【作者】 陈学中；

【导师】 盛昭瀚；

【作者基本信息】 南京大学 ， 管理科学与工程， 2011， 博士

【摘要】 计算实验方法在社会科学各领域的研究和应用呈现出惊奇的发现,应用计算实验方法研究组织管理问题具有创新性的意义。组织管理问题的研究已经有许多传统的方法论和理论,如交易费用经济学、组织行为理论、战略理论、组织生态学和演化理论等。这些传统的方法论和理论对现实世界的实际运行情况进行解释和分析。基于多主体系统的计算实验方法论更注重现实世界的多样性、非线性和适应性,其模型运行的结果对现实世界的解释力更强。本文梳理了相关文献,提出组织演化问题研究方法论的多路径整合思想,研制人工组织模型框架,并以人才集聚和组织起源为实例进行了模型设计和实验分析,得出了一些结论。(1)组织演化问题研究方法论的多路径整合思想。从横向集成的角度,把传统方法论与计算实验方法论结合起来,既要承认和充分利用人类已经取得的先进的、正确的科学成果,又要研究新的具有发展前景的方法论。从纵向综合的角度,需要将“自下而上”和“由上而下”的方法论结合起来。基于复杂适应系统思想的计算实验方法论,是自下而上的集聚方法。而传统的组织理论多考虑了系统的整体性,更多地是由上而下的方法论。自下而上的方法论,适用于一般自然界或生物界,是正确合理的。然而人是具有智慧和创造能力的高级动物,常常会发号施令,只研究自下而上的方法论,显然是无法解释这些客观现实或现象。因此,本文提出“自下而上”和“由上而下”方法论的纵向结合,形成基于复杂适应系统思想的组织演化问题研究方法论的多路径整合思想。(2)人工组织模型框架。组织是一个典型的复杂适应系统,可以将组织分为人类组织和人工组织。在分析了个体的局限性、传统组织的特征以及组织环境的复杂性和不确定性的基础上,根据组织的实体构成要素,提出面向任务的组织的一般模型框架,并分析了组织内和组织间各要素构成的网络。人工组织理论是应用数学方法和计算方法,将人类组织和人工组织看作是计算实体,(理论上)建立关于组织的新概念、新理论和新知识,(技术上)研制在计算机多主体系统中模拟的组织模型的验证和分析的工具及程序,(实践上)通过这些工具和知识,将人工组织模型的实验结果反过来应用于实际的人类组织的实践。根据复杂适应系统的特点,设计了适应性主体的结构,探讨了主体的知识获得、学习模型以及主体的交互模式和行为机制,提出了人工组织模型的背景假设、组成要素、设计原则和学科基础,建立人工组织的基本模型框架,并将组织起源、组织内部演化和组织适应性看作是人工组织模型的主要研究内容。(3)基于多主体系统的人才集聚模型。论述了人才集聚的概念和理论依据,建立了人才流动的静态博弈模型和人才集聚的演化博弈模型,分析了人才集聚过程中存在的主要障碍和政策导向。在剖析借鉴人工社会系统的基础上,建立了基于多主体系统的人才集聚模型,描述了模型的目标、环境、主体及模型的设计。通过对人才集聚模型的运行,得到了较丰富的实验资料。我们从信息不对称环境、区域弱联盟模式、区域强联盟模式、完全信息环境、封闭系统的孤岛等状态下,对模型实验结果,进行了分析和论证,得出了在区域经济社会与个体之间的相互影响和作用机理。(4)基于多主体系统的组织起源模型。分析了组织起源理论的传统观点和交易费用理论,探讨了新古典经济学和新制度经济学关于组织边界的界定问题。由于组织的社会网络是组织立身的基本环境,我们探讨了社会网络分析的基本概念、特征和分析指标,进而引出了团队的构建,特别是创新团队构建的社会资本观。在组织起源和团队形成等研究的基础上,根据多主体系统建模的思想,对组织起源模型的环境、主体和主体行为规则进行了分析和设计,并在计算机上实现了基于多主体系统的组织起源模型。模型运行的结果呈现出组织作为复杂适应系统的多样性。同时,可以验证组织的存在不只是由交易费用的存在而决定的,更应该是个体能力的局限和组织能力的强大,完全符合系统科学中1+1大于2的观点。特别是模型呈现出的多样性,对传统理论的假设条件,可以作为输入参数或规则,加入模型,从而得到对已有理论直接客观的验证或理论创新。更多还原

【Abstract】 Computational experiment methods are increasingly applied to various fields of social sciences and show surprised findings, so it is of innovative significance using computational experiment methods to research organizational management problems. There have been many traditional methodologies in organizational management research, such as transaction cost economics, behavior analysis, strategic theory, organization ecology, and evolutionary theory, which explaining the real world. Computational experiment methods pay more attention to the diversity and nonlinearity and adaptability of the reality. The comprehensive and integrated methodology based on the computational experiment is put forward to and the model framework of the artificial organization is developed. Then talents aggregation model and organization emergence model based on multi-agents system are designed.(1) The comprehensive and integrated methodology based on the computational experiment is proposed. From the point of view of the horizontal integration of the traditional methodology and computational experiment methodology, we not only admit and make full use of the advanced and correct scientific achievements of human beings, but also study the new potential methodology. From the longitudinal comprehensive angle, we need to combine "bottom-up" and "top down" methodology together. The computational experiment methodology based on complex adaptive system belongs to bottom-up methods, while most of the traditional methodologies are top down methods because the traditional organization theory considers more the integrity of the system. The bottom-up methodology is used right in the general nature or biological world. But it could not explain the reality of organizations well only using the bottom-up methodology because the man as the advanced animal has the wisdom and the ability of creation and often give orders. Therefore, we should combine "bottom-up" and "top down" methodology and integrate the computational experiment methodology and the traditional methodology to form the comprehensive and integrated methodology based on the computational experiment.(2) The artificial organization model framework is developed. The organization is a typical complex adaptive system and can be divided into human organization and artificial organization. The general model framework of the task oriented human organization is put forward to on the basis of the entity components of the organization by analyzing the limitations of individuals and the characteristics of traditional organizations and the complexity and uncertainty of organization environment. The various kinds of networks inside the organization and between organizations are analyzed.The artificial organization theory establishes new concepts and new theory and new knowledge of organization and develops analysis tools and programs for validating simulated organization model based on the multi-agents system and the application of mathematical method and computational method. Inversely, the result and conclusion of running the above computer model of the artificial organization is applied to instruct the practice of the man. According to the characteristics of complex adaptive system, the structure of the adaptive agent is designed and the knowledge acquisition and the learning model and the interactive mode and the behavior mechanism of the adaptive agent are discussed. The basic model framework of artificial organization is established and the background assumption and the design principles and the discipline base and the basic elements of the artificial organization are analyzed. The main study contents of the artificial organization are the organization emergence and the evolutions inside the organization and the organizational adaptation.(3) The concept and theory basis of talents aggregation are discussed and the static game model of talents flow and the evolutionary game model of talents aggregation are established to analyze the main existed obstacles and policy guidance. The talents aggregation model based on multi-agents system is developed and the goal, the environment, the agent and the design of the model are described. The conclusion on interactive impact and interaction mechanism between the regional economic society and individuals are discussed in the mode of information asymmetry environment, and regional strong alliance, and regional weak alliance, and complete information environment, and the closed system such as the island on the analysis and verification of experimental results and the plentiful experiment materials getting through running the talents aggregation model.(4) The traditional view of the organization origin theories and transaction cost economics are analyzed and the organization boundary definitions given by the new classical economics and the new institutional economics are discussed. We describe the basic concepts and features and analysis index of the social network analysis because of the social network of organization is the foundation environment that organization survives, and then we explore the construction of the team, especially the innovation team built from social capital concepts. The organization emergence model based on the multi-agents system is developed by analyzing the environment and agents and behavior rules. The running results of the computational model present the diversity of organization as a complex adaptive system. At the same time, it is proved and verified that the existence of the organization is not only because of the existence of transaction costs but also because of the limited ability of individual and the strong ability of organizations. This conclusion exactly matches the principle of system science "1+1> 2". Especially the diversity of the output of the computational model shows that the traditional theory can be verified directly or even new theory can be found by inputting the assumptions and rules of traditional theory as parameters and running the model and exploring the model output.更多还原