基于决策单元空间关系特征的DEA方法及其应用研究

Data envelopment analysis (DEA) is an important tool for organizations’ efficiency evaluation and improvement. In traditional DEA framework, decision-making units (DMUs) are assumed to be homogeneous, independent and continuous, while it is difficult for the DMU in real-life world to match all the assumptions because of the existing of space factors. Then, the existing DEA models might not be utilized to deal with these DMUs. In this project, the spatial relationships among DMUs are studied. Firstly, three main spatial relationships will be abstracted from the real problems as following: spatial non-homogeneity, spatial interdependency and spatial discreteness, and the impact mechanism that each type of spatial character on the efficiency evaluation and efficiency improvement will be considered. Secondly, we will construct a set of DEA models for evaluating different type of spatial relationships among DMUs. Then, the feasibility and effectiveness of these models will be demonstrated. Finally, we will apply the proposed models and methods to the cases of air environmental efficiency to verify their effectiveness. This project will correct and improve the traditional DEA methods, which will also promote its application in real cases. Besides, our study can also provide some management supports for measuring the efficiency of DMU based on the influence of spatial factors.

数据包络分析（DEA）是对组织单元效率进行评价和改进的重要研究方法。传统DEA评价理论基于决策单元的同质性假设、独立性假设以及连续性公理，现实中空间因素普遍存在，导致决策单元很难完全满足相关假设和公理，无法使用现有的DEA评价方法进行处理。本项目以决策单元的空间关系特征为研究对象。首先，从现实问题中抽象出几种关键的决策单元空间关系特征，将其刻画分类为空间非同质性、空间相关性以及空间离散性特征，研究其对效率评价和效率改进的影响机理。其次，分别研究基于不同决策单元空间关系特征的效率评价与效率改进方法，构建出系统的DEA模型和方法体系，并论证其可行性和有效性。最后，将模型和方法应用于我国空气环境效率评价及治理，以验证其有效性。本项目将对传统DEA评价体系进行修正和完善，推进DEA方法的理论发展和应用实践，为实际运作中包含决策单元空间特征的效率评价、效率改进和管理决策提供支持。

数据包络分析（DEA）是一种重要的非参数效率评价方法。现实绩效评价工作中，由于决策单元之间可能存在一定的空间关系，DEA理论中关于决策单元同质性的假设往往难以得到满足。本项目以决策单元之间空间关系为研究对象，本着“现实问题刻画”到“理论模型构建”再到“实际应用分析”的研究思路。首先，抽象出决策单元之间的空间异质性和空间相关性两个科学问题，剖析其对DEA理论假设的影响。其次，分别针对决策单元之间的空间异质性和空间相关性特征，构建一系列满足DEA基本假设的评价方法与模型体系，并通过实证数据论证其有效性。最后，将理论和方法研究成果应用于空气环境评价、环境效率、商业银行、国际港口、高等教育等不同的评价领域，验证并修正了相关的理论研究成果。本项目是对经典DEA评价方法的有益补充，相关成果推动了绩效评价理论和方法的研究前沿，拓展了DEA评价模型的应用场景，为管理者的决策提供更加具有现实指导意义的评价结果。