Framework and Overall Structure of TMS

Advanced traffic management systems that dynamically monitor and control network traffic flows have been the focus of recent transportation research. A fundamental characteristic of dynamic traffic management systems is their ability to react to current or anticipated traffic conditions in real-time. A widely accepted approach on how to provide the above functionality does not exist. For this reason a structure that facilitates the representation of both reactive and proactive systems is employed to allow for experimentation. This structure can represent different designs of dynamic traffic management systems with varying levels of sophistication.

  
Figure 4-1: Generic structure of dynamic traffic management systems

Figure 4.1 illustrates the main components of TMS and their interactions with MITSIM. The role of the network state estimation is to obtain the best estimate of the current network state utilizing the data obtained from the surveillance system. Once the state of the network has been estimated, the generation of the control and route guidance strategies can be done in two general approaches: reactive or proactive. The reactive approach consists of pre-determined rules or control laws that depend only on the current network state. Typical examples of such systems are the ALINEA ramp metering model by [Papageorgiou et al.(1990)] and the CA/T incident response plans []Cather:1996. For the proactive case, the structure can represent a system that is able to: (i) utilize the real-time traffic information provided by the surveillance system; (ii) predict future traffic conditions; and, (iii) optimize traffic control and routing strategies.

The majority of the proactive systems proposed in the literature (see Chapter 1 for more details) are based on a sequential approach which uses an iterative process to generate control and route guidance. Examples of such systems are [Gershwin et al.(1978)]; [Gartner et al.(1980)]; [Reiss and Gartner(1991)]; and [Papageorgiou(1980)]. Most recently, the work by [Mahmassani et al.(1994), FHWA(1995)] and [Ben-Akiva et al.(1996), Ben-Akiva et al.(1997)] attempts to develop proactive routing and/or control strategies using an iterative process. In this approach, a control and routing strategy which accounts for predicted future traffic conditions is generated. Given a proposed strategy, traffic conditions on the network are predicted and the performance of the candidate strategy is evaluated based on some convergence test. One of two actions are taken based on the evaluation: (i) if a satisfactory strategy has been identified, the strategy is implemented; or, (ii) if additional strategies need to be tested, another generation-prediction iteration is conducted.

The structure of TMS, as shown in Figure 4-1, is designed to support evaluation of the traffic management systems that are based on the above sequential process. The main functions that are supported by TMS include route guidance, traffic control, and incident management.