Model for Adjustment of Aggregate Forecasts using Fuzzy Logic

Abstract

This research suggests a contribution in the implementation of forecasting models. The proposed model is developed with the aim to fit the projection of de mand to surroundings of firms, and this is based on three considerations that cause that in many cases the fore casts of the demand are different from reality, such as: 1) one of the problems most difficult to model in the forecasts is the uncertainty related to the information available; 2) the methods traditionally used by firms for the projection of de mand mainly are based on past behavior of the market(historical demand); and 3) these methods do not consider in their analysis the factors that are influencing so that the observed behaviour occurs. There fore, the proposed model is based on the implementation of Fuzzy Logic, integrating the main variables that af fect the behavior of market demand, and which are not considered in the classical statistical methods. The model was applied to a bottling of carbonated beverages, and with the adjustment of the projection of de mand a more reliable forecast was obtained.