Two-echelon van-robot routing problem with sharing-curbside satellites

High population density and commercial-activity density in urban areas make land use for urban logistics systems even more challenging. Herein, a concept named sharing-curbside satellite (SS) is involved in two-echelon city logistics systems. Traditional vans are deployed in the 1st-echelon network, whereas ground-based robots are employed in the 2nd-echelon network. As a type of nondedicated satellites, the SS shares curbside spaces with the local traffic flow, and each SS can have multiple time windows for direct transshipment between vans and robots. The SS can provide a new mode for urban deliveries through temporary and nondedicated satellites at the neighborhood level. In this study, the two-echelon van-robot routing problem with SSs (2ERP-SS) is defined. The SS synchronization involves vans being used as part of SSs, each SS has multiple time windows, cargoes are transshipped directly between vans and robots, and the available transshipment capacity decreases over time. We develop a mixed-integer linear programming model. We provide a large neighborhood search (LNS) combined with a beam search algorithm, and employ an adaptive LNS (ALNS) for comparison. The effectiveness of the mathematical formulation and heuristics are evaluated through computational experiments, and practical management insights are elucidated.