A fuzzy ranking approach to data envelopment analysis

One problem with data envelopment analysis (DEA), a prominent evaluation method in social sciences, is its low discriminating power when evaluated units are insufficient or inputs and outputs are too many relative to the number of units. To deal with this problem, we incorporate fuzzy set theory with classical DEA so that a broader aspect of evaluation could be taken into account. We propose a method to encapsulate the efficiencies of an unit in different aspects as a fuzzy efficiency. A fuzzy efficiency is further compared with other fuzzy efficiencies to determine its strength and weakness based on extended fuzzy preference relation. With the strength and weakness of an unit, we aggregate them into a total performance index so that a complete ranking of units is obtained. The method proposed in this paper demonstrates a high discrimination in DEA applications compared to classical DEA.