Evaluation Based on Pessimistic Efficiency in Interval DEA

In Interval DEA (Data Envelopment Analysis), efficiency in has been proposed and its bounds are obtained from the optimistic and pessimistic viewpoints, respectively. Intervals are suitable to represent, uncertainty of the given input-output data and decision makers' intuitive evaluations. Although the intervals give elements a partial order relation, it is sometimes complex, especially in case of many elements. The efficiency measurement combining optimistic and pessimistic efficiencies in Interval DEA is proposed. They are compared front the view that both of them represent the difference of the analyzed DMU (Decision Making Unit) from the most efficient one. The proposed efficiency measurement is mainly determined by the pessimistic efficiency. The optimistic one is considered if it is inadequate comparing to the pessimistic one. Such a pessimistic efficiency based evaluation is more similar to our natural evaluation and DMUs are arranged as a linear order.