Abstract:Problem definition: E-commerce giants scale up their order-picking operations by adopting robotic fulfillment centers (RFCs). In RFCs, automatic guided vehicles (AGVs) transport movable shelf racks to pickers’ workstations, instead of having human pickers travel to pick items. Unfortunately, such an apparent relief for pickers turns out to be a curse: Pickers become the bottleneck in the order-picking process. They suffer from high-intensity, stationary, and repetitive work, and consequently, physical and mental problems. To ease this tension, we collaborate with a major e-commerce firm to study how the RFC improves picking efficiency by considering heterogeneous real-time picker performance. Methodology/results: We first propose a distributionally robust human-centric picking performance prediction model (DHPP) to predict two real-time performance metrics of pickers: picking time and fatigue-endurance level. We incorporate probabilistic constraints to address the imperfect information of the random features associated with the picker's behavior. We show that the DHPP can be reformulated into a tractable second-order cone program. Based on the predicted performance metrics, we then propose a mixed 0\textendash1 program to optimize the picker and order assignment. Managerial implications: Our computational study shows that the DHPP model significantly outperforms the state-of-the-art forecasting models regarding prediction accuracy. Our simulation, calibrated with real data from JD.com, shows that our strategy helps to reduce the number of remaining unfulfilled items by 12.5% and improves average pickers' picking productivity by 8.9%. These results suggest that our strategy can significantly improve pickers' welfare by increasing their income and alleviating their mental stress and illness.

Human-Centric Order Picking: Performance Prediction and Robot Assignment at a Robotic Fulfilment Center