The assortment planning is one of the most important areas for retail and for category managers. The creation of a single assortment requires the analysis of an enormous amount of variables and products. Therefore, for managers, the development of informatics tools to create product assortments is crucial. In this sense, it was developed a simulation model based on a mixed integer nonlinear programming problem to create a proposal to the product assortment at a big Portuguese retailer network. The model uses the customer segmentation to direct the assortment, and also applies the demand substitution parameter in order to determine a more realistic solution. In general terms the tool produces a feasible solution within 2 minutes, and the results compared with the currently adopted solution, show a significant improvement in profit performance category, while maintaining the sales levels. The discrepancy between results is greater the larger the number of assortment possibilities, which increases as the available space on the shelves decreases. Since the increase in the number of assortments leads to an increase in possibilities that the manager needs to analyze, the probability that the solution proposed is not optimal tends to augment.

--

Product assortment, Category management, Simulation, Mixed Integer Nonlinear Programming