| Traffic transport capacity plays an important role for the development of national economy. The higher transport capacity could improve efficiency, reduce the travel time and promote the circulation of materials. The transportation situation has been an important index to judge the growth potential of a country.The inharmony between the rapid development of the social economy and the slow speed of transportation constrcution has become a serious problem all over the world. Although governments around the world have invested heavily on transportation, the traffic jam phenomena have not been solved fundamentally. The traffic congestion and environment pollution created by gas emission have caused a huge economic loss.How to utilize the current traffic resources adequately, how to use the scientific theory to guide the traffic planning, designing, management and controlling, all of these have become urgent problems now. Thus, the traffic flow theory is emerging as an interdiscipline.Besides the engineering application, the research of traffic flow theory also has the important scientific significance. The transportation system is composed of a huge number of interacting vehicles. The traffic system is a system which is far from equilibrium. The research of traffic flow will help us to understand the laws of complex systems which are far from quilibrium better and promote the cross of the subjects such as statistical physics, nonlinear dynamics, fluid mechanics, applied mathematics, traffic engeering.The contents of the paper are as follows:1. Traffic flow at a signal controlled T-shaped intersection has been investigated by the cellular automata model. The three phase traffic signal is used to control the conflicts of vehicles. Firstly, traffic signal with fixed period and signal phase order is used. The phase diagram and capacity of the system are investigated and compared with that of previous unsignalized model. Simulation results shows that fixed signal phase order strategy does not perform better in some special cases.2. Thus, we introduced another type of signal controlling strategy, i.e., adaptive signal phase order strategy. The phase diagram, the capacity and the average travel time of the system are investigated and compared with that of fixed signal phase order strategy. The simulation results show that the traffic adaptive signal strategy is better than the fixed signal phase order strategy.3. The state of car is updated in parallel in previous two-dimensional cellular automata models. A stochastic version of the Biham-Middleton-Levine (BML) model with random update rule is studied. It is shown that under periodic boundary condition, the system exhibits a sharp transition from moving phase to jamming phase. The intermediate stable phase observed in the original deterministic BML model disappears due to the random update rule. Under open boundary condition, the coexistence of moving phase and jamming phase can be observed. The size of the moving phase is roughly the same under different system sizes. We have presented a mean-field analysis for the moving phase, which successfully takes into account the correlation and produces good agreement with simulation results.4. Effects of randomization on urban traffic dynamics are studied based on the BML model. It is found that the average velocity exhibits a first-order phase transition from moving phase to jamming phase under periodic boundary conditions. The intermediate stable phase identified in the original deterministic BML model disappears with the introduction of randomization. The average velocity in the moving phase and the critical car density decrease as the randomization probability increases. We have developed a mean-field theory which successfully predicts the average velocity in the moving phase. Under open boundary conditions, there are only two phases and the maximum current phase does not occur. The dependence of the average velocity, the density and the flow rate on the injection probability in the moving phase have also been obtained through the mean-field theory.5. Effects of violating traffic light rule on urban traffic dynamics are studied based on the BML model. There are two models according to wether the violators are fixed or not. In model Ⅱ, the violators are selected randomly at each time step. In model Ⅱ, some drivers play the role of violator until the end of simulation after choosing as violator randomly at the beginning of simulation. Although the two models are similar, the results are different.In model I, the intermediate phase could be found at any violator’s ratio pv Simulation results show that the violator increases the average velocity of free flowing phase while decreases the threshold from free flowing phase to jam. However, the threshold does not decrease monotonously with pv. In the free flowing phase, the cars are distributed randomly and homogeneously. Thus the velocity of free flowing phase could be obtained by ignoring the correlation.In model Ⅱ, the intermediate phase could only be found when pv=1. Because the leading car in the jam front is always hindered by cars of other direction. Simulation results show that the threshold decreases with the increase of pv. However, at pv=1, it suddenly increases. Because there are long tails stretching in the upstream direction of the stripes. Once the tails reach and interact with the upstream stripes, jam might be induced. A new kind of configuration with stripe slope different from that of BML model has been found in the free flowing phase. We have developed an analytical investigation which successfully predicts the average velocity in the free flowing phase.6. The BML model is defined on the periodic regular lattice and each lattice site represents a crossing. However, there is a road between two successive crossings in the real city network. Thus a cellular automaton model of vehicular traffic in a Manhattan-like urban system is proposed.The system exhibits three diffierent states, i.e. moving state, saturation state and global deadlock state. A metastability of the system is observed in the transition from the saturation state to the global deadlock state. With a grid coarsening method, vehicle distribution in the moving state and the saturation state has been studied. Interesting structures (e.g. windmill-like ones, T-shirt-like ones, Y-like ones) have been revealed. The effect of an advanced traveller information system (ATIS), the traffic light period and the traffic light switch strategy have also been investigated. |
PDF Full Text Request