矩阵型网络DEA模型、方法及应用研究

发布时间：2018-02-10 10:41

本文关键词： 网络DEA 矩阵型网络生产可能集多目标规划系统效率子系统效率实证研究　出处：《吉林大学》2014年博士论文　论文类型：学位论文

【摘要】：效率评价在经济社会生活中是一项非常重要的任务，同时也是一项有一定难度的任务。本文研究的就是如何对经济社会生活中决策单元的效率进行评价，采用的是基于数据包络分析的方法。数据包络分析（DEA）是评价具有可比性决策单元之间相对效率的数学规划方法，从1978年第一个DEA模型建立以来，DEA方法已逐渐发展为相对效率评价的最重要的方法之一。传统的DEA方法将决策单元视为一个整体，不考虑决策单元内部的投入产出关系，这种方式无法有效的反映出决策单元真实的效率。本文研究的目的就是将决策单元打开，，对决策单元内部进行分析，提出有效的能够计算其效率值的DEA模型，使得DEA方法能够推广到复杂系统的效率评价方面。矩阵型网络系统是目前经济社会常见的一类复杂系统，比如循环系统，基于生命周期的供应链系统等，这类系统是指没有开始子系统和结束子系统的一类网络系统，它的每个子系统都具有外部投入和产出，每两个子系统之间都存在双向投入产出的关系。本文针对矩阵型网络系统的结构特点，构建更为科学有效、符合管理实践需求的矩阵型网络DEA模型。根据研究方法的不同，本文的主要内容可分为五章，主要内容为：矩阵型网络系统的结构分析、NCCR模型的建立及分析，NCCGSS模型的建立及分析、NSBM模型的建立及分析、模型在经济社会中的实际应用。第三章首先提出矩阵型网络结构的具体结构形式，然后对矩阵型网络结构进行分析，基于数学逻辑和经济学范式，建立相应的生产可能集理论及相应的多目标规划模型。第四章到第六章针对矩阵型网络系统的特点，构建NCCR模型，NCCGSS模型及NSBM模型，对模型提出数值计算和有效性判别方法；证明矩阵型网络DEA有效与Pareto有效解之间的等价关系，同时给出了数值算例分析。第七章是实证分析。以农业循环经济系统效率评价、基于投入产出表第二产业效率评价及电力企业效率评价为应用对象开展所建模型与方法的实证研究，分析了评价结果的现实意义，也显示了本文的研究具有一定的实际应用价值。总之，本文从理论上初步构建矩阵型网络DEA模型的方法体系，从实践上为宏观及微观循环经济系统的管理效率评价与决策提供可操作的定量手段支撑。
[Abstract]:Efficiency evaluation is not only a very important task in economic and social life, but also a difficult one. This paper studies how to evaluate the efficiency of decision making units in economic and social life. Data Envelopment Analysis (DEAA) is a mathematical programming method for evaluating the relative efficiency of comparable decision making units. Since the first DEA model was established in 1978, it has gradually developed into one of the most important methods for evaluating relative efficiency. The traditional DEA method regards the decision making unit as a whole and does not consider the input-output relation within the decision making unit. This method can not effectively reflect the true efficiency of the decision making unit. The purpose of this paper is to open the decision making unit. The DEA model which can calculate the efficiency value of the decision making unit is put forward by analyzing the decision making unit. The DEA method can be extended to the efficiency evaluation of complex system. Matrix network system is a kind of complex system, such as circulatory system, life cycle based supply chain system and so on. Each subsystem has external input and output, and there is a bidirectional input-output relationship between each subsystem. The DEA model of matrix network meets the requirement of management practice. According to the different research methods, the main contents of this paper can be divided into five chapters. The main contents are as follows: the structure analysis of matrix network system and the establishment and analysis of NCCGSS model; the establishment and analysis of NSBM model; the practical application of NCCGSS model in economic society. In the third chapter, the concrete structure of matrix network structure is put forward, and then the matrix network structure is analyzed. Based on mathematical logic and economic paradigm, the corresponding production possibility set theory and the corresponding multi-objective programming model are established. From 4th to 6th, according to the characteristics of matrix network system, NCCR model and NSBM model are constructed, numerical calculation and validity judgment method are proposed, and the equivalent relationship between DEA efficiency and Pareto efficient solution of matrix network is proved. Numerical examples are also given. Chapter 7th is the empirical analysis. The empirical research is carried out on the evaluation of agricultural circular economy system efficiency, the evaluation of the secondary industry efficiency based on the input-output table and the efficiency evaluation of electric power enterprises. The practical significance of the evaluation results is analyzed, and it also shows that the research in this paper has some practical application value. In a word, this paper preliminarily constructs the method system of matrix network DEA model in theory, and provides the operational quantitative means for the management efficiency evaluation and decision making of macro and micro circular economy system in practice.
【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：N941.4

【参考文献】