Abstract: Scientifically planning dedicated lanes for automated trucks within a transportation network is crucial for ensuring safe and efficient freight transportation using automated trucks. This paper investigates a new optimization problem in designing a dedicated automated truck network, considering traffic flow variations across different time periods within the transportation network. The study proposes dedicated lane setup schemes and automated truck transportation networks for different time periods, ensuring rapid and efficient automated truck transport while minimizing the negative impact of the dedicated lanes on regular traffic. Firstly, the problem is formulated as a mixed-integer nonlinear programming model and subsequently transformed into an equivalent mixed-integer linear programming model. To solve it effectively, a tailored adaptive large neighborhood search (ALNS) algorithm with real-number encoding is developed to address the problem’s characteristics. Six tailored operators, such as the repeated proportional destroy operator, are designed to enhance its effectiveness. Finally, extensive numerical experiments show that, in small-scale instances, the average difference between the optimal solutions and those obtained by the tailored ALNS is only 1.40%, and the computation times are all less than 4 seconds. For large-scale instances, compared with the ALNS considering only random destroy and repair operators, the tailored ALNS obtains better solutions, with an average improvement of 8.46%, confirming the effectiveness of the designed operators. Additionally, considering time-varying road conditions, as opposed to ignoring them, reduces negative impacts by an average of 1.36%, verifying the necessity of including time-varying road conditions in the study.

Key words: automated truck, dedicated lane network, time-varying road conditions, route planning, mixed integer programming, adaptive large neighborhood search