The Research Of Intersection Signal Timing Optimization Based On The Ordered Weighted Averaging Operator

Intersection is one of the most fundamental elements to the urban road network. Vehicles’ diverging, emerging and crossing behaviors at intersections are main factors restricting road network capacity. For the sake of traffic management and local (or regional) traffic optimization, signal control is a commonly used method. The key of signal control is defining the signal timing plan, which directly affects road network capacity. Therefore, it is worth researching signal timing. This paper develops an ordered weighted average-operator-based (OWA) signal timing optimization model.Firstly, this paper defines basic parameters and process of signal control, introduces OWA theory and illustrates three traditional signal timing methods, i.e. TRRL method, ARRB method and HCM method. Based on existing researches, an OWA-operator-based signal timing optimization model is proposed. According to the definition of OWA operator, this method sorts signal timing cycles by the three traditional methods in a descending order, and then gives a weighted computation of the cycles by any of the following method: entropy-maximization-based subjective weighted method, normal-distribution-based objective weighted method and average-value-similarity-based objective weighted method.Secondly, this paper presents an empirical study of OWA-operator-based signal timing, given the observed data at the intersection of Ronghua Bridge. Besides, simulations of traditional and optimized signal timing method by VISSIM are conducted, and a comparison among evaluation index by factor analysis is given in order to find out advantages and disadvantages of these signal timing methods. Results show that OWA-operator-based signal timing optimization method is significantly better than traditional methods. Average delay and queue length are optimized, and the efficiency is improved.Finally, this paper uses entropy-maximization-based subjective weighted method to investigate signal timing plans at different orness levels. In this way, an orness level interval is determined, which obtains the better timing plans.