Complex network-based storage assignment strategy for item communities

Order picking is a labour-intensive process in warehouse operations, and one of the methods to improve picking efficiency is to optimise the storage location of items, known as the storage location assignment problem (SLAP). Most existing SLAP approaches are based on single-item attributes or pairwise item correlations, neglecting the complex interdependencies among multiple items. The objective of this study was to propose a more efficient storage location assignment strategy for picker-to-parts systems to reduce the total order picking distance (TD). Therefore, this study introduced the concept of item community and proposed a complex network-based item communities storage assignment strategy (CN-ICSAS), in which items within each community were strongly correlated and were stored in locations as close as possible. Building upon this, a new SLAP model was constructed, along with a two-stage solution algorithm: network community storage (NCS), which employs complex network clustering to identify item communities and determine initial storage locations, and adaptive variable neighbourhood search with simulated annealing (AVNS-SA), which further minimises the TD. Experiments were conducted using real data and numerical instances to compare CN-ICSAS with existing storage assignment strategies. The results demonstrated that the proposed method significantly outperforms existing strategies in various cases.