A Methodology for Replenishment of Products on Supermarket Sector

Abstract

This paper presents a methodology for the strategic planning of product replenishment in stores of a supermarket network. The purpose of such strategic planning is to reduce stockouts levels and prevent overstocking, in addition to enhancing logistics service to customers. A quantitative method for predicting time series, Radial Basis Function Neural Networks, is used. After that, we applied the new methodology composed by quantitative methods, based on literature and by qualitative methods, based on the company’s staff day-to-day practices. The association of these qualitative and quantitative methods, very simple and efficient, is the main contribution of this paper. The results were highly satisfactory, reducing the Distribution Center (DC) to store stockouts levels from 12% to 0.7% on average in the hypermarkets, and from 15% to 1.7% in the supermarkets. It is worth emphasizing that the methodology proposed here can be applied to any company facing this challenging forecasting problem.