Date of Original Version
Abstract or Description
Tactical planning models for the Fast Moving Consumer Goods (FMCG) industry can quickly become intractable due to the extremely large number of Stock Keeping Units (SKUs). We propose an SKU decomposition algorithm that is aimed at being able to solve cases containing thousands of SKUs. The full tactical planning model is decomposed into a set of single SKU models. These models are then solved sequentially. The capacity used by other SKUs is removed from the available capacity and, at a certain penalty cost, a violation of the capacity is initially allowed. By slowly increasing the penalty cost, the capacity violations are decreased until a feasible solution is obtained. Using the algorithm it was possible to obtain solutions within a few percent of optimality for example cases containing 10 and 25 SKUs. It was also possible to solve a larger 100 SKU case for which the full space model was intractable. The main advantage of the algorithm is that the required CPU time scales approximately linearly with the number of SKUs.