Dynamic Modeling And Complicated Characteristics Of Two-Dimensional Road Traffic Flow Based On Cellular Automata

With the deepening of city-oriented development, the urban population and cars increase sharply, traffic problems have become a serious bottleneck to restrict the development of modern cities. Therefore, it is an urgent issue to how to scientifically supervise, guide and effectively control traffic jams. Because the traffic system is a classical self-driven system far from equilibrium, it has highly nonlinear and uncertainty. So, how to combine research methods and multi-disciplinary theories including mechanics, physics, nonlinear science and computer simulation etc., and how to establish reasonable traffic flow models to explore the formation mechanism of the macroscopic features and congestion phenomenon, which are great significance for enhancing the level of the basic theory and application in traffic flow. In this dissertation, based on the cellular automata models of traffic flow, several improved two-dimensional cellular automaton models are proposed to investigate road traffic flow, which include no traffic light effect, multi-accidents effect, slow-to-start effect in four-directional traffic How and city road network by considering OD property. The macroscopic characteristics of traffic How and the formation mechanism of phase transition in traffic jams are studied by numerical simulation and quantitative analysis. In addition, we put forward some valuable suggestions for urban transportation planning, response of sudden traffic accidents and control of traffic jams.The main contents of the dissertation are as follows.I. An improved BML cellular automata model is established to simulate traffic flow in urban road by considering no traffic light effect, the influence of asymmetric numbers of the different direction vehicles and other factors are investigated, and the microscopic mechanisms of traffic jams are also discussed.Based on BML cellular automata model, an extended two-dimensional cellular automata traffic flow model is proposed to describe the effect of no traffic light by adopting random sequences update. The effect of system size and asymmetric numbers of different directions vehicles are studied by numerical simulation. Through analyzing the fundamental diagram of traffic flow and typical space-time patterns, several important conclusions are drawn. The new model can exhibit the first-order phase transition process from freely moving phase to completely jammed phase. However, the critical density of our model is smaller than that in the BML model, which indicates urban traffic without traffic lights is more easily induce traffic jams. The critical density is independent of the system size, only when the width and the length of system are larger than certain critical values, respectively. The asymmetric numbers of different direction vehicles can increase the stability of the system and effectively reduce traffic congestion. The above results can be contributed to the optimization and control of traffic flow.Ⅱ. An extended cellular automata model is constructed to simulate multi-accidents traffic flow by considering the existence of traffic accident in urban road. The influence of the number of traffic accidents and the space distributions of incident location are investigated.Through incorporating the modeling idea of Nagatani, an improved two-dimensional cellular automata traffic flow model is established by considering the existing of multi-accidents in urban road. The influence of the number of traffic accidents, the space distributions of incident location and other factors are studied via numerical simulation. Furthermore, the characteristics of space-time dynamics are also discussed. The simulation results show that with the increase of traffic accident numbers, the congestion areas of system also increase, and traffic jams is more easily occurred. The space distributions of incident location have a significant impact on urban traffic flow. Especially, when two traffic accidents happened in the same lane, the larger gap between two incident locations is more easily reduce traffic congestion. Finally, some suggestions are put forward to improve and solve traffic accidentsⅢ. An extended BML model is proposed to simulate complex characteristics of four-directional traffic flow by considering the effect of slow-to-start, and the macroscopic phenomenon of traffic flow and the microscopic mechanism of two phase coexistence induced by slow-to-start effect are investigated in detail.Based on BML model, an extended BML model is established to simulate complex characteristics of four-directional traffic flow by considering the effect of slow-to-start effect. The influence of slow-to-start effect, traffic light period and other factors are investigated via numerical simulation. Furthermore, the typical features of space-time patterns, phase transition and "deadlock" phenomenon are also discussed. The simulation results show that the effect of slow-to-start plays an important role on traffic flow. Our model can exhibit a novel phase separation phenomenon, i.e., the coexistence of multiple free flow stripes and multi-local jams can be observed, which is in good accordance with the real traffic flow. In addition, the value of maximum velocity Vmax in the moving phase is obtained by using the mean field analysis, which is in good agreement with simulation results.Ⅳ. Through incorporating the idea of OD (Origin and Destination) modeling, an extended traffic flow cellular automata model is established to simulate dynamic route-choosing between residential area and the workspace by considering OD properties in urban roads, the influence of number of the workspaces and the space distributions of two workspaces are studied.Based on the modeling idea of OD model, an improved two-dimensional traffic flow model is proposed to simulate dynamic route-choosing between residential area and the workspace by considering OD properties, which describe real urban traffic flow. The influence of number of workspaces and the location distributions at a certain total area are studied by numerical simulation.. The simulation results show that with the increase of the number n of workspace, the jam critical density decrease, and the traffic jams are more easily occurred. Compared with other distributions, when two workspaces locate the same straight line, the traffic system are more likely to cause traffic jams. In addition, as n=2and θ=0, with decrease of gap between two workspaces, the traffic jams can be effectively alleviated. The above results can be contributed to the reasonable layout of residential area and working area as well as the control of traffic jams.The final chapter of this dissertation is devoted to a summary of the thesis and a prospect of further study of the traffic flows.