Joint optimal pricing strategy of shared autonomous vehicles and road congestion pricing: A regional accessibility perspective

Abstract

Current studies on shared autonomous vehicles (SAVs) and road congestion pricing (RCP) mainly focus on their independent effect on mobility, and rarely consider the joint impact of the two policies on accessibility. In light of this incentive, this paper establishes a bi-level programming model with multi-objective functions and multi-decision variables to solve the joint optimal pricing problem of SAVs and RCP considering accessibility. Because of the complex relationship between the decision variables and the objective functions, the proposed bi-level programming model belongs to a black-box problem. A machine learning algorithm based on multi-objective Bayesian optimization was developed to solve the proposed model, by taking Jiangyin City as an empirical case study. The results show that the combined implementation of SAVs and RCP plays an essential role in enhancing the performance of land-use and transportation systems. However, this advantageous outcome relies on an appropriately appropriate pricing strategy. Failing to account for the influence of the pricing strategy on implementation efficacy, SAVs and RCP could result in a 22.2 % and 5.0 % degradation in regional accessibility and total flow time, respectively. In contrast, the joint optimal pricing strategy of SAVs and RCP can not only mitigate traffic congestion and improve transportation network efficiency but also optimize urban spatial distribution and promote compact urban development, ultimately achieving a 14.0 % improvement in regional accessibility and 18.5 % reduction in total flow time. Sensitivity analysis reveals that with the increase in SAV market penetration rate, regional accessibility sees a steady rise. Conversely, the total flow time initially escalates before experiencing a decline. The method proposed in this paper serves not only in determining the joint optimal pricing strategy of SAVs and RCP to enhance accessibility, but also in analyzing and optimizing the accessibility impact of other land-use and transportation policies.