Dynamic Pricing for Managed Lanes: Synthesis of Current Best Practices and Framework for Integration with Connected and Automated Vehicles
Ye, Mengmeng, Computer Engineering - School of Engineering and Applied Science, University of Virginia
Chen, Tong, Civil & Env Engr, University of Virginia
Congestion pricing is an effective way to manage travel demand on transportation facilities. The most common (and fastest-growing) form of congestion pricing in the US are variably priced high occupancy/toll (HOT) facilities which run parallel to non-priced general purpose (GP) lanes. The parallel general purpose (GP) lane density or volume can serve as a predictive tolling component for HOT lane demand in subsequent time periods. This thesis explores the future of dynamic pricing for managed lanes in two parts. First, motivated by the opening of the I-66 Inside the Beltway managed lanes in northern Virginia, this research proposes a combination tolling framework (utilizing both historic and real-time traffic data) for pricing of managed lanes that do not have parallel GP lanes. Second, motivated by connected and automated vehicle (CAV) technology, this research explores a reservation-based tolling system for managed lanes.
This thesis first reviews best practices among existing HOT lane facility operations in the US, synthesizing a series of academic and agency expert interviews which address topics including tolling basis, toll update frequency, toll elasticity of demand, maximum and minimum pricing caps, occupancy restrictions, toll signage, facility access, and incident management on managed lanes which have parallel GP lanes. However, on facilities without parallel GP lanes, existing dynamic algorithms (which rely partly on GP lane metrics) fall short in providing sufficient predictive power. In this thesis, we propose a tolling framework (based on expert insights discussed in best practices) for HOT facilities which do not have parallel GP lanes. This framework considers both user preference for the price-certainty of a time-of-day pricing scheme as well as the flexibility of a dynamic pricing scheme to accommodate real-time fluctuations in traffic demand. Without GP lane metrics, the framework utilizes historical traffic on the managed lanes as a predictive tolling component.
Currently, dynamic tolling schemes are largely reactive to the real-time traffic flow. With the introduction of CAV technology, it is possible to enhance these existing mechanisms of tolling. The second part of this thesis explores a reservation-based dynamic pricing scheme, which allows connected vehicles to reserve the managed lane in advance (at a discount). The pricing scheme is designed to be operational with mixed (connected and non-connected) fleets, as the dynamic toll updates in discrete intervals (much like in current practice). A hypothetic two-lane (one HOT lane and one GP lane) highway system is assumed to demonstrate the conceptual tolling framework.
Simulation results suggest that, compared with traditional dynamic pricing algorithms, the proposed reservation tolling system increases the predictability of the upstream traffic’s demand for the downstream HOT facility, and the tolling scheme is able to handle saturated traffic conditions (by ensuring a reasonable level of service in the managed lane) and keep HOT lane density at a desired density. The system is also demonstrated to be functional with a mixed fleet with a two-tiered second-best pricing scheme. Under most traffic conditions, the proposed tolling scheme is effective in keeping traffic demand (both CAVs and non-CAVs) on the HOT lane near goal density. Likewise, under most traffic conditions, the proposed tolling scheme ensures reservation equity for non-CAV users (non-CAVs choose the HOT lane at nearly the same rate as CAVs, despite CAVs’ ability to reserve in advance).
MS (Master of Science)
HOT Lane, Highway Reservation System, Second-Best Pricing, Predictability
All rights reserved (no additional license for public reuse)