Research Of Network Traffic Congestioin Based On Complex Networks Theory

With the rapid development of economy and the deepening of the urbanization, the size of the urban road network is larger and more complex. However, the traffic congestion is becoming more and more serious. To control the degree of congestion, many major cities usually reduce the traffic congestion through widening the road and building new roads, but it is not efficient. Previous studies suggested that the transmission performance of the network is mainly related to the routing process. Therefore, to enhance the network throughput and reduce the traffic congestion that is a key research subject for solving the traffic problem.System science finds out that the function of the system is dependent on its structure.The structure characteristics of the road network are crucial to the traffic flow process.Therefore, to deeply understand the complexity of the structure and function of the road network is very important to discover efficient dynamic traffic guidance strategies in theory and reality. On the basis of this, this dissertation will focus on the structural and functional characteristics of the urban traffic system, study the distribution of gravities of the traffic network and mine the gravitational mechanism during dynamic traffic process, by using the complex networks theory and gravitational-field theory. Then, a traffic gravitational-field model for roads is defined and some dynamic routing strategies based the gravitational-field model are presented. Specifically speaking, the main researching contents and achievements of this dissertation are described as follows:1. Using the complex networks theory, this dissertation studies the dual topological representation of urban road networks, researches the dual graph modeling principle for street-street relationship, and summarizes the strengths and weaknesses of this dual graph.On the basis of this, a dual topology of arc-arc relationship is proposed. This dissertation illustrates the principle of the dual graph and its advantage in dynamic traffic analysis. The dual topology based on arc-arc relationship is helpful to describe the real traffic process and further mine the structural and functional complexity of street networks, so as to improve the accuracy of dynamic traffic simulations.2. From the point of gravitational-field theory, the relation between the structure and function of networks is studied; the gravity distribution of traffic network under shortest-path routing strategy is discussed. It concludes that the traffic flow is always influenced by the gravities of different roads. The results indicate that the congestion distribution and betweenness distribution both considerably reflect the effect of roads on traffic process;these two indices both describe the attraction of roads for traffic flow.3. For further studying the gravity distribution of traffic networks, an evaluation method of road importance is defined with considering the degree, betweenness centrality and closeness of nodes, which reflects the contribution of nodes to the structural and functional characteristics of road networks. On the basis of this, an approach of gravitation analysis for traffic networks is proposed based on the evaluation of road importance and a corresponding gravitation formula of roads is defined. Finally, this dissertation discusses the optimization problem of traffic gravitational-field and obtains the optimization ideas. The results indicate that the proposed evaluation method of road importance effectively describe the significance of roads within the road network and their gravitation strength during transmission process.The attraction of roads for the traffic flow is the key factor that affects the network throughput and traffic congestion, the non-uniform distribution of traffic load can be considered as the result of the non-uniform distribution of roads' gravities. Therefore, this dissertation concludes that the traffic gravitational-field is the potential driving force causing traffic problem, and how to optimize the traffic gravitational-field within dynamic transmission process is helpful to improve the network throughput and reduce the traffic congestion.4. Combined the theories of complex networks and the gravitational field, this dissertation deeply study the dynamic traffic process of networks and define a gravitation model of nodes for the traffic flow with considering the transmission performance and traffic state of the node and its neighbors. The results show that the defined gravitational-field model better describe the attraction mechanism of nodes for traffic flow during dynamic traffic process; the proposed routing strategy considerably improve the transmission performance of networks, balance the network load and reduce the traffic congestion.5. Considering the factors including the betweenness centrality and capacity of nodes,the distance and so on, a traffic gravitational-field model for describing dynamic traffic process is defined and correspondingly the gravitation model for routing paths is constructed.Based on this, from the point of global paths, the global optimal gravitational-field routing strategy is proposed considering the gravities of all paths. Theory and experiment show that this routing algorithm takes all paths into account and make full use of more nodes. The traffic load of many important nodes is reduced. The proposed gravitational-field routing strategy further improves the network throughput and decreases the network's traffic congestion.