Ranking Efficient DMUs with Stochastic Data by Considering Ine cient Frontier

Data Envelopment Analysis (DEA) models which evaluate the eciency of a set of decisionmaking units (DMUs) are unable to discriminate between ecient DMUs. The problemof discriminating between these ecient DMUs is an interesting subject. A large numberof methods for fully ranking both ecient and inecient DMUs have been proposed.Through real world applications, analysis may encounter data that are not deterministicor on have a stochastic essence but whose distribution can be dened by collecting data in successive periods and by statistical methods. In this paper, a method for ranking stochastic ecient DMUs is suggested which is based on the full inecient frontier method. Using a numerical example, we will demonstrate how to use the result.