网络DEA理论方法与应用研究

发布时间：2018-03-04 00:39

本文选题：数据包络分析　切入点：网络结构　出处：《中国科学技术大学》2016年博士论文　论文类型：学位论文

【摘要】：在实际生产活动中,许多决策单元往往具有网络生产结构,比如银行、医院、保险公司、机场、铁路等等。对这些决策单元的效率求解以及对这些决策单元进行排名一直受到海内外学者以及企业决策者的关注。数据包络分析(data envelopment analysis, DEA)方法,作为一种非参数的方法,有无需预先设定生产函数的参数以及无需预先设定投入产出权重等优点。因此,自1978年被提出以来,DEA方法受到越来越多学者的关注,同时也在许多领域被广泛应用。本文以具有网络结构的决策单元为研究对象,以现有的网络DEA模型为基础,研究具有网络结构的决策单元的排名区间问题以及网络结构效率的效率评价问题,并且将网络DEA方法应用到环境效率以及并购效率评估领域。因此,本研究有效拓展了网络DEA理论并推动了相关应用研究的深入。本文一共分为六章,主要内容总结如下：第一章是绪论。首先,我们介绍了有关DEA的基本理论：即投入和产出、决策单元、生产可能集、生产前沿面和效率等基本概念；使用DEA方法做效率评价的基本思路；一些最基本的DEA模型以及DEA方法的应用。然后,引入本文的主要研究领域-网络DEA的理论方法以及应用。我们介绍了网络DEA理论的研究意义、研究现状以及应用领域。最后,我们概括了本文的研究方法、研究内容和研究意义。第二章是基于两阶段关系模型的环境效率评价方法。在实际的生产过程中,决策单元往往具有两阶段的网络结构。比如说,中国区域工业系统由生产子系统和污染物处理子系统两个串行子系统构成。针对此种具有网络结构的决策单元的环境效率评估问题,本文将两阶段关系模型引入环境效率评估,构建出新的基于两阶段关系模型的环境效率评估模型。中国各个省份的工业系统的算例验证了提出模型的合理性。第三章是考虑决策者偏好的基于网络SBM模型的环境效率评价研究。中国将资源节约和环境保护写进法律,目标是逐步构建资源节约型、环境友好型社会。在中国,资源节约和环境保护同等重要。本研究考虑每个区域工业系统的内在网络结构,以及中国的资源利用和环境保护的政策,同时弥补了以往使用网络DEA模型评估环境效率时使投入和产出仅仅径向投影的缺陷。本研究提出了一个新的网络SBM模型(NSBM)。这个模型考虑了投入产出以及中间产品的非径向投影以及中国现有的资源环境政策,测度中国区域工业系统的环境效率,并做了实证分析。第四章是基于两阶段排名区间的分析方法。传统的DEA模型在评估每个决策单元的效率值时仅仅寻找其最优的权重,Salo and Punkka(2011)为了解决这个问题,提出了考虑所有可行的权重的一系列混合整型模型。但是他们将每个决策单元视为一个“黑盒子”而忽视了决策单元的内部结构。针对此问题,本文提出了基于两阶段排名区间的分析方法,这种新的方法考虑了决策单元的网络结构,不仅计算了总系统的排名区间而且计算了两个子系统的排名区间。并且,计算出的排名区间可以用来分析总系统和两个子系统排名的稳定性。我们用了两个算例来演示模型。第五章是非合作博弈两阶段生产系统DEA并购效率评价。如何选择并购对象及预测并购可行性是企业管理者在并购决策过程中常面临的问题。针对企业并购预测问题,本文以两阶段生产系统的决策单元为研究对象,研究数据包络分析方法在预测并购可行性方面的应用。该生产系统有两个明显的特点：(1)决策单元由两个阶段串行子系统构成,(2)两个子系统一个处于主导地位,而另一个处于从属地位。两个或者两个以上决策单元并购为一个虚拟决策单元。本文在分析由决策单元整体效率与其子系统效率之间关系的基础上,引入非合作博弈的思想,提出了基于非合作博弈的DEA模型评估虚拟决策单元两阶段生产系统的并购效率。文章提出的这种方法能够在考虑子系统之间的主从关系的基础上,有效地分析虚拟决策单元内部子系统的并购有效性水平。第六章是总结和展望。首先对前面各个章节的内容进行总结、归纳；然后列出本文的主要创新点和不足；最后根据本文的不足之处指出了几个未来可以研究的后续问题。本文的创新性总结为以下几个方面：(1)本文首次将两阶段关系模型引入环境效率评估,构建出新的基于两阶段关系模型的环境效率评估模型,并且将两阶段关系模型从规模报酬不变的情形扩展到规模报酬可变的情形,启发式搜索的方法(heuristic search procedure)被用来求解新提出的模型。(2)与以往将每个地区的工业系统视为一个“黑盒子”不同,我们将中国每个地区的工业系统分为生产子系统和污染物处理子系统。考虑每个区域工业系统的内在网络结构以及中国的资源利用和环境保护的政策,本研究提出了一个新的网络SBM模型(NSBM)测度中国区域工业系统的环境效率,这个模型考虑了每个子系统中投入和非期望产出的相对重要性来体现决策者偏好或中国的资源环境政策。(3)在本研究中,通过考虑决策单元的内部结构扩展了Salo and Punkka(2011)的方法。我们将它们的方法扩展到两阶段生产系统,计算两阶段生产系统的排名区间。然后,我们通过使用已有文献中的算例验证了这种方法。因此,通过这种方法,决策单元的“黑盒子”被打开了。这样,可以为决策者提供更多关于总系统和两个子系统排名区间的信息。(4)为解决具有网络结构的决策单元内部各子系统之间可能存在主从的非合作博弈关系的并购效率评估问题,本文将非合作博弈思想引入,首次提出了两阶段的非合作博弈并购效率评价模型。这种方法能够有效地处理具有主从博弈子系统的网络生产系统的并购效率评价问题。
[Abstract]:In actual production, many decision-making units often have network structure of production, such as banks, hospitals, insurance companies, airports, railway and so on. The efficiency of solving these decision unit and has been ranked by scholars at home and abroad and business decision-makers attention to these decision unit. Data envelopment analysis (data envelopment, analysis, DEA) the method, as a nonparametric methods, there is no need to set parameters of the production function and without presetting the advantages of input and output weights. Therefore, since 1978, the DEA method is put forward, more and more attention of scholars, but also is widely applied in many domains. In this paper, a decision unit has a network structure as the research object, the DEA network model based on existing efficiency ranking interval problems with decision making unit network structure and network structure efficiency evaluation The price problem, and the DEA method is applied to the network environment and the efficiency of M & A efficiency evaluation field. Therefore, this study extends the network DEA theory and promote the further application of the research. This paper is divided into six chapters, the main contents are summarized as follows: the first chapter is the introduction. Firstly, we introduce the basic theory about DEA namely: input and output, decision-making units, production sets, the basic concept of production frontier and efficiency; basic idea of using DEA method to evaluate the efficiency of application; some of the most basic DEA model and DEA method. Then, the main theoretical research field is introduced into the network and the application of DEA. We introduce the significance of DEA network theory, research and application. Finally, we summarize the research methods, research content and research significance. The second chapter is the two stage model based on Relationship The environmental efficiency evaluation method. In the actual production process, decision making units often have network structure of two stage. For example, China regional industrial system by the production subsystem and pollutant processing subsystem of two serial subsystems. In view of this evaluation with decision-making units of the network structure on the environmental efficiency of the two stage the relationship between environmental efficiency evaluation model, and proposes a new evaluation model of two stage model of the relationship between the environmental efficiency based on industrial system. Chinese all provinces of the examples to verify the rationality of the proposed model. The third chapter is to consider the preference of efficiency evaluation of network environment based on the SBM model. China will conserve resources and protect the environment written into law, the goal is to gradually build a resource-saving and environment-friendly society. In Chinese, resource conservation and environmental protection are equally important. This study considers each The internal network structure of regional industrial system, policy and resource utilization China and environmental protection, and remedy the defects of the previous use of DEA network model to assess environmental efficiency of input and output only radial projection. This paper proposes a new network model SBM (NSBM). This model takes into account the non radial projection the input and output of the intermediate products and China existing resources and environmental policy, measure China regional industrial system of environmental efficiency, and do empirical analysis. The fourth chapter is the analysis method of two stage ranking based on interval weight. The traditional DEA model to find the optimal efficiency evaluation in only when the value of each decision unit, Salo and Punkka (2011) in order to solve this problem, put forward a series of mixed integer model all feasible considering weights. But they will each decision unit is regarded as a "black box "While ignoring the internal structure of the decision-making unit. Aiming at this problem, this paper proposed two stage analysis method based on interval number, this new method considering the network structure of decision making units, ranking interval of the two subsystems of the total system and calculate the ranking interval is calculated. And the calculated ranking interval can be used to rank the stability analysis of total system and two subsystems. We use two examples to demonstrate the model. The fifth chapter is a non cooperative game two stage production system DEA acquisition efficiency evaluation. How to choose and purchase and acquisition target prediction feasibility is the enterprise managers often face in making decision for enterprise problems. M & a prediction problem, the decision making unit two stage production system as the research object, study on the application of data envelopment analysis to predict the feasibility of mergers and acquisitions in the production system. There are two distinct characteristics: (1) the decision unit is composed of two stage serial subsystems, (2) the two subsystems of a dominant position, and the other one is in a subordinate position. Two or more than two mergers and acquisitions decision unit as a virtual decision unit. Based on the analysis of the relationship between the overall efficiency and efficiency of decision making units on the subsystem, the introduction of non cooperative game theory, put forward the evaluation of virtual decision-making unit two stage production system of M & A efficiency DEA model based on non cooperative game. The method proposed in this paper can considering the master-slave relationship between the subsystems, effectively analyze the level of effectiveness of mergers and acquisitions virtual decision-making unit internal subsystem. The sixth chapter is the summary and outlook. The chapters of the content of the summary, induction; main innovations and shortcomings of this article and list; according to the The inadequacies of this article points out the future problems for further research. This paper summarizes the innovation as follows: (1) the two phase relation model into the environmental efficiency evaluation, construct a new two stage model based on environmental efficiency evaluation model, and the two stage model from returns to scale the same situation is extended to the case of variable returns to scale, heuristic search methods (heuristic search procedure) was used to solve the proposed model. (2) and the industrial system in each region is regarded as a "black box", we will Chinese industrial system in each region is divided into sub production processing subsystem system and pollutants. Considering the internal network structure of each region of the industrial system and policy Chinese resource utilization and environmental protection, this paper proposes a new network model SBM Type (NSBM) measure China regional industrial system of environmental efficiency, this model considers the relative importance of inputs and undesirable outputs of each subsystem to reflect the environmental policy preference or China. (3) in this study, by considering the internal structure of the decision-making unit extends Salo and Punkka (2011) methods. We will expand to two stages of their method of production system, calculation of the two stage production system ranking interval. Then, we use the example to verify this method. Therefore, by this method, the decision unit of the "black box" was opened. In this way, can provide more information about the total system and two subsystems of ranking interval for decision makers. (4) to solve with possible non cooperative game between master-slave decision-making units subsystems in network structure of the M & A efficiency rating To estimate the problem, this paper introduces the non cooperative game theory, and first proposes the two stage non cooperative game M & A efficiency evaluation model. This method can effectively handle the M & A efficiency evaluation problem of the network production system with master slave game subsystem.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F224

【相似文献】