Integrating freight transport into first-and-last-mile ridesharing services with modular autonomous vehicles

Modular autonomous vehicles (MAVs), which can be physically connected as a platooned MAV (PMAV) with flexible capacities and cost savings, are widely recognized as effective solutions for shared mobility services. Motivated by this potential, this study investigates a novel first-and-last-mile ridesharing problem with integrated freight requests to address imbalanced demands and reduce overall operational costs simultaneously. To preserve service quality and passenger acceptance, we incorporate mechanisms for passenger requests to specify their willingness to share rides with freight. Due to the separable advantage of a PMAV fleet, passengers and freight can be appropriately allocated to adaptable MAVs in a PMAV by respecting passengers’ travel requirements. To address the problem mathematically, we adopt it as a route-based set-covering (RSC) model. A column generation algorithm is applied to solve the linear relaxation of RSC, and an improved bidirectional label setting algorithm is employed to efficiently tackle multiple pricing subproblems for finding boosted columns. To obtain high-quality integer solutions, an exact branch-and-price (BP) method is finally utilized. To accelerate the labeling process, we design a tailored ranking technique to eliminate symmetrical labels and an enhanced dominance rule to prevent redundant computations. Furthermore, to handle large-scale cases, we develop an adaptive large neighborhood search (ALNS) metaheuristic, strengthened by a problem-specific MAV matching strategy. Extensive computational experiments demonstrate the effectiveness of the methods for finding high-quality solutions in realistic instances from Singapore. The results reveal that the implemented solution can reduce total operational costs by 53.67% in comparison with separate passenger and freight transport.