Storage Assignment Methods Based on Dependence of Items

According to historical customer orders, some items tend to be ordered at the same time, i.e., in the same orders. The correlation of items can be obtained by the frequency of these items present at the same orders. When the dependent items are assigned to adjacent storage locations, order pickers will spend less time to complete customer orders, compared to the other storage assignment which treats items independently. This research provides optimization models to make sure that the storage locations of highly dependent items are nearby. Although the provided nonlinear integer programming can be transformed to a linear integer model, it is still too complex to deal with large problems. For a large number of items, two heuristic algorithms are proposed according to properties of optimal solutions in small problems. Numerical experiments are conducted to show the results of the proposed algorithms with the comparison of the random and class-based storage assignment.