, and random network. Secondly, this thesis simulates the performance characteristics of urban traffic flow under different network structures by a GS-CTM dynamic traffic assignment simulation platform which bases on GS-CTM, and compares the network traffic flow index include average journey velocity of network, average traffic density of network, total network impedance, and total delay time. The simulation results show that random network is a good network structure in three network structures because its highest performance efficiency, followed by small-world network and regular network is the worst. (3) This thesis simulates the evolution of urban traffic congestion under different network structures. Respectively, this thesis analyzes the evolution of traffic congestion under different networks from the perspective of global and local. A global perspective refers to analyze the characteristics including the source of traffic congestion bottlenecks, scale of traffic congestion, and dissipation time of traffic congestion under different network structures; a local perspective refers to analyze impact of some links to performance characteristics of urban traffic flow. The simulation results show that random network can carry a larger traffic demand, has a stronger ability to resist traffic congestion and better network performance.