DEA灵敏度分析的进一步探讨与应用

针对相对误差对 DEA方法的影响 ,利用线性规划方法 ,提出一种保持 DEA方法有效性分类的模型 ,然后分别得到有效单元和无效单元保持有效性分类的充分和必要条件 .在此基础上给出了最坏条件下 DMU保持有效性分类的充分条件 .最后通过一个测度创新型企业竞争力的例子对其进行灵敏度分析 ,验证所得到的结论 ,从而表明此方法是有效的.

Further Approach to the Sensitivity Analysis of DEA and Its Application

One model of DEA(data envelopment analysis) sensitivity analysis is presented in this paper. Using this model, we developed the sufficient and necessary conditions to keep the classification of being efficient or inefficient when only one DMU changes. Nevertheless, the sufficient condition of keeping classification in the worst case is proposed. In the last section, an empiric example on measuring competitiveness performance of innovative firms is offered to verify the result. Thus, the proposed method is proved feasible and valid; competitiveness.