Genetic Algorithms for Traffic Signal Control and Queue Management of Oversaturated Two-Way Arterials

Abstract

The formulation and solution of a dynamic signal control and queue management algorithm for two-way oversaturated arterials are presented. The algorithm is structured to find optimal control and queue management in at least one direction: the primary direction. The other direction is optimized only subject to the fulfillment of the constraints on the primary direction. The main idea of the procedure is to manage queue formation and dissipation through proper design of signal control parameters such that queues are always contained within respective links and that spillbacks are prevented. The two-way control procedure presented accounts for all possible traffic flow regimes that may form as a result of queue management schemes. Genetic algorithms were used to solve the problem. The results show that although the algorithm provides optimal control and queue management in the primary direction, it was also able to manage queues in the secondary direction so that the occurrence of queue spillbacks is prevented. Progression was attainable only in the primary direction. However, the algorithm was able to effectively deal with evolving traffic queues in both directions. In all but one link, traffic was successfully contained in the respective links, thus preventing spillback.