QQ客服 官方微博 反馈留言

题名	普适环境下服务发现的语义支持机制
作者	李玉明
学位类别	博士
答辩日期	2008-06-06
授予单位	中国科学院软件研究所
授予地点	软件研究所
关键词	普适计算 服务发现 近似匹配 上下文敏感 QoS 服务选择 Pareto
其他题名	semantic support mechanisms for service discovery in pervasive computing
中文摘要	随着计算能力和通信能力的增强，计算设备所占用的体积也越来越小，各种新形态的传感器、移动设备及无线网络设备日趋普及。这极大地促进了以无线、移动和嵌入式设备为基础的普适计算模式的形成和发展。普适环境下的服务发现机制可使得用户能够使用各种移动设备无缝的、随时随地的共享和访问各种服务信息。 普适环境中的网络异构性和动态性，以及服务的多样性和异构性，均对服务发现提出了新的挑战。目前学术界和工业界在服务发现方面进行了不少的探索和研究，当前的一些服务发现协议或系统比如SLP、UPnP、INS等，一般基于语法的方法描述服务，主要关注服务的功能性需求，但基于关键字来匹配请求和服务描述，在普适环境下常常会导致较差的匹配结果。 针对服务发现在普适环境中的新需求，本文基于XML定义了一种轻量级的服务语义建模语言SML，SML定义了丰富的数据类型，以模板和属性的方式定义各领域的实体，可以表达丰富的语义知识；同时，本文以轻量级的推理引擎Jess为依托，将用服务建模语言定义的各应用领域的服务模板和语义知识自动转换成Jess的推理规则和事实。本文定义了一种类似XPath的服务查询语言规范，并在支持精确匹配的基础上，提出了服务的近似匹配策略，提供了各种的近似计算规则。鉴于不同用户对服务的各属性有不同的偏好程度，还提出了基于用户偏好的服务匹配策略。动态的上下文信息是服务匹配过程的重要考虑因素。本文以Jess规则来匹配服务和用户的上下文，选择适合于用户当前情况的服务。对服务各种QoS的描述支持也是服务语义建模语言提供的功能之一，为此，本文还提出了一种基于Pareto最优的服务选择策略，根据服务的QoS以及服务与请求的匹配程度来选择Pareto最优的服务。本文的上述工作已实现到服务发现系统Service CatalogNet Extended中。
英文摘要	With the increasing capability of computing and communication, computational devices become smaller and smaller; and new sensors, mobile devices and wireless network equipments become prevalent. This has greatly engendered a new computing paradigm, called pervasive computing, which is based on connected wireless, mobile and embedded devices. Service discovery is essential in pervasive computing environments so that people can seamlessly share and visit varied services through mobile devices at anytime anywhere. The heterogeneous and dynamic characteristics of network and the complexity and diversity of services have brought new challenges for service discovery in pervasive computing. Some work has been done for solving this problem from both academy and industry. Current protocols or systems, such as SLP UPnP, INS, etc. adopt syntactic methods to describe services, focusing on function requirements; they match service requests with service descriptions based on keywords, which often lead to a poor quality result in pervasive environment. This thesis focuses on semantic support mechanisms for service discovery in pervasive computing. By arguing the importance of semantics in service discovery, a light-weighted service semantics modeling language SML is proposed. SML, providing various built-in data types, can define entities in specific application domains with templates and attributes. On the basis of search engine JESS, a lightweight reasoning engine is constructed so that all service templates and semantic knowledge are automatically converted to Jess facts and rules. Meanwhile, a service query language similar to XPath is defined. Besides the exact service matching, the thesis proposes an approximate service matching policy, which provides a set of rules for approximately calculating. During the matching process, the client’s preference on service attributes is also considered. Context is one of the important factors in pervasive computing, so the services whose contexts accord with user’s contexts are matched in the aid of JESS rules. QoS description is one of the functions of SML and is used in service selection. The thesis proposes a Pareto optimal policy for service selection based on the QoSs of the services and the matching degrees between service and request. All above work has been implemented into a service discovery system named Service CatelogNet Extended.
语种	中文
公开日期	2011-03-17
页码	81
内容类型	学位论文
源URL	[http://124.16.136.157/handle/311060/5808]
专题	软件研究所_软件工程技术研究开发中心 _学位论文
推荐引用方式 GB/T 7714	李玉明. 普适环境下服务发现的语义支持机制[D]. 软件研究所. 中国科学院软件研究所. 2008.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。