基于神经网络的高管团队决策绩效的评价研究

VIP免费

3.0 牛悦 2024-11-19 4 4 1.25MB 74 页 15积分

侵权投诉

摘要

本文在国家社科基金资助项目“团队过程视角下的高层梯队特征对企业行为

和绩效的影响研究”（项目编号 11BGL014）和上海市教委科研创新重点项目“基

于人力资本的高管团队认知特征对战略一致性的作用机制研究”（项目编号：

10ZS96）的课题框架体系基础上，利用人工神经网络（Artificial Neural Networks）

非确定性因果关系映射的特点，建立了一个模拟现实复杂决策环境的决策绩效评

价模型，用来对高层管理团队（Top Management Team, TMT）决策绩效进行评价

研究，为后续的相关研究提供了崭新的研究视角和研究方向。

随着全球经济一体化与日趋激烈的市场竞争，企业面临的问题日益复杂化，

这对企业的高层管理者提出了崭新的挑战，高层管理团队作为一个特殊团体其表

现（尤其是决策绩效）对现代企业组织的成功起着日益重要的作用。

1984 年Hambrick 与Mason 提出的“高层梯队理论（Upper Echelon Theory）”

吸引众多管理学者开展了大量有关高层管理团队与其绩效、战略一致性和决策等

方面的研究。对于之前高管团队决策绩效的研究，可以总结为两点：第一，研究

模型大多是利用线性回归函数；第二，以往评价模型只关注了团队构成特征对决

策绩效的关系，或团队结构特征与决策绩效的关系，或决策过程与决策绩效的关

系。但是没有一个完整的、涉及 3个以上决策绩效影响因素的评价模型。而近年

来，人工神经网络的兴起为研究 TMT 决策绩效提供了新的视角。

本文首先探讨了以往 TMT 决策绩效的评价方法和评价因素，在此基础上，根

据高层梯队理论和决策绩效评价的相关理论，建立了本文的 TMT 决策绩效评价体

系。利用神经网络非确定性因果关系映射的特点，选取显著影响决策的因素和评

价其绩效的因素，作为网络输入和输出向量。并通过发放调查问卷请高管团队成

员根据已有的决策经验为每项因素赋值，再利用 MATLAB 人工神经网络工具箱建

立ELMAM 神经网络模型，并对网络模型进行训练。训练好的神经网络模型经测

试，可以应用于企业对高管团队决策绩效的评价，有实际的应用价值。

从建模的分析结果来看，本文得到了以下结论：

第一，将神经网络理论引入到高管团队决策绩效模型上来，建立以高管团队

特征为自变量，决策绩效衡量标准为因变量的神经网络模型是可行的。

第二，从模型的评价结果来看，神经网络完成了决策绩效评价任务，这也充

分显示了神经网络在建立非线性、非确定性因果关系映射模型中的强大生命力。

神经网络不像线性模型，要求因素集的完整性、严密性，其神经元可以自动寻找

给定因素集的相关信息进行建模，以达到模型的正确性，具有容错性。

第三，本文建立了评价 TMT 决策绩效的模型，企业可以用来评价其 TMT 的

决策绩效。

关键词：人工神经网络高管团队决策绩效评价

III

ABSTRACT

This paper, as the sub project of "Team process perspective on top echelon to the

enterprise behavior and performance characteristics study" sponsored by National

Social Science Fund Project (PN:11BGL014) and "Human capital based cognitive

characteristics on the TMT strategy consistency mechanism research" sponsored by

Shanghai Municipal Education Commission, Scientific Research Innovation Key

Project (PN:10ZS96), breaks through the shortage of former TMT decision study by

using the neural network capable of uncertain causality mapping features and

establishing a nonlinear, simulated reality of complex decision-making environmental

decision making performance forecasting model, and used to evaluate the Top

Management Team (TMT) decision performance. In the meanwhile, this research

provides a new perspective and research direction.

Due to the global economy integration and the increasingly fierce market competition,

the problems faced by the enterprise are growing complexity, the TMT as a special

group, whose performance (especially decision making performance) is essential for

modern enterprise organizational success.

"High Echelon theory" proposed by Mason Hambrick in 1984, attracted numerous

management scholars, who initiated many studies such as TMT and its performance,

strategic consistency and decision making etc. The former studies on TMT decision

making can be summed up in two points: firstly, the decision making models are mostly

based on the linear regression function; secondly, the former decision making evaluation

model either focus on the team compositional features to the decision making

performance, or team structural characteristics to decision making performance, or the

relationship between decision making process and decision making performance.

However, there isn’t a complete model involving three or more influencing factors of

decision making performance. Until recent years, the rise of artificial neural network for

the TMT study provides a new perspective.

Based on the this Upper Echelon Theory analysis and decision making performance

evaluation summaries, this paper firstly discusses the former evaluation methods and

influencing factors of TMT decision making performance and then build TMT decision

making performance evaluation system. By using the uncertain causality mapping

features of Neural Network, this paper elects the factors which can affect the

performance of TMT decision and decision-making performance measurement.

Executives number every factor based on their experiences during the survey. Then

using MATLAB toolbox to build and train ELMAM Artificial Neural Network model.

The trained model after testing can be used to solve the problem of evaluating

decision-making performance. Therefore, this study has actual practical value.

From the analysis of the modeling result, this paper got the following conclusion:

1. The neural network theory is introduced to executive team decision performance

model, and establishes team executives as independent variable features;

decision-making performance measurement standard for the dependent variable neural

network model is feasible.

2. From the prediction result, neural network completed forecasting task, it also show

the neural network in the uncertain nonlinear, establish causality mapping model of the

powerful vitality. Neural network does not like linear model, the requirement of

integrity, rigor factor, and the neurons can be looking for a given set of factors related

information model, in order to achieve the validity of the model, has the fault tolerance.

3. We build an evaluation model, which can be used to evaluate the decision making

performance.

Key Word: ANNS, TMT, Decision making Performance Evaluation

摘要

ABSTRACT

第 1 章绪论 ................................................................................................................ 1

§1.1 研究背景 ............................................................................................................ 1

§1.2 研究的目的和意义 ............................................................................................ 2

§1.3 主要研究内容 .................................................................................................... 2

§1.4 研究思路 ............................................................................................................ 3

第 2 章相关理论及其文献综述 .................................................................................. 5

§2.1 TMT 决策相关理论 ............................................................................................ 5

§2.1.1 群体决策理论 ............................................................................................. 5

§2.1.2 团队决策理论 ........................................................................................... 10

§2.1.3 TMT 决策模式 .......................................................................................... 15

§2.2 团队决策绩效评价的相关理论 ...................................................................... 21

§2.3 人工神经网络 .................................................................................................. 21

§2.3.1 人工神经网技术发展的主要历程 ........................................................... 22

§2.3.2 ELMAN 人工神经网络的结构 ................................................................ 23

§2.3.3 ELMAN 人工神经网络学习算法 ............................................................ 25

§2.3.4 ELMAN 人工神经网络决策评价研究的可行性 .................................... 26

§2.4 MATLAB 简介 .................................................................................................. 28

第 3 章高管团队决策绩效的评价探讨 .................................................................... 29

§3.1 以往 TMT 决策绩效评价方法的总结 ............................................................ 29

§3.2 以往 TMT 决策绩效评价因素的总结 ............................................................ 31

§3.3 本文对决策绩效评价方法和因素的选取 ...................................................... 33

§3.4 环境变量的影响 .............................................................................................. 34

第 4 章评价模型体系及变量设计 ............................................................................ 35

§4.1 ELMAN 神经网络模型引入决策评价 ............................................................ 35

§4.2 神经网络模型所应用的测评指标体系 .......................................................... 37

§4.3 问卷的设计及数据来源 .................................................................................. 38

§4 .4 影响决策绩效的因素解释 .............................................................................. 41

§4.5 评价决策绩效的因素解释 .............................................................................. 49

第 5 章评价模型的建立及数据分析 ........................................................................ 51

§5.1 模型的结构确定 .............................................................................................. 51

§5.2 模型 MATLAB 计算实现 ................................................................................ 53

第 6 章结论及展望 .................................................................................................... 57

§6.1 结论 .................................................................................................................. 57

§6.2 展望 .................................................................................................................. 57

附录 ........................................................................................................................ 59

参考文献 ........................................................................................................................ 63

在读期间公开发表的论文和承担科研项目及取得成果 ............................................ 69

致谢 ........................................................................................................................ 70

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

15 积分 4人已下载

立即下载 VIP免费下载

摘要：

I摘要本文在国家社科基金资助项目“团队过程视角下的高层梯队特征对企业行为和绩效的影响研究”（项目编号11BGL014）和上海市教委科研创新重点项目“基于人力资本的高管团队认知特征对战略一致性的作用机制研究”（项目编号：10ZS96）的课题框架体系基础上，利用人工神经网络（ArtificialNeuralNetworks）非确定性因果关系映射的特点，建立了一个模拟现实复杂决策环境的决策绩效评价模型，用来对高层管理团队（TopManagementTeam,TMT）决策绩效进行评价研究，为后续的相关研究提供了崭新的研究视角和研究方向。随着全球经济一体化与日趋激烈的市场竞争，企业面临的问题日益复杂化，...

展开>> 收起<<

基于神经网络的高管团队决策绩效的评价研究.pdf

共74页,预览8页

还剩页未读，继续阅读

基于神经网络的高管团队决策绩效的评价研究

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

推荐作者

热门标签

举报选择: