【作者】 黄逸民；

【导师】 王树青；

【作者基本信息】 浙江大学 ， 控制理论与控制工程， 2002， 博士

【摘要】 智能管理信息系统是在传统管理信息系统的基础上，应用人工智能技术，设计实现的具有智能化、集成化、协调化特点的新型计算机管理系统，智能管理信息系统的设计与实现是系统工程研究与人工智能应用研究的热门问题。Agent技术作为分布式人工智能研究中的新成果，多Agent系统的应用研究已成为人工智能领域的一个重要研究方向，并且Agent技术在网络管理、智能机器人、分布式智能控制、信息检索与服务等应用领域取得了巨大的成功。本文的研究在国家自然科学基金项目“基于多智能体的企业信息集成理论及应用研究”(No.79970035)资助下进行的，作者将Agent技术引入到智能管理信息系统研究中，以多Agent系统为基础研究基于多Agent的智能管理信息系统的理论与应用，设计智能管理信息系统的框架模型，给出其实现策略。作者采用面向Agent的软件工程技术来研究和实现管理信息系统中的信息处理、信息查询、辅助决策和数据挖掘等功能子系统，并给出具体应用实例。本文的主要研究工作内容总结归纳如下： 1．介绍了智能管理信息系统的概念，综述了Agent技术和多Agent系统的原理、发展过程及其应用发展现状。 2．通过对智能管理信息系统的系统分析，提出基于Agent技术实现智能管理信息系统的方法，设计整个系统的体系结构和Agent的内部结构并给出Agent的定义，分析了整个系统的设计实现的过程与方法，研究了管理信息系统中Agent之间的协调机制、复杂问题的求解机制和多Agent系统的通讯机制。 3．从企业的组织机构和流程分析入手，提出了支持动态企业模型的基于多Agent的智能管理信息系统，并研究了该系统如何实现对企业的组11 浙江大学博士学位论文织机构和业务流程动态调整的支持。同时从智能人机交互系统的应用入手，分析了智能人机交互系统应具备的功能，提出了基于Agent技术的智能人机交互系统的体系结构，并重点讨论了接口 Agent的设计、机器学习的算法和智能人机交互系统的工作流程。 4．将Agent技术应用于信息查询系统中，从智能综合查询系统功能设计和实际应用出发，分析基于Agent技术的智能综合查询系统的体系结构和系统的工作流程，给出其采用的模糊优化查询算法，总结了采用Agent技术实现智能综合查询系统的主要优点。 5．采用面向Agent的系统分析与设计的方法，分别根据决策支持和数据挖掘的应用模型，分析了基于Agent技术的智能决策支持系统和数据挖掘系统的体系结构，研究了各自的实现技术与方法，总结了各自的主要优点。 6．作为基于Agent的智能管理信息系统理论的一个应用实例，根据检测中心的业务分析并结合其实际特点和管理现状，综合运用CORBA技术和 Agent等技术，构建了一套检测中心智能管理信息系统，详细分析整个应用系统的功能设计、体系结构、实现方法及应用效果。 最后总结了全文的研究工作，归纳了基于多Agen上技术去设计和实现智能管理信息系统的优势，并对该研究方向的未来发展进行了展望。更多还原

【Abstract】 IMIS (Intelligent Management Information System) is an advanced computer management system of high coordination, intelligibility, and information integration by means of artificial intelligence. It develops from management information system (MIS) and becomes an important research field of system engineering and" artificial intelligence. With Agent bearing fruits in network management, intelligent robot, distributed intelligent controlling, information indexing and service, the application research of Multi-Agent System (MAS) has accordingly become a more and more important research orientation. This work is supported by National Nature Science foundation of P. R. C with the project entitled "Multi-intelligent-based enterprise information integration theory and application research" (No. 79970035). The main contribution of this work is that Agent is introduced into IMIS, and based on the Multi-Agent, the theory and application of Multi-Agent-based IMIS are studied, the IMIS model is proposed and its implementation strategy is further discussed. The writer designs and implements information management system, information query system, decision supporting system and data mining system of IMIS by means of orient-Agent software engineering. This dissertation is devoted to theoretical and practical study of IMISbased on MAS. The main content is listed as follows:1. The concept of IMIS is firstly introduced. The historical development of Agent and MAS is reviewed from theory and practice, and the prospect as well.2. This paper proposes Agent-based IMIS after a careful study of IMIS, designs system architecture, and the innate structure and conceptualization of Agent, discusses in detail the implementation procedure and methodology. Based on previous research related with Agent, this paper suggests Agent-based IMIS coordination method, complex problem solving strategy and Agent’ s communication mechanism.3. This paper, through the analysis of enterprise organizing and process, proposes the Multi-agent-based IMIS in supporting the dynamic enterprise model, and then discusses in detail how to realize the dynamic adjustment of the enterprise organizing and business process. In addition, the application study of human-computer interaction leads to a discussion of the functions it should have. Great emphasis is attached to the designing of interface Agent, calculation of machine learning and the working procedure of intelligent human-computer interaction.4. Agent is introduced into the information query system (IQS), following the system designing, the intelligent IQS architecture, working procedure and fuzzy optimization query algorithm is further analyzed, and the main advantagesof intelligent IQS are summarized.5. Agent is introduced into decision supporting system (DSS) in analyzing the subsidiary decision models and methods. We, thereby, propose the Agent-based intelligent DSS architecture, Agent innate structure, Multi-Agent working strategy, and the implementation skill. The main benefits of adopting Multi-Agent in realizing intelligent decision supporting system are also elucidated. Agent is also introduced into data mining system (DMS) in analyzing the data mining models and methods. The Agent-based intelligent DMS architecture is proposed, and the DMS process and main advantages are provided.6. As an example of the Agent-based IMIS, a set of center measurement is constructed fully employing technology such as CORBA and Agent based on the business analysis in the measurement center with present features and management situations. The system designing, architecture, implementation and application are also discussed.The dissertation concludes with a summarization of advantages of IMIS designing and implementation, and a series of vistas in this research orientation.更多还原