Solving DEA models in a single optimization stage: Can the non-Archimedean infinitesimal be replaced by a small finite epsilon?

The overall assurance interval for the non-Archimedean Epsilon in DEA models; a partition base algorithm

Mehrabian, et. al. (1998) presented a procedure for determining overall assurance interval of ε . Solving n linear programs are needed for this propose, where n is the number of Decision Making Units involved in the evaluation. This paper proposes an efficient algorithm that can determine the overall assurance interval of ε by solving a few number of linear programs. KeywordsData Envelopment An...

Using Non-Archimedean DEA Models for Classification of DMUs: A New Algorithm

A new algorithm for classification of DMUs to efficient and inefficient units in data envelopment analysis is presented. This algorithm uses the non-Archimedean Charnes-Cooper-Rhodes[1] (CCR) model. Also, it applies an assurance value for the non-Archimedean                          using only simple computations on inputs and outputs of DMUs (see [18]). The convergence and efficiency of the ne...

The Uniqueness of the Overall Assurance Interval for Epsilon in DEA Models by the Direction Method

The role of non-Archimedean  in the DEA models has been clarified, so that the associatedlinear programs can be infeasible (for the multiplier side) and unbounded (for theenvelopment side) with an unsuitable choice of  . This paper shows that the overallassurance interval for  in DEA models is unique by the concept of extreme directions. Also,it presents an assurance value for  using only s...

A new method for solving fuzzy DEA models

using non-archimedean dea models for classification of dmus: a new algorithm

a new algorithm for classification of dmus to efficient and inefficient units in data envelopment analysis is presented. this algorithm uses the non-archimedean charnes-cooper-rhodes[1] (ccr) model. also, it applies an assurance value for the non-archimedean                          using only simple computations on inputs and outputs of dmus (see [18]). the convergence and efficiency of the ne...