Análise do uso da modelagem bootstrap aplicada a gestão de estoques: estudo de caso em uma distribuidora de alimentos

Abstract

Stock management models disseminated in the literature presuppose normal distribution. However, depending on the case, this premise may not be respected, given the distribution of the data, compromising the applicability of the model to real problems. Thus, this study proposes a adapted stock management model based on continuous review system, using the boostrap statistical model to define the safety stock. In order to ascertain the applicability, both models (traditional and adapted) were subjected to simulation with data from a food distributor located in the state of Rio Grande do Norte, considering the years 2018 and 2019 and a 99% confidence level. To compare the results of the models, five performance indicators were proposed, being them, stock level, periods with ruptures, total amount of ruptures, number, number of orders and total cost. The results showed that the traditional model has a higher capacity to meet the demand, generating less ruptures, however it maintains a higher stock level, generating a high storage cost (% more than the adapted model). On the other hand, the model adapted with the bootstrap maintains a level of inventory closer to the real, but presents a bigger failure in relation to stock ruptures. So it is interesting to aggregate the knowledge generated through the comparative of these models, to generate expertise in the practical application of stock management.