The Study Of Traffic Dynamical Behaviors For Complex Networks

The rapid development of information technology represented by the Internet has brought thehuman society into the network age．At present，the structure and dynamics of complex networkshave lately attracted a widespread amount of interest and attentions f in all fields. Congestionphenomenon is a typical dynamical process which takes place on complex networks such ascommunication network and traffic network. Therefore, to find optimal strategies for traffic routingand searching is one of the important issues we have to address.On the base of summarizing and concluding previous research results about congestion andoptimized routing and searching strategy, the issue of congestion and routing strategy are studiedfor a variety of complex network model. Several improved routing strategies and searchingalgorithms are proposed in order to increase the network capacity and optimize the transportationperformance. The main contents and contributions of this dissertation are summarized as follows.Firstly, since global information is usually unavailable in large-scale networks, several localrouting strategies were proposed. An order parameter is used to measure the network capacity bythe threshold value of phase transition from free flow to congestion. Simulation results show thatthe capacity of maximum traffic flow is greatly enhanced by the optimal controlled parameter inthree kinds of routing strategies.Secondly, we propose an evolutionary model for weighted network according to characteristicsof real network. The model gives power-law distributions of degree and weight, which has a higherclustering coefficient, as confirmed in many real networks. The congestion problems on theweighted scale-free networks is investigated based on the local routing strategy using node strengthand node degree, analyzing the problems of the packet traffic flow.Thirdly, consider the deficiency of current routing strategies, an improved local algorithmbased on breadth-first search algorithm and high degree seeking algorithm was present. Thesimulations are made on classical networks models and real networks. Our results indicate that theproposed search algorithm has better search results.