Differential Congestion Pricing Strategies for Heterogeneous Users in the Mixed Traffic Condition

Zhihong Yao,Yifan Chen,Ziyuan Gu,Yuliang Zhang

doi:10.1155/2022/1829104

Abstract

Congestion pricing is one effective demand management strategy to alleviate traffic congestion. This work investigates pricing schemes for mixed traffic flow systems where the human-driven vehicles (HVs) and autonomous vehicles (AVs) coexist. The emerging and integration of autonomous vehicles can help improve the overall transportation efficiency and safety. Given the coexistence of HVs and AVs in the near future, there is need to adjust the existing traffic management strategies to adapt to the mixed traffic conditions. In this study, congestion pricing is imposed on the HVs and the AVs differently, that is, a distance-based toll to the HVs while a delay-based toll to the AVs. We consider six user groups based on the value of time (VOT) and the vehicle types. Compared with the unified distance-based toll, the advantage of delay-based toll is demonstrated first. Then, a surrogate-based optimization framework, namely the regressing Kriging (RK) model, is formulated. Three pricing schemes are investigated and compared: equity-oriented (EQ), environment friendliness-oriented (EN), and revenue-oriented (RE) schemes. Results show that the RE scheme collects the highest revenues; however, its cost-efficiency is weakened. The EQ scheme reduces the variance in the average travel costs among user groups, thus solving the equity issue.

Highlights

With the rapidly increasing population, traffic congestion has become a major problem, especially in urban areas
We propose a surrogate-based optimization to solve the dynamic pricing problem for mixed traffic with drivers of heterogeneous value of time (VOT). e human-driven vehicles (HVs) are charged by distance while autonomous vehicles (AVs) users pay according to the average link delay. e VOT-based tolls are adopted among different groups. e results of three pricing schemes, namely equity-oriented (EQ), environment friendliness-oriented (EN), and revenue-oriented (RE) schemes are compared
We choose it as the pricing zone. ere are 75 nodes and 218 links in the object network. e information of daily income of residents is extracted from the Victorian Integrated Survey of Travel and Activity (VISTA)

Summary

Introduction

With the rapidly increasing population, traffic congestion has become a major problem, especially in urban areas. Due to the limited land resources and high costs, it is unsustainable to address this issue by merely building new infrastructures. Given this situation, congestion pricing, one of the promising demand-oriented strategies, is widely advocated to mitigate congestion. The first-best pricing is difficult to be implemented because it is unrealistic to charge all links in a network strictly by the marginal cost surplus of traffic [2]. The second-best pricing model was proposed wherein just a subset of links are charged based on the traffic conditions [3]. A branch of congestion pricing practices emerged since 1975, when the area licensing scheme (ALS) was successfully operated in Singapore. Low public acceptance was the main reason which led to the failure of these cases [4]

Objectives

Methods

Results

Conclusion