数据包络分析（DEA）作为组织效率评价的一种重要方法已经被广泛应用于各个领域。然而，不论在理论研究，还是现实应用中，现有DEA方法对决策对象和应用条件存在诸多限制，一定程度上制约了DEA理论本身的发展及其应用范围的拓宽。而实践中,DMU组合的效率评价问题大量存在，而现有DEA方法无法合理有效对其进行处理。因此，本项目拟开展DMU组合的效率评价方法研究，具有重要的科学和现实意义，主要内容有：（A）DMU数量组合的效率评价方法；（B）DMU关系组合的效率评价方法；（C）DMU规模组合的效率评价方法；（D）DMU范围组合的效率评价方法。这些研究一方面提出一些新的效率评价方法，另一方面改进一些现有DEA方法，在一定程度上丰富和完善了DEA理论体系。研究成果拟应用于诸如研发项目选择，营销网络的构建以及供应链内部子系统的费用分摊等决策问题，这些研究将进一步推动DEA方法从理论层次走向更广阔的应用层面。

Data envelopment analysis (DEA) as an important appraoch for performance evaluation and efficiency improvement of organizations, has been widely used in various fields. However, both in theoretical research, and in real-world applications, the existing DEA method has many restrictions on decision making object and application conditions, what restricts the development of DEA theory itself, as well as the widening scope of its application to some extent. However, there are a number of questions of DMU combinations' efficiency evaluation in practice, while the existing DEA method can not reasonably deal with such questions. So this project attempts to carry out the methodological research on efficiency evaluation of DMU combinations, thus it has an important scientific and practical significance, and its main contents are: (A) the methodological research on efficiency evaluation of DMU's quantity combination; (B) the methodological research on efficiency evaluation of DMU's relationship combination; (C) the methodological research on efficiency evaluation of DMU's scale combination; (D) the methodological research on efficiency evaluation of DMU's scope combination. These studies propose some new efficiency evaluation appraoches on one hand, and improve some existing DEA appraoches on the other hand, what enrichs and improves the DEA theoretical system to some extent. These research achievements will be applied to some decision-making questions such as R&D project selection, construction of marketing network as well as cost allocation among internal subsystems of a supply chain, and these studies will further promote the DEA method from a theoretical level to a broader application level.