分类号： 学 校 代 码：10069
密级：研究生学号：120190402
基于讨价还价博弈-DEA的企业效率评价方法
研究
Research on the Method of Enterprise Efficiency Evaluation Based
on Jointment of Bargaining Game and DEA
研 究 生 姓 名：孟海宁
专 业 名 称：企业管理
指导教师姓名：王庆 教授
论文提交日期：2022年 5 月
学位授予单位：天津商业大学 独创性声明及使用授权声明
一、学位论文独创声明
本人所撰写的学位论文是在指导教师的指导下独立完成的研究成果。除已明确标
注或得到许可外，所撰写的学位论文中不包含他人已申请学位或其他用途所使用过的
成果，不包含其他个人或者集体已经发表或者撰写过的研究成果，本人的指导教师对
此进行了审定。对本文的研究做出重要贡献的个人或者集体，本文已在文中做出明确
的说明并表示谢意。如有不实，本人承担相应责任。
二、学位论文及研究成果使用授权声明
本人同意授权天津商业大学以非赢利方式保存、使用本人的学位论文的电子版及
纸质版。授权天津商业大学将本论文的全部内容或部分内容提供给有关方面编入数据
库进行检索、出版及提供信息服务。同意学校向国家有关部门或机构送交论文的复印
件和磁盘。本人在校期间取得的研究数据、相关成果等知识产权归天津商业大学所有。
注：涉及保密的学位论文在解密后适用本授权。
学位论文作者签字：
指 导 教 师 签 字：
日 期：2022 年 5 月 26 日摘 要
科学的效率评价是企业提高效率的关键前提，为企业提供一个科学的效率评价方
法具有重要的现实意义。因此，本文通过对现有相关文献进行查阅和分析，结合数据
包络分析（Data Envelopment Analysis，DEA）和讨价还价博弈论相关理论，给出了一
个更科学、更合理的讨价还价博弈-DEA 企业效率评价方法。
企业生产活动中有着一系列复杂过程，因此考虑多指标效率评价方法是有必要
的。在众多的效率评价方法中，DEA是一种可以用于评价具有多投入和多产出的同质
决策单元的相对有效性方法，这使得 DEA 在企业效率评价过程中被广泛应用。同时
传统DEA 存在自利性的问题，即决策单元自由地选择权重，这会导致评价结果不能
被决策单元所接受。DEA 求解公共权重可以解决这一问题。在众多的 DEA 求解公共
权重的方法中，基于讨价还价博弈-DEA 求解公共权重的企业效率评价方法的优势是
经过博弈双方不断讨价还价，所得的公共权重是纳什均衡解，评价结果容易被决策单
元所接受。
本文采用文献研究法、讨价还价博弈论方法、数据包络分析法和数学仿真法，研
究了讨价还价博弈-DEA 企业效率评价方法的优缺点，提出了改进的讨价还价博弈
-DEA 的企业效率评价方法的模型，并用实例验证了本方法的科学性。
针对传统的讨价还价博弈-DEA 仍存在的两个问题，本文给出了相应的解决方法：
（1）针对模型中决策单元效率范围不精准的问题，本文确定了一个更合理、更
精准的决策单元效率范围。通过研究文献发现，使用纳什讨价还价博弈-DEA求解公
共权重是通过在决策单元的效率最大值和效率最小值之间找到一个纳什讨价还价效
率值，达到一种均衡解，使得评价结果被决策单元所接受。效率范围不准确会使所求
的公共权重不科学。因此，本文采用决策单元CCR-DEA 的效率值作为上界，表示决
策单元最有利的状态；最小的交叉效率值作为下界，表示决策单元最不利的状态。
（2）针对模型的目标函数没有体现决策单元讨价还价能力差异的问题，本文在
模型的目标函数中加入贴现因子来体现每一个决策单元讨价还价能力的差异。通过研
究文献发现，很多学者在使用纳什讨价还价博弈模型时会采取四个公理假设，其中一
个假设为博弈者之间的讨价还价能力相同。但在现实中，决策单元的权重向量不同，
表示决策单元的实力不同，那么每个决策单元的讨价还价能力不同。本文不考虑对称
性公理对博弈双方具有相同讨价还价能力的假设，并在模型的目标函数中加入贴现因
子来体现每一个决策单元讨价还价能力差异。
通过对企业进行实例测算和结果分析，表明本文提出的方法具有一定的可行性和
科学性。
关键词：效率评价；数据包络分析；讨价还价博弈；公共权重ABSTRACT
Scientific efficiency evaluation is the key premise for enterprises to improve efficiency.
It is of great practical significance to provide a scientific efficiency evaluation method for
enterprises. Therefore, based on Data Envelopment Analysis (DEA) and bargaining game
theory, a more scientific and reasonable bargaining game-DEA enterprise efficiency
evaluation method is proposed in this paper by consulting and analyzing existing relevant
literature.
There are a series of complex processes in the production activities of enterprises, so it
is necessary to consider multi-index efficiency evaluation method. Among many efficiency
evaluation methods, DEA is a method that can be used to evaluate the relative effectiveness
of homogeneous decision making units (DMUs) with multiple inputs and outputs, which
makes DEA widely used in enterprise efficiency evaluation methods. At the same time,
traditional DEA has the problem of self-interest, that is, DMUs choos the weight freely,
which leads to the evaluation results cannot be accepted by the decision making unit. DEA
can solve this problem by solving the common set of weights (CSW). Among the numerous
methods of solving CSW by DEA, the advantage of the enterprise efficiency evaluation
method based on bargaining game-DEA solving CSW is that the two sides of the game
continue to bargain, and CSW obtained is the Nash equilibrium solution, and the evaluation
result is easy to be accepted by DMUs.
In this paper, literature research method, bargaining game theory method, DEA and
mathematical simulation method are used to study the advantages and disadvantages of
bargaining game-DEA enterprise efficiency evaluation method, put forward an improved
bargaining game-DEA enterprise efficiency evaluation method model, and verified the
scientific nature of the method with an example.
In view of the two problems still existing in the existing bargaining game-DEA, this
paper provides corresponding solutions:
(1) Aiming at the problem that the efficiency range of decision making unit in the
model is not accurate, this paper determines a more reasonable and accurate efficiency
range of decision making unit. Through the research literature, it is found that using Nash
bargaining game-DEA to solve CSW is to find a Nash bargaining efficiency value between
the maximum efficiency and the minimum efficiency of the decision making unit to achieve
an equilibrium solution, so that the evaluation results are accepted by the decision making
unit. Inaccurate efficiency range will make the desired common weight unscientific.
Therefore, the efficiency value of CCR-DEA is used as the upper bound to represent themost favorable state of DMU. The minimum crossover efficiency value is the lower bound
and represents the most unfavorable state of DMU.
(2) In view of the problem that the objective function of the model does not reflect the
difference in bargaining power of decision-making units, this paper adds a discount factor
into the objective function of the model to reflect the difference in bargaining power of
each decision-making unit. Through research literature, it is found that many scholars adopt
four axiomatic assumptions when using Nash bargaining game model, one of which
assumes that the bargaining power of players is the same. However, in reality, different
weight vectors of DMUs represent different strength of DMUs, so each decision making
unit has different bargaining power. This paper does not consider the assumption that the
two sides of the game have the same bargaining power, and adds a discount factor into the
objective function of the model to reflect the difference of bargaining power of each
decision-making unit.
It is proved that the method proposed in this paper is feasible and scientific.
Key words：Efficiency evaluation, Data envelopment analysis, Bargaining game, Common
set of weights