Multi-criteria classification. Approach from the study of neural networks

Abstract

An efficient inventory control system does not treat all items in stock in the same way, but applies methods of control and analysis corresponding to the relative economic importance of each product. This gives rise to a question: ―How do we classify products in stock with an effective and efficient impact on business management?‖ It is becoming general practice to differentiate stock management with dependence on the characteristics of articles in which it is composed, and in literature reviewed, a recommendation for ABC methods of classification based on a quantitative variable or parameter. This paper presents and applies, into a Cuban shipyard, an alternative approach based on the neural networks analysis, in order to classify products in stock, and to support efficient, customer-oriented supply management. The study reveals the theorical relevance and the practical feasibility of the method considered regarding decision-making process.