Um estudo sobre análise de eficiências usando análise envoltória de dados sob ambiente difuso

Abstract

Among the methodologies used to measure productive efficiency, Data Envelopment Analysis (DEA) is one of the most commonly used not only in the literature but also in real world applications. Therefore, a deeper study of this methodology becomes a crucial part in helping Decision Making Unit (DMUs) to achieve productive efficiency. Classical DEA methods allow us to measure efficiency, for the set being analysed, and classify the DMUs as either efficient or as inefficient, while also allowing us to verify which efficient cases can be used as reference for the inefficent cases. The use of Classic DEA models require precise data for both input and output values. However, input and output values in the real world are usually imprecise, approximate, or vague. One of the objectives of this work is to obtain a deeper knowledge of the classical DEA methods and Fuzzy Data Envelopment Analysis (FDEA). Another objective is to report a comparitive study between the classic DEA methods and the FDEA methods by applying them to a set containing data from Brazil’s Northeast ports. Our results showed that ports classified as efficient by classical DEA methods, including the ones used as reference within the set, are not always classified as efficient by FDEA methods.