Scheduling the replenishment of just-in-time supermarkets in assembly plants

In recent years, many OEMs, especially in the automotive industry, have installed so-called supermarkets on their shopfloors to feed parts to assembly lines in a flexible and just-in-time manner. Supermarkets are small logistics areas within the factory where parts are intermediately stored to be transferred, often in the form of presorted kits, to nearby workstations frequently and in small lots. While this greatly alleviates inventory concerns at the assembly line, care must be taken that the supermarket itself always be adequately stocked. In this paper, we tackle the problem of determining when which part types should be taken from central receiving storage to the supermarket in what quantities, such that, on the one hand, shopfloor traffic remains manageable, while, on the other hand, inventory costs are not excessive. We formalize the problem, investigate the computational complexity, and develop a bounding procedure as well as a heuristic decomposition approach. Computational tests show that our procedures work very well on instances of realistic size. Moreover, we study the tradeoff inherent in the problem between delivery frequency and in-process inventory.