Resilience-based transportation system planning optimization through dedicated autonomous vehicle lanes configuration

During the transition phase when connected and autonomous vehicles (CAVs) and human driven vehicles (HDVs) coexist on the road, it is essential to devise scientific lane management strategies for CAVs in enhancing the operational efficiency of the transportation system (TS). It is worth noting that most current studies do not incorporate the resilience requirements of the TS to effectively respond to seismic events in the planning process. In this study, a novel resilience-based planning optimization methodology through dedicated autonomous vehicle lanes (DAVLs) configuration is proposed in the form of bi-level structure. The optimal configuration scheme for generating DAVLs in the upper level has the optimization objective of minimizing the overall impedance of the TS while meeting the system’s resilience constraints. In terms of the lower level, it quantifies the TS’s functions, fully considering the impact of CAVs on the time value, fuel consumption, and link flow capacity. In order to balance the optimization performance and computational costs, a heuristic algorithm combining genetic algorithms and successive averaging method are integrated to solve the proposed bi-level programming model effectively. On this basis, the proposed methodology adopts a real-world large-scale transportation network regarding deterministic and stochastic earthquake damage scenarios. The sensitivity analysis outcomes show that the market penetration of CAVs and the predefined system resilience threshold have different mechanisms of actions on the optimal configuration strategy of DAVLs and the system performance of TS.