Experimental Study On Chaotic Phenomena Based On Two Traffic Flow Models

The chaos like a bridge connects the determinism and probability, makes us can know the world in a more practical view. The study of chaotic phenomena in traffic flow can promote the development of traffic flow theory and solve some hard problems in the traffic field. The study of chaotic phenomena based on the traffic flow models, can avoid some kinds of complicated factors and get traffic flow that we wanted easily through changing the parameter combinations. Then we can get some conclusions from the experimental results to guide the study of chaos in simulation traffic flow and practical traffic flow.In this paper, chaotic phenomena in two traffic flow models are studied roundly. The two models are Optimal Velocity Model (OVM) and a frequently used nonlinear traffic flow model named Bierley Model. Firstly, the traffic flow time series are generated from the dynamic equations. Then we use phase space reconstruction technique and some kinds of chaos identification methods to identify the states of traffic flows and conclude the changing rules of the traffic flow states with the changing of parameters. At last, we compare the research results of the two models, sum up the common rules, and analyze the reasons why the differences generated.Through the research we discovered that chaotic phenomena exists in the both models. But chaos state is not the ultimate state of the system, and only is a transitional state before the system state changes to the periodic state (quasi-periodic state) or collision state. While studying the influence of the parameter changes on the traffic flow states, we summed up some conclusions that is useful to find the generation conditions of chaotic phenomena in traffic flow. Thus if we can find the numerical conditions, that the system state changes from chaos state to periodic state or collision state, and the controllable factors, then the methods of chaos control in traffic flow will be found. In addition, we also found that the OV model is closer to the real situation of traffic flow, and is more suitable to do the research of chaos in traffic flow. There are some shortages in the Bierley model.