Integrated Variable Dea Method And Its Application

The DEA theory and method, which is used to appraise the multi-input and multi-output efficiency has been constantly developed and improved in practice. This research is based on the current results of study regarding DEA both in domestic and outside and four aspects about the DEA theory and method have been studied preliminarily in this dissertation. As results of this study, the basis of the DEA theory has been perfected, the validity and veracity of the DEA efficiency measurement have been improved and the DEA method, as an evaluating method, has been promoted to be a decision-making tool. The four aspects are as follows:Firstly, the output-oriented DEA efficiency and the equivalence of Pareto solutions for multi-objective programming has been proved, at the same time the multi-input (multi-output) variables correlation's impact on DEA efficiency value has been discussed.Secondly, in order to solve the problem of the DMU variables correlation's impact on DEA efficiency value, the CCA method of the multi-unit statistics is introduced aiming at the imperfect of CCA-DEA method by Friedman andSinunay-Stern-------Synthesis-variable DEA method. A kind of new method ofefficiency measurement has been put forward and a computing procedure of the new method has been provided by data processing.Thirdly, measuring the efficiency of real samples and making a comparative analysis of the results produced by traditional and improved DEA method proves the validity and feasibility of the Synthesis-variable DEA method.Fourthly, utilizing the Projection Theorem of Traditional DEA method in conjunction with the trait of Synthesis-variable DEA method and adjusting the input-output variables of non-efficiency DMU, the non-efficiency DMU has been made to be an efficient DEA method. As a result, the DEA method, as an evaluating method, has been promoted to be a decision-making tool.