A hierarchical decomposition approach to retail shelf space management and assortment decisions

Abstract

Shelf management is a crucial task in retailing. Because of the large number of products found in most retail stores (sometimes more than 60 000), current shelf space management models can only solve sub-problems of the overall store optimization problem, since the size of the complete optimization problem would be prohibitively large. Consequently, an optimal allocation of store shelf space to products has not yet been achieved. We show that a hierarchical decomposition technique, consisting of two interwoven models, is suitable to overcome this limitation and, thus, is capable of finding accurate solutions to very large and complex shelf space management problems. We further conclude that other important variables (such as product-price) can be included into the methodology and their optimal values can be determined using the same solution technique. Our methodology is illustrated on a real-life application where we predict a 22.33% increase in store profits if our model's solution is implemented.