have been extensively used in studying the performance of traffic control and route guidance strategies. In the literature, various control elements of traffic management systems have been studied individually using simulation. Considerable literature exists, for example, on evaluating design of ramp metering for freeway traffic control. [Payne(1973)] developed a ramp metering regulator and evaluated its performance using a macroscopic traffic flow model. The objective of the control was to minimize the deviation of traffic conditions from some nominal conditions. Feedback optimal control was obtained by solving a linear quadratic cost problem. A simulation study was conducted for a 7.8-mile section of the San Diego freeway in Los Angeles. Similar approaches for ramp metering have also been used recently by [Papageorgiou et al.(1990)], [Stephanedes et al.(1992)], [Shepherd(1993)], [Hellinga and van Aerde(1995)], [Diakaki and Papageorgiou(1995)], etc.
Simulation-based studies have also been conducted for other traffic management strategies such as urban traffic signal controls [Sibley(1985), Yauch et al.(1988)], route diversion [Stephanedes et al.(1989), Barcelo and Ferrer(1995)], variable speed limit signs [Smulders(1990)], and mainline metering [Haboian(1995)]. Common characteristics of these studies are a focus on a single element of the traffic management system and the use of small networks of a particular type (i.e. freeway, arterial, or urban streets).
Only a few simulation studies evaluated integrated traffic management systems. For example, [Reiss and Gartner(1991)] conducted simulation studies to assess the performance of IMIS (an earlier version of INFORM). Nine scenarios were studied by varying the traffic demands and the locations and severity of incidents.
New developments in adaptive traffic control, dynamic traffic assignment, and route guidance systems require more powerful integrated simulation tools for evaluation and design refinement. For example, these systems focus on the collection and utilization of real-time traffic information to adjust signal timing and route advisory. One can hypothesize that their success depends in large part on whether or not reliable information can be obtained and used in time to modify control and guidance to best satisfy the needs of changing traffic conditions. A modeling system capable of realistically simulating the traffic flow in the network and its dynamic interrelationship with the control and route guidance system under a candidate design is a necessity for both the development and evaluation of such systems. SIMLAB is a simulation model designed to meet this need [Ben-Akiva et al.(1995)].
Qi Yang