Efficiency measures in DEA with grey interval data under the hypotheses of data consistency

Some methods have been developed to calculate the interval efficiencies of decision making unit (DMU) in DEA model with interval data, in which there are two shortcomings. One is that the evaluated DMU and referenced DMUs are not be dealt with fairly, as they are not counterparts in locations of inputs and outputs within possible ranges. Another is that efficiency intervals may be too wide to provide valuable information. This paper proposes the hypotheses of data consistency in DEA model. Under the hypotheses, linear programming(LP)models to solve the upper and lower bounds of interval efficiencies are established. Lengths of efficiency intervals under the hypotheses are shorter, which produces more reliable and informative evaluation results and DUMs are deal with more fair.