Ryuichi Tani, Agachai Sumalee, Kenetsu Uchida

    This paper proposed an optimization problem that determines the deployment pattern of dedicated lanes to connected autonomous vehicles (CAVs) considering the stochastic traffic demand and the stochastic traffic capacity. The difference between CAVs and regular human-piloted vehicles (RHVs) is driving behavior. The driving behavior of CAVs is expected to be more standardized than that of RHVs. Therefore, we assume that when the penetration ratio of CAVs increases in the lane flow, the mean lane capacity will increase, and the lane capacity variance will decrease. The mean and the variance of lane travel time decrease when the penetration ratio increases. Following this assumption, the difference in the stochastic properties between CAVs and RHVs is considered in a traffic assignment model. The traffic assignment model is formulated as a variational inequality problem. The network design problem with equilibrium constraints was solved by a simulated annealing algorithm in a test network.

Key words:Connected autonomous vehicles (CAVs)；travel time reliability；stochastic lane capacity；mixed traffic flow

DOI:https://doi.org/10.1080/23249935.2021.1954107

Date:2022-12