【作者】 徐选华；

【导师】 陈晓红；

【作者基本信息】 中南大学 ， 管理科学与工程， 2005， 博士

【摘要】 网络技术的发展和网络技术在社会各行各业中的广泛应用,一方面给整个社会的思维方式和工作方式带来了深刻的影响,另一方面给企业的经营方式和政府政务的管理方式带来了深刻的变化,社会组织的重大经营和管理决策逐渐在网络上向群体决策转变。然而,这种影响和变化也使得参与决策的群体变得越来越复杂,其规模越来越大,网络环境下复杂大群体的出现给传统群体决策及其支持系统带来了新的内容。本文在对传统群体决策及其支持系统的研究基础上,结合系统的核心部件“模型驱动”,进一步研究网络环境下模型驱动的复杂大群体决策支持系统(Model-Drive Complex-Large Group Decision Support System in Environment of Network,MCLGDSS),提出了系统的体系结构和功能结构,对系统实现的理论、方法、开发技术和应用方式等方面进行了深入的探讨、研究、实现和应用。 论文的主要工作主要包括以下几方面: (1) 对网络环境下复杂大群体的出现背景及其基本特征进行了分析,对国内外群体决策和群体决策支持系统各理论研究现状与发展趋势进行了综述与分析。利用层次模型理论提出了研究的基本思路和系统框架。把模型驱动和复杂大群体决策方法作为研究的基石,以此为基础研究和构建群决策处理机制、网络环境下复杂大群体决策支持系统的体系结构、系统功能结构及系统各功能的实现方法、系统实现方法和应用方式,为各行业的社会组织在网络时代进行群体决策提供支持。 (2) 模型驱动方法研究 在广泛分析面向决策问题求解时模型需求的基础上,结合问题求解实际说明了模型存在的形式、描述及运行方式,提出了基于复杂决策问题求解的模型方案解决办法,通过将一个原子模型(单个模型)分解成方法(算法)对象和数据对象的组合和对原子模型进行复合(其中提出纵向和横向两种复合方式)来建立模型可重用体系和结构,同时提出了基于Multi-Agents技术的模型复合实现方法,形成网络环境下群体决策支持系统中的模型驱动方法。 (3) 网络环境下复杂大群体决策方法研究 根据网络环境下复杂大群体的特点,提出了将决策问题求解流程划分为二个阶段,在第一个阶段,各决策成员利用模型和知识求解形成自己的决策方案,这更多还原

【Abstract】 The development of network technology and the extensive application of network technology in every industry or trade of society have brought deep influence on the one hand to working way and the thought way of entire society, have brought deep change on the other hand to the way of managing of enterprise and the management way of government affairs. The significant decision of operating and management of social organization gradually turns into group decision in network. However this kind of influence and change also make the group who participate in the decision become more and more complex, its scale is more and more large. Under network environment, the appearance of the complex large group has brought new content to traditional group decision and its support system. Based on the research of traditional group decision and its support system and combing the systematic key part "model drive", this paper has further studied the model-drive complex-large group decision support system in environment of network (MCLGDSS ), and put forward systematic architecture and function structure. The theory, method, developing technology and application way etc. of the system realization have been gone on thorough discuss, study, realization and application.The major works of the paper mainly includes following few aspects:( 1 ) The appearance background and its basic feature for complex large group in network environmental have been analyzed. The studying situation in present and developing tendency of each theory of the domestic and international group decision and group decision support system have been summarized and analyzed. By use of layer model theory, the basic thought route and systematic frame of the research have been put forward. Regard model-drive and complex-large group decision method as the footstone of the research, based on it, the process mechanism of the group decision and the architecture of the complex-large group decision support system in network environment have been studied and constructed. The systematic function structure, the realizing method of all the systematic functions, the systematic realizing method and application way have also been studied. These will offer support to carry out group decision for the social organizations of every industry or trade in network times.( 2 ) The research of model-drive methodOn the foundation of model demand of extensive analysis oriented decision problem solving, combing the fact of problem solving, explain the form of model existence, description and performing way. The solution of model scheme based on complex decision making problem solving has been proposed. The architecture of the model reuse has been established through decompounding a atom model ( single model ) into a method ( algorithm ) object and a data object which are formed a combination and compounding atom models (in which the longitudinal and lateral two methods are put forward ) come to establish model can put system and structure in an important position. At the same time, the model compound realization method based on multi-agents technology has been put forward. It forms the model-drive method in the group decision support system of the network environment.( 3 ) The research of complex large-group decision method in network environmentAccording to the characteristic of the complex-large group in network environment, it has been proposed to divide the solving process of decision making problem into two stages. In first stage, each decision member solves to form own decision scheme using model and knowledge, it is equivalent to the traditional individual decision support system. In second stage, the group opinions are synthesized to form group decision. On the synthetic method, a kind of new function formula has been put forward to be used to measure the clustered degree of two decision members. Based on this function formula, a clustering category algorithm of the complex-large group members has been designed. Furthermore, the preference vector of the complex-large group, the algorithm of group coherence index and the algorithm of group decision value have been designed. By use of above results, the final result of group decision making is be obtained. Further the multi-schemes large lroup lecision making is obtained. In order to make the process of group decision making to go on smoothly, the processing mechanism of group decision making has been designed which includes the device of processing process of group decision making and the device of clustering process of group members.( 4 ) The research of the architecture of complex-large group decision support system in network environmentIn the support of model-drive and decision making method of complex-large group, combing the characteristic and developing tendency of network, the layerstructure of the systematic platform has been studied using layer model theory. This platform is divided into three layers. It is the decision front layer, decision service layer and decision resource layer. To normalize decision process, the information flow of systematic built-in has been designed. The systematic function structure and the function modules of the each layer of the system have been designed. The decision globe control subsystem, data management subsystem, model management subsystem and knowledge management subsystem have emphatically been studied and designed, which includes the processing process of each subsystem and its function realization method. With this, form the architecture of complex-large group decision support system in network environment.(5 ) The research of system realization method and its application wayIn system realization method, the component configure and the component operation mechanism of group decision have put forward based on the component developing technology and multi-agents technology, and the gross structure of system component and its internal relation have been studied. According to the need of application system developing, the using way of system component in application system developing and the coordinative mechanism of each system component in solving of decision problem have put forward to make user to be able to quickly developing the application system with group decision function.(6) The empirical studySelect certain representative famous copper group company’s overall budget management group decision support system as the empirical case. Through the form of project cooperative research, the demand of the copper group company’s overall budget management group decision has systematically been analyzed. The gross structure of overall budget management group decision support system of the group’s key enterprise "Copper stock corporation" has been designed, in which the group decision basic condition and decision members have been explained. The group decision process and systematic function structure has carefully been designed. The realization method of every subsystem has carefully been studied. Through system developing, the result shows that the expectant effect has been reached.更多还原