Anticipatory traveler information and pricing for real-time traffic management.

摘要

Anticipatory traveler information and traffic management systems aim to relieve recurrent and nonrecurrent congestion by using predicted traffic conditions to guide tripmakers through the network, and to design control strategies to improve system performance. This study addresses fundamental algorithmic problems associated with providing predictive travel time and reliability information for anticipatory traveler information systems, and regulating dynamic congestion prices for anticipatory traffic management systems (e.g. for managed lane applications).;Anticipatory information is derived from forecasts of network states. When predictions are accurate, anticipatory information is generally expected to be more effective than prevailing information because it accounts for the spatial and temporal changes in traffic conditions. Behavioral studies suggest that traveler acceptance of route guidance information depends considerably on whether such information is experienced to be valid and reliable. Anticipatory traveler information provision is more consistent, and therefore easier to be accepted by travelers, than prevailing information, as it is based on the traffic conditions predicted to prevail at the time the tripmaker reaches a particular location. Similarly, anticipatory traffic management systems utilize forecasts of the traffic conditions and then determine the applicable control strategies based on the prediction. Dynamic congestion pricing is examined as an important application of such management strategy.;To provide consistent traveler information and regulate toll road prices effectively, this study examines (1) travel time prediction, which projects the network states into the near future and calculates the consistent anticipatory route information; (2) travel reliability measure, which captures the probabilistic nature of flow breakdown and provides the reliability information on-line; (3) an anticipatory pricing scheme, which sets the tolls dynamically to prevent traffic breakdown before it occurs; and (4) heterogeneity in users' response to real-time information and dynamic prices, as well as its impact on system performance. The methods are implemented in a real-time traffic estimation and prediction system (TrEPS) and tested on real world networks.