Studies On Complex Networks And Traffic Dynamical Behavior

As the rapid development of computers, studies on complex networks become more and more popular in academia in nation and abroad. Basic theories and research methods of complex networks may provide a useful way to study various real systems or the complexity of the networks, and be of great significance in people's work and life. In this paper, combined complex networks with main characteristics of transport networks, by using statistical physics, mathematics and computer simulation methods, we investigate network modeling, load distribution and cascading failures as the three aspects related to network analysis and research.On network modeling, we introduce a combined model for non-directed networks and an asymmetrical model of directed and weighted networks. The combined model considers the attraction of the degree, distance constraints and aging through the evolution of the networks. Simulation results demonstrate that the combined model can evolve into random networks and scale-free networks. The asymmetrical model studies some topology properties by theoretical analysis and numerical simulations, which recover the power law behavior of the distributions of the degree, strength and weight discovered in many real networks.There are plenty of traffic dynamical behaviors on complex networks. In this paper we focus on the load distribution and cascading failures. Based on four typical network structures (i.e., regular networks, random networks, small-world networks and scale-free networks), we investigate the effect of three traffic assignments, i.e., user equilibrium, system equilibrium and system optimization, on the system total cost. It indicates that, with the increase of the traffic volume, scale-free networks and random networks have a better performance to relieve the increase of the system total cost. In addition, simulation results also show that, under the traffic assignments of system equilibrium and system optimization, load distributions exhibit a power-law behavior in scale-free networks, which is similar to that under the traffic assignment of user equilibrium. But the exponents are much larger than that under the traffic assignment of user equilibrium; as a result, the heterogeneity of the load distribution can be decreased under the traffic assignments of system equilibrium and system optimization. Finally, by considering three different traffic assignments, we study cascading failures within four various network structures, which can theoretically support the effective strategies to control the safe of the traffic and transportation networks.