As the adoption of warehouse automation continues its upward trajectory, “parts-to-picker” order picking systems, epitomized by the Robotic Mobile Fulfillment System (RMFS), are increasingly deployed across diverse enterprises. Considering that current automation systems are economically unfriendly to the traditional warehouse transformation, and other shortcomings such as low space utilization, a separated bin-picking system has been developed and adopted in companies. The key improvement compared to RMFS is the adoption of a novel Automated Guided Vehicle (AGV) equipped with temporary racks, which can pick material bins from traditional racks through a grabbing device without redesigning and alterations of the fixed racks. However, research on related decision-making and optimization is still lacking. To fill this gap, we explore the joint optimization of order sequencing and temporary rack shelving and formulate a mixed integer programming model, taking into account the SKU-based (Stock Keeping Unit) workload balancing among multiple picking stations. To solve the model, we propose an interactive order driven heuristic combined neighborhood search algorithm. Based on the real data from an auto-parts distribution center, the case study is conducted to demonstrate the effectiveness of the proposed method. The proposed method can reduce the number of rack movements by 15.39% and improve the utilization rate of temporary racks by 27.49% compared to the two-step method typically used in practice. Sensitivity analysis is also performed to provide valuable managerial insights for the operational improvement of the separated bin-picking system. Note to Practitioners—This paper is inspired by an emerging “parts-to-picker” order picking system, which can more effectively reduce retrofit costs of warehouses and improve efficiency compared to currently used Automated Storage and Retrieval System (AS/RS) and Robotic Mobile Fulfillment System (RMFS). However, although the system has been applied in practice, there lacks research on related decision-making and optimization at the operational level. In this paper, we investigate the joint optimization problem of order sequencing and temporary rack shelving, considering the SKU-based workload balancing among multiple picking stations. A mixed integer programming model is formulated. Then, to deal with large-scale cases, an interactive order driven heuristic combined neighborhood search algorithm is proposed to improve the computation efficiency. Practitioners can implement the proposed method and algorithm for routine order processing of automated warehouses that use separated bin-picking systems. Based on the sensitivity analysis, we also present some managerial insights, which can provide guidance for practitioners when implementing the operational improvement of the system. For instance, practitioners could prioritize increasing the capacity of temporary racks over that of picking stations if they would like to improve the system’s efficiency.
