This paper addresses an operational planning challenge in two-tier robotized sorting systems (T-RSS), an emerging alternative to traditional conveyor-based sorting in e-commerce delivery stations. Designed to be compact and space-efficient, T-RSS use an upper tier to sort parcels from loading stations to drop-off points, which connect to roll containers on a lower tier where parcels are consolidated and transported to shipping docks for outbound trucks. While this architecture supports fast parcel processing, it also tightly couples the two tiers: parcel sorting decisions directly determine roll-container flows and downstream movement to outbound docks, creating a need for integrated planning across sorting and outbound operations. We address this challenge by developing a pathbased sort planning model that jointly determines parcel-to-loading station, parcel-to-drop-off point, roll container-to-shipping dock, and truck-to-shipping dock assignments to efficiently route parcels through the hub. To enable scalability, we propose an offline–then–online solution framework that replaces parcel-level planning with an aggregated, lower-fidelity commodity-flow model in the offline phase, and then uses the resulting sorting paths online to guide real-time parcel routing. Results of computational experiments demonstrate the value of decision integration by comparing fully and partially integrated planning schemes across tiers, and offer practical insights on how to best exploit the operational potential of T-RSS in last-mile delivery stations
Citation
Kim, J., Khir, R. (2025). Optimizing Two-tier Robotized Sorting Systems for Urban Parcel Delivery. Optimization Online.