【作者】 胡雄鹰；

【导师】 张金隆； 胡斌；

【作者基本信息】 华中科技大学 ， 管理科学与工程， 2010， 博士

【摘要】 传统企业组织理论研究重心在集中式组织,对于分布式组织理论及其模拟方法较少研究。本文先探讨了分布式组织一般理论,提出了分布式组织的模拟方法,分别研究了网络组织、项目群、分布式团队和分形企业的组织结构设计和优化的模拟方法。首先,归纳阐述了分布式组织结构设计和优化的模拟方法。对分布式组织从两个角度进行分类。一种分法是将现有的具有分布式特征的企业组织归纳为特许经营、承包公司、网络组织、战略联盟、产业集群等五类。另一种分法是将原有企业组织按分布式扩展为项目群、网络组织、分形企业和分布式团队等四类。归纳了分布式组织结构设计和优化的模拟方法。指出可用于分布式组织结构设计和优化的分布式模拟技术主要为Petri网和MAS。其次,引入Petri网表示网络组织结构,并对网络组织结构设计和优化进行了模拟研究。在不同的具体组织结构下网络组织完成相同任务时的绩效是不同的,使绩效达到最大时的组织结构即为该网络组织的最优结构形式。生产型网络组织的绩效目标设定为产品总生产能力最大化,将网络组织结构映射成Petri网形式后,将问题转化为在Petri网上求解网络最大流问题。服务型网络组织的绩效目标设定为服务总成本最小化,将问题转化为在Petri网上求解网络最短路问题。绩效最优时的组织结构即为所求。再次,引入Petri网表示项目群组织结构,并对项目群组织结构设计和优化进行了模拟研究。将项目运行参数转化为Petri网的不同变迁、库所与托肯参数值,分析了项目群中子项目之间的资源共享状况,研究了他们对项目工期、绩效的影响,最后给出项目群中不同资源的共享策略,以及基于不同绩效目标的项目群最优组织结构。然后,基于MAS建立了一个分布式团队角色-活动互动模拟模型,研究分布式团队角色-活动互动机制,提出了分布式团队结构设计和优化的模拟方法。根据贝尔宾团队概念,将成员角色分为3类共9种。将活动分为创新、应急和常规三类。模拟程序的核心设计理念是如何将拥有不同能力的成员组合成团队使之完成给定任务的绩效最高。模拟结果如下：任务完成得越快,则下次下达的任务会越多越难,即成员能力与任务难度得到协同演化；当引入团队间合作时,绩效会增加；有竞争机制时总绩效会增加,但会出现马太效应；由于成员能力提高导致团队能力更快提升,团队可以完成任一单一个体无法完成的复杂任务；成员随机结盟的合作行为涌现出团队,同样道理,团队间协作行为涌现出更高一级的组织形态来协作完成更难的任务。最后,归纳出分形企业的概念,提出分形企业组织结构设计和优化的模拟方法。分形企业由若干相对独立的单元组成,这种单元就是分形企业的分形元。基于分形控制结构在分布式并发模拟平台ExSpect上实现了对分形企业组织结构优化的模拟研究。还基于MAS实现了分形企业的组织结构,并在此基础上进一步做了分形企业组织结构优化工作。最后通过价格季节性随机波动环境下的分形企业采购实例,利用蒙特卡洛模拟方法,展示在业务竞争异常激烈的不确定经济环境下,由确定性经济环境下分析出的结论对于分形企业经常失效,并对其原因进行分析。更多还原

【Abstract】 Traditional organization theory focused on the centralized organization while the distributed organization theory and its simulation study was less. This paper first discussed the general theory of distributed organization and put forward simulation methods of distributed organization, especially studied the organizational structure design and optimization of the network, project group, distributed team and Fractal Company by simulation approach.First of all, the simulation methods of the distributed organization’s structure design and optimization were summed up. Distributed organization was classified from two perspectives. One classification was to summarize the enterprise organization available as five categories, such as franchising, contracting company, network, strategic alliance, industry cluster by its distributed nature. The other classification is to expand the existing organization to project group, network, fractal enterprise and distributed team by distributed methods. The simulation methods of distributed organization’s organizational structure design and optimization were summarized. It was pointed out that the distributed simulation technology for design and optimization of the distributed organization’s organizational structure was mainly HLA, Petri nets and MAS.Secondly, Petri net was introduced to denote the network’s organizational structure by which the design and optimization of network’s organizational structure was studied. When the specific organizational structure of different networks completed the same task the performance is different, so that when the performance reached the maximum the network’s optimal structure formed. The performance target of Production network organization was set to maximize the total capacity. When the organizational structure of network was mapped to the form of Petri nets, the issue would turn into solving maximum flow of network in Petri nets. The performance goal of Service-oriented network organization was set to minimize the total service cost. The problem is transformed into solving the most short circuit of network. The optimal organizational structure was gotten when the performance reached a maximum.Thirdly, Petri net was introduced to denote the project group’s organizational structure. And the design and optimization of it was studied. Operating parameters of the project were translated into the different transition, place and token parameter values of Petri net, analyzed the resource-sharing status in the project group between the subprojects, and studied their impact on project schedule and performance, at last provided the different resource sharing strategies in the project group and the optimal organizational structure of project group based on different performance goals.Fourthly, the roles-activities’interactive simulation model of a distributed team was established based on MAS, the roles-activities’interactive mechanism of a distributed team was studied and the simulation method of distributed team’s structural design and optimization was put forward. According to Belbin’s team concept, the role of the membership was divided into three categories 9 species. The activities were divided into three categories including innovative, emergent and routine one. Simulation program’s core design philosophy is how members who have different abilities formed teams to complete a given task so that the performance was of the maximum. The simulation results are as follows:the faster the task is completed, the more difficult the next task would be assigned, that is, members’ability and task’s difficulty would be co-evolution; as team co-operation was introduced, the performance would increase; as a competitive mechanism existed, the overall performance would increase, also Matthew effect would appear; because the improvement of their ability led the ability of the team to upgrade more quickly, the team can complete the complicated task which a single individual could not complete; the cooperation of the members emerged in the team by acts of random alliance, similarly, a higher level of organizational form would emerge to complete collaboratively more difficult task by the collaborative behavior among team. And finally, the concept of Fractal Company was summarized, the simulation approach of design and optimization of Fractal Company’s organizational structure was proposed. Fractal Company is made by a number of relatively independent units, which are the fractal elements of Fractal Company. The simulation research of Fractal Company’s organization structure optimization was realized based on fractal control structure on a distributed concurrent simulation platform ExSpect. The organizational structure of Fractal Company is also achieved based on MAS, and further structural optimization of Fractal Company was operated on this basis. At the end of chapter, fractal corporate procurement example at randomly fluctuating seasonal price was given. It showed that under a very competitive and uncertain economic environment, the conclusions from the certain economic environment often failed to the Fractal Company by Monte Carlo simulation and the causes was analyzed.更多还原