节点文献
基于本体的网格元数据模型及其相关技术研究
Study on Otology-Based Grid Metadata and Related Technology
【作者】 杨洪山;
【导师】 陈家训;
【作者基本信息】 东华大学 , 控制理论与控制工程, 2008, 博士
【摘要】 网格的发展和应用对网格资源的发现和管理提出了新的要求,网格资源管理的目标是要把通过网络连接起来的各种自治资源和系统进行集成并统一管理,提供可扩展的、安全的、一致的、普及的、高效的大规模资源共享和计算服务。因此,网格资源管理也就成为了网格研究中的核心问题,但是网格资源所具有的动态性和异构性特点对资源的发现和调度提出了很高的挑战。要求网格资源管理的方式更加具有智能性,以适应网格资源动态性变化、资源异构、资源管理规模和范围巨大的需求。网格资源本身的异构性、应用的差异和地域的差异等也带来了资源描述元数据的异构性,这对网格资源的发现和互操作都带来很大的影响。本论文对相关技术进行了全面综述,总结了这些领域的研究现状和发展趋势,同时指出了现有工作的不足。结合相关领域的最新成果,针对其不足之处,提出了基于本体的网格资源元数据语义模型,研究了元数据抽取和语义化算法,以及基于语义元数据的网格资源注册和管理技术,并探索了资源语义匹配算法和资源发现系统框架。并且将研究成果应用于政务网格中,解决电子政务领域的资源共享和业务协同问题。首先,本文对网格资源进行抽象分析、对网格资源元数据进行模型化分析,提出了网格资源抽象模型和元数据抽象模型,并采用本体建模技术设计了元数据本体。结合领域本体实现了机器可读的网格资源语义元数据,为基于元数据的语义资源发现奠定了基础。其次,本文对网格资源的元数据管理进行了论述,提出了网格资源元数据的方案,对元数据的获取、语义化处理和元数据注册,以及元数据的互操作进行了系统的研究。结合本体和语义标注技术设计了网格元数据语义化处理算法,使网格资源元数据的语义可以通过本体中的概念准确的表述,实现了对网格资源语义元数据的采集和管理。再次,网格中语义元数据的建立为网格资源发现奠定了基础,为了实现网格内资源的语义发现和管理,本文还研究了网格资源的语义匹配策略和发现算法。论文分别对概念的语义相关性和语义距离进行研究,设计了本体中概念语义相似度计算的算法。通过本体概念语义相似度的分析和语义辨析的研究,提出了基于语义的网格资源相似度算法。在此基础上,实现了网格资源和用户需求的语义匹配,并对算法的扩展进行了探讨。然后,对网格资源发现的机制、发现方式、发现服务的设计要素和操作进行了论述。提出了基于元数据和本体的网格资源发现模型,研究了多本体的协作技术,设计了网格资源语义发现系统的框架,并对框架结合个人资源代理的扩展方面进行了探讨。最后,分析了电子政务建设中资源共享和业务协同的需求,提出了政务网格体系,通过网格技术实现政府部门之间软硬件资源的共享。结合上海市电子政务建设的情况,在几个政府部门之间搭建了政务网格的实验环境,初步建立了电子政务网格资源元数据和其语义化,开发了政务网格资源的编目和注册系统,并且开发了基于本体对资源元数据进行了统一的语义化处理和转换的系统。实现了网格资源的目录服务和资源的语义发现,同时对政务网格资源的调用和交换做了部分实现。本论文对于网格资源的元数据模型和网格中基于元数据和本体实现资源的语义发现进行了系统的研究,分别建立了网格资源抽象模型,元数据抽象模型、元数据本体模型和元数据语义模型;研究了网格资源语义元数据的管理技术,设计了网格资源元数据方案、元数据抽取和元数据语义化算法,并对元数据注册的机制和元数据的互操作进行了系统研究;并研究了基于元数据和本体的网格资源匹配策略和发现算法,实现了基于语义相似度进行网格资源匹配和发现;研究了网格资源发现模型,设计了网格资源语义发现的系统框架;最后,把课题的研究成果应用于电子政务建设中,设计了政务网格体系,搭建了政务网格实验环境,对元数据的管理、资源发现和调度进行了应用实验研究,并且已经在一些实际项目建设中应用了本文的部分研究成果。
【Abstract】 The development and application of Grid have put forword new requirements on resource management continuously.The goal of Grid resource management is integrating all kinds of autonomic resources and applications,and managing them unitedly to provide expandable,security,unanimous,pervasive,high efficency and wide scope resource sharing and computing service.So that Grid resource management has become one of the core problems in Grid research area.But the dynamic and heterogenous characteristics of Grid resources bring high challenge on resource discovery and allocation.It requires the resource management mode to be more intelligent,and adapt the requiremtn of resource dynamicly changing,heterogenouse, large scale and scope.The difference of Grid resouces,applications and regions also make the resource description mentadatas heterogeneity.This may bring a lot of influence on Grid resource discovery and interoperability.This paper described the related technology completedly,summarized the researching status and development tendency briefly,and indicated the deficiency of the current works.Combining the newliest production of related researching area and aiming at the lack of them,we promoted an ontology-based Grid resource metadata semantic model,investigated the metadata extracting and semanticnizing algorithm and Grid resource registry and management technology.We explored the resource semantic matching arithmetic and discovery system framework.Also,we applied the researching results in Government Grid to solve the problem of resource sharing and business cooperating problems in e-Government area.First,this paper made abstract analysis on Grid resources and made modeling analysis on metadata.It put forward abstract model of Grid resource and metadata.So we designed metadata ontology using ontology modeling technology,and realized the semantic Grid resource metadata,which is computer readingable,combining domain ontologies.This provided basis for semantic resource discovery.Next,this paper discussed the Grid resource management technology.It provided Grid resource metadata scheme and do some systemic research on metadata extraction, metadata semanticnizing transaction,metadata registry and metadata cooperation. Combining ontology and semantic annotation,we designed the Grid metadata semantic transaction arithmetic,so that the semantics of metadata can be expressed by concepts of ontology exactly and realized the semantic metadata collection and management.Then,the construction of semantic metadata in Grid establish basis for Grid resource discovery.But we need study the semantic matching policy and discovery arithmetic to realize the semantic resource discovery and management.The paper had studied the semantic relativity and semantic distance between concepts in ontology,and designed the semantic similarity algorithm between concepts.By the research of semantic similarity and semantic discriminating between concepts,we brought forward the Grid resource similarity algorithm based-on semantics.Laying the foundation of the research,we realized the semantic matching among Grid resources and customer’ s requirements,moreover we have discussed the expansion of aligrithm.Moreover,we discribed the discovery mechanism,discovery mode,discovery service’ s key element and discovery operation of Grid resources.So we put forward a Grid resource discovery model based on metadata and ontology.We have researched the muliti-ontologies collaboration technology and designed Grid resource discovery system framework.Furthermore,we have discussed the expansion of the framework by combining personal resouce agent.Finally,we combined the requirement of resource sharing and business collaboration,and put forward the Government Grid,which realized the hardware and software sharing among government departments by Grid technology.According to the Shanghai e-Government construction condition,we built a testing environment of Government Grid among a few departments.In the validation of our research,we designed the Government Grid resource metadata ontology and e-Government ontology, and developed metadata editor,registry and catalog service systems.Moreover,we programmed the ontology-based metadata semantic transaction and transform system. We realized the Grid resource catalog service and semantic discovery.At the same time, we partly applied the resource dispatching and exchanging.The paper made sytemic researched on Grid resource metadata model and ontology-based semantic Grid resource discovery.Separately,we built the Grid resource abstract model,metadata abstract model,metadata ontology model and semantic metadata model.We studied the semantic metadata management technology in Grid, designed Grid resource metadata shceme,metadata extraction and semantic transaction algorithm,and done research on metadata registry and interoperability.Also we developed Grid resource matching and discovery algorithm based on metadata and ontology and realized the resource matching and discovery by semantic similarity.We researched the Grid resource discovery model and designed the semantic resource discovery sytem framework.Lastly,we using the researching results in e-Government construction,designed the Government Grid System,built a validitory environment,and do applying experiment of metadata management,resource cataloging,discovery and alloction.Further more,some of the research achievements have been used in the practice projects.
【Key words】 Grid; Grid Resource; Metadata; Abstract Model; Semantic; Ontology; Resource Descoverry; Government Grid; e-Government;