Dynamic integrated freeway traffic signal networks: Problems and proposed solutions

During peak periods in large urban areas, the operation of most signalized networks and freeways is heavily interdependent. Congestion within such networks can be reduced if the control strategies for each component subnetwork are coordinated to reflect this interaction, allowing the various traffic controls to cooperate rather than compete. To date the drastically different operational characteristics of each subnetwork, and their complex interactions, have prevented accurate modelling of integrated neetworks. Consequently, each subnetwork is modelled and suboptimized in virtual isolation, and any potential benefits of integrated control are forfeited. This paper examines the fundamental reasons for this lack of integrated models and explores possible solution approaches. Initially, the operational characteristics of integrated networks are discussed and the ability of current state-of-the-art modelling techniques to represent these characteristics is examined. Subsequently, the fundamental requirements of a model for integrated networks are established more formally and the potential of alternative approaches to meet these requirements is evaluated. Finally, detailed criteria for developing an improved modelling approach for integrated traffic networks are proposed and presented.