150 Paper No. 97-0857 TRANSPORTATION RESEARCH RECORD 1644 Intelligent transportation systems (ITS) are being designed to provide real-time control and route guidance to motorists to optimize traffic net- work performance. Current research and development efforts consist of a dynamic traffic assignment capability that can predict future traffic con- ditions and a real-time traffic adaptive control system (RT-TRACS) for generation of signal control strategies. Although these models are intimately connected, so far they have developed independently of one another. A framework is presented here for integrating the two models into a combined system with a practical approach for realizing it. First the static case involving the interaction between travelers (demand) and trans- portation facilities (supply) under recurrent conditions is discussed. This model is applicable in the design and planning of transportation systems management actions. The framework is then extended to the quasi- dynamic and the dynamic cases, which involve incorporation of advanced ITS technologies in the form of advanced traffic management systems and advanced traveler information systems. An innovative application of this framework to advanced traffic-adaptive signal control is presented using the hierarchic structure of RT-TRACS. Traffic engineers have always recognized the fact that performance of the transportation system is the result of a complex interaction between the physical supply of transportation facilities and the indi- vidual trip-makers’ decisions. Supply includes management policies and operational controls that are applied in these facilities. In the parlance of economists, the flows in a transportation network are the result of an “equilibration” process between the demand for trans- port services and the supply of transport capacity. This was the basis for the development of “static” models for designing and planning transportation system management actions (1,2). In this study the static framework is extended to the quasi-dynamic and dynamic cases, where on-line control and guidance are being pro- vided in response to real-time traffic information. The application of advanced technologies in sensing, communications, and computation in intelligent transportation systems (ITS), coupled with advanced modeling concepts, provides great opportunities to improve the per- formance of traffic networks under both recurrent and nonrecurrent conditions. Current research and development efforts are directed toward development of a dynamic traffic assignment (DTA) capabil- ity to predict future traffic conditions and a real-time traffic adaptive control system (RT-TRACS) for generation of signal control strate- gies. Although these models are intimately connected, they have developed independently of one another so far. In the proposed frame- work the two models are integrated into a combined system and a practical approach for realizing it is presented. TRAFFIC MANAGEMENT PROBLEM: STATIC CASE The traffic management problem can be viewed as the interaction between the urban traffic manager and the individual trip maker and their (sometimes differing) perceptions of the performance of, or the supply provided by, the transportation system. A schematic illustration of the interactions in a transportation network is pre- sented in Figure 1. The physical transportation system and the socioeconomic activity system interact, via the equilibration process, to produce a set of flows on the links of the network. Two major feedback loops affect this process. In Loop I, the traffic manager assesses the system’s performance according to his or her measures of effectiveness (MOEs) and intervenes in the physical transportation system to achieve his or her desired objectives (or, more accurately, the objectives of the community he or she repre- sents), such as reducing total travel times, minimizing pollution, and increasing safety. The trip makers, on the other hand, assess the flows according to their own perceptions, which may be dif- ferent from those of the traffic manager, and propagate an adjust- ment in the travel demand pattern via Loop II. Performance of the system is a result of the combined interaction of the demand and the supply feedback loops. The problem can be set in the following context: given (fixed) demands for travel in an urban area and a (fixed) supply of trans- portation facilities, the traffic manager must consider a variety of management strategies to induce a traffic flow pattern that will meet, in an optimal way, the overall objectives of the community. The manager’s criteria for evaluating actions may include public interest measures of performance such as travel time, energy con- sumption, noise, and pollutant emissions. On the other hand, indi- vidual trip makers are assumed to minimize only their own travel costs (usually the travel time), subject to the constraints imposed by the physical system supply, the manager’s actions, and the interactions resulting from other trip makers’ decisions. The prob- lem can be cast as system optimization (the manager’s objective) for flows that result from user optimization (the trip-makers’ objective). This leads to a compound mathematical optimization formulation, as follows: subject to flow conservation constraints P P p ik jk h r ik I a   , , , , ( 29 = ( 29 ( 29 ( 29 ∑ with 2 min arg min M F Z M Z F 1 2 1 , ( 29             ( 29 Integration of Dynamic Traffic Assignment with Real-Time Traffic Adaptive Control System NATHAN H. GARTNER AND CHRONIS STAMATIADIS Department of Civil Engineering, University of Massachusetts, Lowell, MA 01854.