A DEA Model Based On Group Decision And Its Application

DEA (Data Envelopment Analysis) is a nonparametric method used for appraising multi-input and multi-output matters, and it is outstanding among numerous appraisal methods for its inherent objectivity. However, it hasn't any restraints for weight in practical application, and can't reflect the decision-maker's will or preference. Therefore, how to restrain the weights is paid more and more attention to in DEA model. Under this situationn, this thesis constructs a new DEA model with preference restraint based on group decision throery, which makes the appraisal result agree to the fact much more.The thesis is divided into five parts. The first part outlines the thesis, introduces the background, contents and main innovative points. The second part recounts the basic DEA theories, includeing the basic principles and models, and summarize the research and application actuality of DEA models and theories. The third part studies the theories about how to constrct a new DEA model with preference restraint. Firstly, the thesis points out that the preference DEA model is an important direction in DEA theory, then several familiar preference DEA models and DEAHP model are studied, and it is advanced that the applications of these preference DEA models are confined for their simple preference information forms, and then an intergrating-weight preference DEA model with extensive applicability is put forward. The thesis also indicates that it is necessary to study how to intergrate different preference information forms in group decision. Afterward, this thesisanalyses the conversion relationship among four types of familiar preference information, on the basis of which the intergrating methods of group preference information are studied. Finally, a preferce DEA model based on group decision (G-DEA model) is constructed on the basis of above works, which can make appraisal result form clear order and solves out the problem that the appraisal units with same DEA efficiency are difficult to judge which one is better. The fourth part is the application of G-DEA model. The 12 pharmacy enterprises of traditional Chinese medicine are chosen from Xi'an high-tech industrial development zone as appraised units, and the new model, is used to appraise their technological innovation ability. Finally, the corresponding advices are given. The fifth part summarizes the thesis and clarifies the problems need to be studied further in the future.