广义参考集DEA方法及中国西部地区经济有效性图谱分析

Data envelopment analysis method is an important method of comprehensive evaluation, it has a very wide range of applications on management decision-making. The basic condition that the traditional DEA method is applied is that the evaluated objects must be satisfied the economic axioms of convexity and monotonicity, ect, but we found that a lot of cases of applying DEA method do not meet these economic axioms in fact. Therefore, we plan to restructure the theoretical system of DEA method based on the theory of partially ordered set, and give a more general DEA method with a generalized reference set in this project. Secondly, the reference system of the traditional DEA methods is the effective production frontier, but the objects that people hope to compare with is not limited to "standard of excellence", also may be an ordinary standard (such as the admission line), low standard (such as the sufferable base line) or special standard(such as some samples or standards selected) in fact. At the same time, DEA method dose not break away from the dilemma of randomization, so we plan try to solve some practical problems encountered in the traditional DEA by using "sample units" to structure the" reference set". Finally, by using these methods, this project will draw the system graph of economic efficiency in western region, and analyze the efficiency, non-equilibrium and fluctuation of economic development since implementing the strategy of developing the western China.

数据包络分析方法（DEA）是一种重要的综合评价方法，在管理决策中具有极其广泛的应用。由于传统DEA方法应用的前提是评价对象必须满足经济学的凸性、无效性等公理体系，而在实际应用中我们发现许多DEA应用案例并不满足经济学的这些性质，因此，本项目准备以偏序集理论为基础重构传统DEA理论体系，给出一种应用范围更广的广义参考集DEA方法。其次，由于传统DEA方法的"参照系"是有效生产前沿面，而实际上人们需要比较的对象不仅仅限于"优秀标准"，还可能是"一般的（如录取线）、较差的（容忍度）、特殊的标准（标准或样板）"，同时，DEA本身也还没有走出随机化的困境，因此，本项目准备以"样本单元"为"参照系"，尝试解决一些传统DEA方法在实践中遇到的难题。最后，应用上述方法绘制西部大开发以来中国西部地区经济有效性谱系，并以此分析中国西部地区经济发展的有效性、均衡性和波动性等特征。

数据包络分析方法（DEA）是一种重要的综合评价方法，在管理决策中具有极其广泛的应用。由于传统DEA方法应用的前提是评价对象必须满足经济学的凸性、无效性等公理体系，而在实际应用中我们发现许多DEA应用案例并不满足经济学的这些性质，因此，本项目从偏序集理论出发系统阐述了DEA有效与偏好理论之间的关系，给出了一系列基于偏序集理论的DEA模型和性质。这些方法比传统DEA方法的含义和应用范围更加广泛。其次，由于传统DEA方法的"参照系"是有效生产前沿面，而实际上人们需要比较的对象不仅仅限于"优秀标准"，还可能是"一般的（如录取线）、较差的（容忍度）、特殊的标准（标准或样板）"，同时，DEA本身也还没有走出随机化的困境，为了解决这些问题，本项目以"样本单元"为"参照系",对广义数据包络分析方法体系给出进一步拓展和完善，包括给出基于面向产出的广义DEA模型体系，面向输入及输出广义DEA模型的性质探讨，并进一步延伸了广义DEA模型体系。最后，应用上述方法探讨了DEA方法在中国中西部地区经济社会发展中的应用问题。