• Dynamic pooling of workforce and AMRs mitigates workload imbalance across hubs. • Human–robot asymmetries enable collaboration, with AMRs complementing humans. • Rolling-horizon stochastic program adapts capacity to uncertain parcel inflows. • Sequential planning integrates intra-hub scheduling and inter-hub relocation. • Enhanced Benders with relax-and-fix heuristics scales to megacity-sized networks. Parcel logistics services play a crucial and expanding role in global economies. In urban parcel logistics hubs, resources support relay-based activities and ensure parcels are sorted, consolidated, and dispatched onto outbound trucks by specific deadlines. In light of labor shortages in logistics hubs, recent developments have introduced autonomous mobile robots (AMRs) to enhance operational efficiency, enabling carriers to meet time-definite customer service guarantees. Emerging market dynamics, driven by increasing customer expectations for swift delivery across diverse product categories, address the importance of dynamic resource deployment to meet uncertainties of parcel arrivals and resource availability in urban hub networks. This paper focuses on a novel operational problem: dynamic multi-location, multi-resource scheduling and deployment in an urban parcel logistics network. It formulates this novel problem as a stochastic optimization problem and presents its main features. Given the inherent stochastic-combinatorial setting of the problem, the solution methodology proposes a reformulation in a two-stage stochastic model. It develops a rolling-horizon framework to solve it sequentially with updated states and new observations. It also designs a solution approach based on Sample Average Approximation (SAA), Benders decomposition enhanced by acceleration methods, and relax-and-fix heuristics, which can solve large-scale instances of a real-sized megacity. Numerical results, inspired by the case of a large parcel express carrier in China, are presented to evaluate the computational performance of the proposed approach. Our results indicate potential savings of approximately 19% in costs and 22% in resources compared to a static resource deployment strategy. Also, necessary workforce-robot hybrid resource deployment yields the highest cost saving, surpassing the workforce-only and robot-only scenarios by 5% and 3%, respectively.
Xu et al. (Fri,) studied this question.