Cellular Automata for Traffic Flow Modeling

摘要

In this paper, the authors explore the usefulness of cellular automata to traffic211u001eflow modeling. The authors extend some of the existing CA models to capture 211u001echaracteristics of traffic flow that have not been possible to model using either 211u001econventional analytical models or existing simulation techniques. In particular, 211u001ethe authors examine higher moments of traffic flow and evaluate their effect on 211u001eoverall traffic performance. The behavior of these higher moments is found to be 211u001esurprising, somewhat counter-intuitive, and to have important implications for 211u001edesign and control of traffic systems. For example, the authors show that the 211u001edensity of maximum throughput is near the density of maximum speed variance. 211u001eContrary to current practice, traffic should, therefore, be steered away from 211u001ethis density region. For deterministic systems the authors found traffic flow to 211u001epossess a finite period which is highly sensitive to density in a non-monotonic 211u001efashion. The authors show that knowledge of this periodic behavior to be very 211u001euseful in designing and controlling automated systems. These results are obtained 211u001efor both single and two lane systems. For two lane systems, the authors also 211u001eexamine the relationship between lane chaning behavior and flow performance. The 211u001eauthors show that the density of maximum land changing frequency occurs past the 211u001edensity of maximum throughput. Therefore, traffic should also be steered away 211u001efrom this density region.