Research On Theory, Models And Applications Of Network DEA

In the actual production activities, many decision making units have network production structure, such as banks, hospitals, insurance companies, airports, railways and so on. Obtaining the efficiencies and ranks of these DMUs has gotten more and more attention of scholars and decision makers of enterprises at home and abroad. Data envelopment analysis (DEA), as a kind of nonparametric method, has advantages of without the need of pre-setting the parameters of the production function and the weights for inputs and outputs. Therefore, since proposed in 1978, DEA has attracted the attention of many scholars, and has been applied to many areas. The objective of the study is the DMUs with network production structure. This study is based on network DEA method to study the ranking intervals and efficiencies of DMUs with network structure. Besides, this study applied the DEA method to evaluate the environmental efficiency and merger efficiency. Thus, this study greatly promoted the development of theory and application of network DEA.This paper is divided into six chapters. The main contents are summarized as follows:The first chapter is the introduction. First, we introduce the basic theory of DEA: the basic concepts of inputs and outputs, decision-making unit, production possibility set, production frontier and efficiency; the basic idea of how to use DEA method to do efficiency evaluation; some basic DEA models and DEA applications. Then, the main subject of this study is introduced, i.e., network DEA theory and its applications. We introduced the significance of network DEA theory, related literature and application areas. Finally, we summarized the research methods, content and significance.The second chapter is environmental efficiency analysis based on relational two-stage DEA model. In the actual production process, decision making unit often has a network structure. For example, Chinese regional industrial production system consists of two serial subsystems. This study proposed a new relational two-stage model to assess environmental efficiency of such network structure. An example of Chinese regional industrial system shows applicability of the proposed approach.The third chapter is a Network SBM model considering national environmental policy for environmental performance evaluation of industrial systems in China. Chinese resource conservation and environmental protection are legislated into law. The goal of the government is to gradually build a resource-saving and environment -friendly society. In China, resource conservation and environmental protection are equally important. Considering the internal network structure of each regional industrial system, policy of resource utilization and environmental protection in China, and the defects of previous literature on network DEA model to assess the environmental efficiency based on radial projection, this study proposes a new network SBM model (NSBM). This model takes into account non-radial projection of inputs, outputs and intermediates and the current environmental policy to measure policy measure environmental efficiency of Chinese regional industrial systems, and do empirical analysis.The fourth chapter is the ranking interval method for two-stage production systems. Traditional DEA models find the most favorite weights for each DMU when evaluating its efficiency. Salo and Punkka (2011) overcame this problem by proposing a set of mix-integer models by considering all the feasible weights, but they deem each DMU as a black box. This study develops a method to obtain the ranking intervals for the classic two-stage production system. The proposed model calculates each DMU’s ranking intervals for the overall system as well as two subsystems. Thus, it provides more accurate information for decision makers-It may identify best (or worst) DMUs in the overall system and both subsystems over all feasible weights. Besides, the proposed approach provides information regarding the sensitivity of the DMU’s ranking intervals for the overall system and both two subsystems over sets of all feasible weights.The fifth chapter is merger efficiency evaluation of two-stage production system based on non-cooperative game theory. In terms of how to choose the candidate target companies and predicting the feasibility of M&As, this paper applies data envelopment analysis to predict the feasibility of M&As with a two-stage production system. The production system has two distinct characteristics:(1) a decision-making unit consists of two tandem components; (2) one component is in a dominant position, and the other is in a subordinate position. A hypothetical DMU is merged by two or more DMUs. To solve this kind of merger efficiency evaluation problem, this paper firstly applies a DEA approach to evaluate the DMU’s efficiency of the overall system and both components simultaneously. Then based on the idea of non-cooperative game, a game DEA approach is provided to evaluate the hypothetical DMU’s merger efficiency of the overall system and both components on the condition that the current output level and efficiency are constant, which is helpful to analyze how the merged hypothetical DMU to save cost from its components.The sixth chapter is a summary and future research. First, the content of each previous chapter is summarized; then the major innovation and shortfalls of this article are listed; finally, the future research is proposed based on the shortfalls of this study.The innovative of this paper is summarized as the following aspects:(1) We extended the relational model from the constant returns to scale framework to the variable returns to scale version to evaluate environmental efficiency. And a heuristic search is applied to the extended relational model.(2) We divided each Chinese regional industrial system into a production process and a pollutant treatment process. We proposed a new Network SBM (NSBM) model to assess the environmental efficiency of Chinese regional industrial systems considering the internal structure of each regional industrial system as well as China’s policy on resource utilization and environmental protection.(3) In this paper, we extended Salo and Punkka (2011)’s models by considering the two-stage production structure. The proposed models calculate each DMU’s ranking interval for the overall system as well as two subsystems. And, they may be used to analyze the stability of the rankings for the overall system as well as the two subsystems.(4) To solve merger efficiency evaluation problem of DMUs with two-stage production system and non-cooperative leader-follower relationship between two sub-systems, this paper firstly applied a DEA approach to evaluate the DMU’s efficiency of the overall system and both components simultaneously. Then based on the idea of non-cooperative game, a game DEA approach is provided to evaluate the hypothetical DMU’s merger efficiency of the overall system and both components on the condition that the current output level and efficiency are constant.