A decentralized scheme for real-time optimization of traffic signals

A real time, highly decentralized, adaptive traffic signal optimization method, ALLONS-D, based on the rolling horizon dynamic programming technique, is presented. We describe the basic architecture, the system model, and the optimization scheme of ALLONS-D. We compare ALLONS-D with other approaches to signal control via several sets of simulation results. The margins of improvement are significant when compared with the standard Webster's criteria for signal setting. The feasibility of on-line implementation of ALLONS-D is discussed as well as on-going evaluations and extensions.