The Efficiency Research Of Multi-Attribute Dmuand Cross-Efficiency Study On Sample DEA

Data envelopment analysis(DEA) is an important method to evaluate the efficiency of decision making unit(DMU) of the same kind. However, in many cases the DMU may be multi-attribute DMU, then on such issues it encounters difficulties to apply the traditional DEA. At same time, the original DEA cross-efficiency analysis can't base on the sample unit to evaluate. Thus, on the above two issues, the following conclusion is obtained.Firstly, in order to research on the efficiency theory of two attribute DMU, adopting the method that synthesize different attributes production possibility set by using two types of sample DMUs, its Specific production possibility set is structured, and the definition of the relative efficiency of DMU is given. A generalized data envelopment analysis model (Twi-DEA model) and the corresponding conception of the efficiency (Twi-DEA) are presented. The properties of (Twi-DEA)model and the relationship between (Twi-DEA)model and the traditional DEA model are discussed. Furthermore, the projection properties of two attribute DMU and the meaning of (Twi-DEA)model are analyzed.Secondly, the above analysis method is extended to the q(q>2) -dimension, and a series of general analysis methods of the efficiency of multi-attribute DMU and related properties are discussed.Finally, The cross-efficiency analysis method of the original DEA model was extended to samples DEA model, the sample DEA cross-efficiency matrix which have more the preference of decision-maker and the priority method based on sample units are studied.