Optimização do Algoritmo de Previsão de Vendas numa Cadeia de Distribuição

Patrícia Marina Dantas Vieira de Oliveira

Key information

Authors:

Patrícia Marina Dantas Vieira de Oliveira (Patrícia Marina Dantas Vieira de Oliveira)

Supervisors:

António Pimenta Aguiar (António Pimenta Aguiar); Maria de Fátima Guerreiro Coelho Soares Rosa (Maria de Fátima Guerreiro Coelho Soares Rosa)

Published in

07/11/2014

Abstract

The objective of this thesis is the study and optimization of the sales forecast algorithm used by Auchan Portugal, to perform the automatic replenishment of major consumer products. In order to reach this objective, 5 previously suggested amendment proposals were tested and 3 new alterations were formulated. The conclusion of the study revealed that, of the 5 previously suggested alterations, only one was unsuitable to use in the sales forecast. The other 4 suggestions improved the results of the sales forecast algorithm. The new amendment proposals were also analyzed, having shown improvements in the results. Different methodologies that were used to perform necessities/sales forecasts, using the treatment of statistical data previously observed, were also studied. It was revealed that the methods with best results adopted weighted moving averages to treat the statistical data, in order to achieve the necessities forecast. Finally, the estimation of the economic order quantity (EOQ) was performed. The calculated values were different from the average quantities ordered in reality.