Combined dynamic traffic assignment and urban traffic control models

Abstract

Urban traffic control systems evolved through three generations. The first generation of such systems has been based on historical traffic data. The second generation took advantage of detectors, which enabled the collection of real-time traffic data, in order to re-adjust and select traffic signalization programs. The third generation provides the ability to forecast traffic conditions, in order to have traffic signalization programs and strategies pre-computed and applied at the most appropriate time frame for the optimal control of the current traffic conditions. Nowadays, the fourth generation of traffic control systems is already under development, based among others on principles of artificial intelligence and having capabilities of on-time information provision, traffic forecasting and incident detection is being developed according to principles of large-scale integrated systems engineering. Although these systems are largely benefiting from the developments of various information technology and computer science sectors, it is obvious that their performance is always related to that of the underlying optimization and control methods. Until recently, static traffic assignment (route choice) modes were used in order to forecast future traffic flows, considering that the parameters which affect the network capacity are fixed over a given origin-destination matrix. Traffic engineering considers traffic flows as being constant and tries to optimize the control parameters in order to optimize certain parameters and measures of effectiveness. These two procedures, although largely depend on each other and known as the combined traffic assignment and control problem, are usually handled separately. The recent scientific and research developments in the fields of traffic assignment, with the rapid development of the advantageous Dynamic Traffic Assignment models, the new dynamic traffic control strategies and the evolution of ITS tend to modify the way with which networks are being modelled and their efficiency is measured. The current paper aims to present the major findings out of a critical review of existing scientific literature in the fields of dynamic traffic assignment and traffic control. Combined traffic assignment and traffic control models are discussed both in terms of the underlying mathematical formulations, as well as in terms of algorithmic solutions, in order to better evaluate their applicability in large scale networks. In addition, a generic and easily transferable scheme, in the form of a methodological framework for the Combined Dynamic Traffic Assignment and Urban Traffic Control problem is presented and applied on a realistic urban network, so as to provide numerical results and to highlight the applicability of such models in cases, which differ from the standard test networks of the related bibliography, which are usually of simple nature and form. © 2011 Published by Elsevier Ltd.