Imprecise DEA via Standard Linear DEA Models with a Revisit to a Korean Mobile Telecommunication Company

Abstract

Data Envelopment Analysis (DEA) requires that the data for all inputs and outputs are known exactly. When some outputs and inputs are unknown decision variables, such as bounded and ordinal data, the DEA model becomes a nonlinear programming problem and is called imprecise DEA (IDEA). The nonlinear IDEA program can be converted into a linear program by an algorithm based upon scale transformations and variable alterations. Such an algorithm requires a set of special computational codes for each evaluation, because a different objective function and a different constraint with a set of new variables are present for each unit under evaluation. The current paper revisits a published Korean telecommunication analysis, and, by so doing, presents a new and simple approach to execute the IDEA through the standard linear DEA models. This greatly enhances the applicability of IDEA in applications, and the IDEA analysis is no longer limited to obtaining the efficiency scores. The key to the new approach lies in the finding that imprecise data can be easily converted into exact data. Based upon the exact data, models can be developed to determine all possible multiple optimal solutions in imprecise data, and to perform efficiency sensitivity analysis in IDEA.