Research Of The Complex Dynamic Properties Of Traffic Flow Induced By Bottlenecks

Traffic is highly correlated with our human society in modern times, and as a result its experiments and theoretical research are regarded as more and more important. Therefore, how could we utilize the limited traffic resources adequately and effectively, how could we plan, design, control and manage traffic by using scientific theory, and how could we further reduce traffic congestion and solve its following environmental problems, all of which have become urgent problems to traffic engineering as well as to our whole society.On the one hand, traffic flow research is not only an engineering problem, but even a physical problem due to its complex dynamic properties. How to describe traffic phenomena by constructing simple and effective traffic flow models, and further make predictions to new traffic phenomenon, have become objective of traffic flow theory for a long time. On the other hand, the traffic system is also a typical self-driven system far from equilibrium, since it shows typical self-driven system properties such as non-linear character, self-organization, and phase transition behavior. Therefore, theory study on traffic requires cross and development of knowledge in statistical mechanics, fluid dynamics, non-linear dynamics and so on, showing profound scientific meaning.Traffic congestion has long been considered obstruction to development of our human society, not only due to the great economic loss caused by it, but also because of the following environmental pollution problems and so on. In real traffic, traffic congestion is briefly caused by traffic bottlenecks; as a result research study on bottlenecks has gradually become one of the pop topics in traffic flow discipline. Of those various traffic bottlenecks, on-ramp systems are very important and also interesting, because of their simplicity and their capacity to reproduce many traffic phenomena. By now many traffic models have been proposed to study these on-ramp systems.Based on synthetical analysis of previous works in on-ramp systems, our paper involves an interesting problem that with fixed total on-ramp flux, how could we assign it to each on-ramp so as to maximize highway capacity? We want to describe traffic evolution in main road and on-ramps by using deterministic NS model, aimed at study of interaction between main road and on-ramp, and also between each on-ramp. Our simulation results show that for two on-ramps systems, with given total on-ramp flux, only assigning a higher proportion of the demand to upstream on-ramp could maximize high way capacity. However, the origin of optimal assignment has not been revealed in previous works. It is shown in our paper that there are seven regions in phase diagram for two on-ramps system, while optimal assignment only exits in one of them. Through analysis on spatio-temporal diagrams of the main road, it is proposed that optimal assignment happens in the critical point of the system, when there is a random walking shock appears on the main road section between the two on-ramps. Any deviation from the critical point will decrease the overall flux of the whole system. Our theoretical analysis is also in good agreement with simulation results.