【作者】 刘晓强；

【导师】 陈家训；

【作者基本信息】 东华大学 ， 控制理论与控制工程， 2003， 博士

【摘要】 本文针对e-Learning知识共享与适应性应用环境的构建问题，从体系结构、知识资源语义、知识资源管理、适应性知识集成等方面进行了深入研究。 采用层次结构建模方法，构建了一种基于元数据的e-Learning知识环境体系结构——KEEM，并分析了各层的功能和技术支持。通过建立由下到上的网络传输层、数据交换层、语法层、语义层、服务层和应用层，明确了e-Learning知识环境相关技术的集成方法，降低了系统分析和设计的复杂程度。KEEM体系结构具有模块化、开放性和有效性的特点，为建立不同层次的知识共享规范和知识应用方法奠定了体系结构基础。 以语义Web理论为基础，通过采用本体进行知识基建模，结合智能教学系统的有关理论和方法，建立了教育资源本体作为e-Learning知识资源语义体系，并对该体系中的各类本体进行了形式化定义和范例分析。该体系包括概念本体、知识本体、教学策略本体、学生模型本体、教学任务本体和教学应用本体。通过建立UML类图分析了教育资源本体的关联和协作能力。教育资源本体在适应性知识集成、基于本体的深度知识资源获取方面具有很好的支持能力。它为e-Learning全面知识共享和适应性应用提供了知识语义基础。 采用Web服务技术，在分析了分布式系统拓扑结构和Web信息服务方式的基础上，提出了一种新的基于Web服务的分布知识资源管理结构，并从Web服务定义、UDDI服务注册中心和LMS(学习管理系统)资源代理等方面探讨了服务层模型的各部分内容。采用UML顺序图分析了知识资源获取的应用流程，介绍了基于本体获取知识资源的方法。该结构综合了集中访问结构与点到点对等访问结构的优势，可以更好地满足知识资源的分布管理和开放获取。 首次探讨了可视化建模语言Petri网对适应性知识集成的支持。在定义了基本Petri网、着色Petri网、层次着色Petri网的基础上，建立了一个支持适应性知识集成的形式化模型——TeachNet。该模型采用色彩表现学生模型和知识特性，通过导航结构和条件表达式体现教学策略，可以实现在学生模型和教学策略支持下的适应性知识集成。它支持层次化、网状课程体第结构，并且具有形式化、可视化的特点，具有动态反映学生学习过程中的能力。该模型为适应性知识集成研究和可视化知识集成工具开发提供了一种新方法。 本文最后还介绍了一个e-Learning知识环境原型系统的实现，分析了系统总体结构、功能、数据模型以及相关实现技术，简介了几个主要模块。更多还原

【Abstract】 This thesis aims at constructing an environment for knowledge sharing and adaptive application in e-Learning. Some investigations including architecture, knowledge semantic, knowledge management, knowledge services and knowledge aggregation are made.A layered principle is suggested to model the e-Learning knowledge environment based on metadata. A layered architecture which consists six layers down to top: network transport layer, data exchange layer, syntax layer, semantic layer, services layer and application layer is proposed, the functions and techniques in several main layers and the relationships between adjacent layers are analyzed. It unifies the relative techniques such as network application, computer application, knowledge management and education technique in the same model. The layered model reduces the design complexity, and makes the system become modular, open and effective.Based on semantic Web, ontology and knowledge base model are demonstrated, with regard to Intelligent Tutoring System techniques, education resource ontology are presented and defined in formalization, some examples are given. It includes concept ontology, domain knowledge ontology, pedagogical ontology, student model ontology, teach task ontology and teach application ontology, which provide semantic for knowledge sharing and adaptive application. The conjunction of education resource ontology are analyzed in UML (Unified Model Language), which shows education resource ontology can support well in knowledge sharing, knowledge capturing and knowledge adaptive application.On the basis of analyzing the topology of distributed system, a new knowledge management structure are introduced based on Web services technique and some important aspects of it are discussed, such as the method of establishing knowledge services, UDDI service center, LMS (Learning Management System) resource agent. The procedures of knowledge capturing in several conditions are described in UML sequence diagram. This structure benefits from both Client/Server and Peer-to-peer network. It can fulfill the needs of distributed knowledge management and open services better.The visual language Petri nets is used in adaptive knowledge aggregation for the first time. After defining basic Petri nets, Coloured Petri nets and hierarchy Coloured Petri nets, a model for adaptive knowledge aggregation named TeachNet is proposed and defined in formalization. It denotes the character of domain knowledge and student mode in coloured tokens and expresses teaching strategies with navigation structure and condition expressions, so it can support adaptive knowledge aggregation involving student model and teaching strategy visual. It provides a new method for developing knowledge aggregation tools and gives a good reference for knowledge aggregation research.At last, a prototype system for e-Learning knowledge environment are designed, the system structure, functions, data schema and some relative techniques for implementation are discussed.更多还原