State-Dependent Pricing for Real-Time Freeway Management: Static, Reactive, and Anticipatory

Abstract

This paper addresses the problem of appropriately setting time-varying link tolls to maintain target service levels on toll links by applying a linear feedback controller similar to ones proposed for ramp metering control. Specifically, the deviations between link concentrations and a given set of target concentrations on toll links are calculated, and then link tolls are adjusted by some control regulator. With this method, a set of static link tolls can be established if the historical link concentration data are given; a set of time-varying reactive link tolls can be set if link concentrations are extracted from prevailing network states; while a set of time-varying anticipatory link tolls can be determined if predicted link concentrations are available as well. To examine and compare the performance of static, reactive and anticipatory link pricing strategies, these methods are embedded in a real-time traffic estimation and prediction system (TrEPS), which constantly publishes prevailing and predicted link concentrations. The Baltimore - Washington DC corridor network is used as the test bed network with two lanes of the I-95 southbound being converted to be the toll lanes. Experimental results demonstrate that the proposed method is able to efficiently generate time-varying reactive and anticipatory link tolls and the toll road performance, in terms of throughput, path travel times and link concentrations, and that the anticipatory pricing strategy outperforms the reactive one.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 9 citations exist upstream.