A two-phased heuristic for relation-based item location

•Relation-based item storage approach based on an actual warehouse environment.•Item relationships and order characteristics integrated in item slotting generation.•Linear placement approach is adapted to create the Minimum Delay Algorithm.•Developed algorithm capable of solving industrial sized problems.

The order picking process contributes to the majority of warehouse operation cost due to its high labour content. Improving the picking process begins with the item assignment policy, which allocates items to storage locations. This paper deals with the case where multiple orders/batches contain multiple items that can be picked together in a single route. The concept of linear placement, initially proposed in computer science for designing circuit boards, is adapted to create the “Minimum Delay Algorithm (MDA)” that considers item relationships and order characteristics to generate an item storage layout. The method seeks to minimize the total walking distance of order pickers. MDA was evaluated against other methods (linear placement CLP, Cube-per-Order Index, and Interaction-Frequency-Heuristic) with respect to 30 cases, ranging from small test cases up to a real industrial scenario.