Adaptive routing considering delays due to signal operations

This work addresses the problem of determining optimal routing decisions in signalized traffic networks, where arc travel times vary over time and are known only probabilistically (i.e. in stochastic, time-varying (STV) networks) and additional delay due to signal operations is explicitly considered. While prior works in the literature address problems of routing in STV networks, none explicitly considers the additional delay that would be incurred due to signal operations at the intersections of the roadway network. In this paper, we consider an adaptive routing problem, where paths are adapted en route based on revealed information concerning the arc travel times and actual signal timings. We first discuss how concepts from existing procedures can be combined to solve the adaptive routing problem in signalized STV networks, where the signal timing plan and actual timings are known a priori. When actual timings or delays due to signal control are known only probabilistically, such techniques will be inefficient. Thus, we propose a more efficient algorithm for solving this latter problem. Results of numerical experiments conducted on a realworld-based signalized street network are presented. These results show that the solutions obtained by explicitly considering delays due to signal operations will likely be significantly different from those solutions generated by techniques that ignore such delays. 2003 Elsevier Ltd. All rights reserved.