AN APPROACH ON NETWORK TRAFFIC SIGNAL CONTROL UNDER THE REAL-TIME AND OVERSATURATED FLOW CONDITION

Signalized intersections with more than one over-saturated approach present a challenge for traffic engineers when determining the most appropriate apportion of the green time to each approach. This paper applies queuing theory analysis and mathematical approximation techniques to formulate the signal timing problem for oversaturated intersections as a dynamic linear programming model. The objectives of this study was to develop a real-time optimization model which could most effectively allocate the green time to each approach of a signalized intersection under any traffic condition. In this paper, we present an interesting and effective approach which deals with the problem following a so called inverse cause-and-effect procedure. In the formulation model, the traffic flows are presented as smoothed cumulative functions as if there were no interruption from the signals. Instead of optimizing some specific intersections, the objective is to maximize the total vehicular output from the network subject to prevailing capacity and operational constraints. Then, the optimized value of departure flow rate (effect) is converted to the value of split (cause) according to their interrelation. The expected benefit from this research comes from its ability to constantly modify the value of splits to most effectively accommodate changing traffic conditions over the network