Towards a digital twin solution for manual order-picking operations in a wholesale distribution center

The (re)organization of existing and planned warehouse facilities typically seeks to balance system-centric performance indicators (e.g., resource productivity) with customer-focused metrics (e.g., order response time). In pursuit of this objective, there is a growing opportunity to transition from conventional simulation models toward digital twin solutions, which offer enhanced decision-support capabilities. This manuscript focuses on the reorganization of manually executed order picking within a real-world wholesale operation. A simulation-based framework is introduced within a digital shadow environment to optimize the order picking process, following an S-shaped, person-to-goods picking strategy. At the core of the modeling approach is an enriched event graph, which captures the manual picking process at a fine-grained level by representing operational events and their real-time interdependencies. To demonstrate the framework’s effectiveness, numerical results are presented for a typical workday in a major Italian retail cooperative. These results compare alternative control policies, examining their impact on queueing dynamics and picker interference, key contributors to service blocking, resource locking, and starvation within the warehouse’s parallel picking aisles.