New approach to determine common weights in DEA efficiency evaluation model  

作　　者：Feng Yang Chenchen Yang Liang Liangc Shaofu Du 

机构地区：[1]School of Management, University of Science and Technology of China, Hefei 230026, R R. China

出　　处：《Journal of Systems Engineering and Electronics》2010年第4期609-615,共7页系统工程与电子技术（英文版）

基　　金：supported by the National Natural Science Foundation of China for Innovative Research Groups(70821001)and the National Natural Science Foundation of China(70801056)

摘　　要：Data envelopment analysis（DEA） is a mathematical programming approach to appraise the relative efficiencies of peer decision-making unit（DMU）,which is widely used in ranking DMUs.However,almost all DEA-related ranking approaches are based on the self-evaluation efficiencies.In other words,each DMU chooses the weights it prefers to most,so the resulted efficiencies are not suitable to be used as ranking criteria.Therefore this paper proposes a new approach to determine a bundle of common weights in DEA efficiency evaluation model by introducing a multi-objective integer programming.The paper also gives the solving process of this multi-objective integer programming,and the solution is proven a Pareto efficient solution.The solving process ensures that the obtained common weight bundle is acceptable by a great number of DMUs.Finally a numeral example is given to demonstrate the approach.Data envelopment analysis（DEA） is a mathematical programming approach to appraise the relative efficiencies of peer decision-making unit（DMU）,which is widely used in ranking DMUs.However,almost all DEA-related ranking approaches are based on the self-evaluation efficiencies.In other words,each DMU chooses the weights it prefers to most,so the resulted efficiencies are not suitable to be used as ranking criteria.Therefore this paper proposes a new approach to determine a bundle of common weights in DEA efficiency evaluation model by introducing a multi-objective integer programming.The paper also gives the solving process of this multi-objective integer programming,and the solution is proven a Pareto efficient solution.The solving process ensures that the obtained common weight bundle is acceptable by a great number of DMUs.Finally a numeral example is given to demonstrate the approach.

关 键 词：data envelopment analysis（DEA） common weight RANKING multi-objective programming. 

分 类 号：O22[理学—运筹学与控制论] S758.3[理学—数学]