Model-Based Dynamic Pricing Algorithm for Managed Lanes

Abstract

Managed lanes, or high-occupancy toll (HOT) lanes, have been implemented across the United States to improve the utility of existing road facilities. The pricing algorithm in managed lanes is the critical component in ensuring that the desired level of service metrics is met. Some pricing algorithms currently in use are static algorithms, and others adopt a dynamic strategy. Static pricing algorithms do not account for the changes in real-time traffic conditions. However, dynamic pricing algorithms are reactive in nature and do not account for the potential demand for actual future time intervals for which the toll is determined. The HOT-lane system is nonlinear and complex in nature; this complexity leads to high levels of instability in such a system. The complexity has been increasing because of the growth in demand as well as incorporation of other, noninteracting strategies that are implemented to address this demand. Therefore, more sophisticated designs of the pricing algorithm are needed for HOT-lane pricing control. A model-based pricing algorithm is proposed for determining the toll rate dynamically to maximize the throughput on the HOT lanes while maintaining a desired level of service. It consists of both proactive and reactive components to achieve fast response to real-time changes and maintain a steady, maximal traffic flow. The controller structure also incorporates active bottleneck management to maintain optimal traffic flow under stressed conditions. Implementation of this controller has shown high utilization of HOT lanes while ensuring the desired level of service.