面向个人服务云的动态服务合成的上下文计算框架及技术的研究与实现
发布时间：2018-08-28 08:26
【摘要】：泛在网络的发展推动大量可联可控的设备节点进入网络,将这些设备节点的服务能力封装并在统一平台上管理,将产生面向用户的个人服务云。SOA为通过简单服务生成复杂业务提供了规范的机制,上下文计算可依据上下文信息按策略指导应用或业务的执行,SOA与上下文计算相结合使得动态的服务合成成为可能。面向个人服务云的动态服务合成的上下文计算拥有广阔的应用前景。但是,传统上下文计算存在以下问题：(1)上下文情景推理模型定义不完善,不利于情景的更新与用户的操作；(2)情景推理机制不完善,多停留于定性分析而不能得到定量结论；(3)指导业务逻辑的策略机械,不够精确、灵活,且不能适应用户偏好的演变,通过自主学习更新策略。 为解决上述问题,本文首先给出了面向个人服务云的动态服务合成的上下文计算系统框架,将CBR引入上下文计算系统。其次,在此基础上设计了基于CBR的情景模型和情景推理机制,将情景模型映射为CBR中的案例模型,通过CBR既有的案例管理机制实现情景的管理,将情景推理映射为CBR中的检索环节,通过相似性计算得到情景似然度。然后,本文设计了服务信息模型和相应的服务评估机制与反馈处理机制,以前述定量的情景似然度和服务信息模型中的偏好为条件完成精确的服务评估和灵活的服务选择,以用户对服务评估结果的反馈为条件完成服务偏好的自主学习与更新。之后,本文还通过一个智能家居环境下的业务流程验证了上述系统功能。 最后,本文总结了全文的主要工作,阐明了今后的主要研究方向。
[Abstract]:The development of ubiquitous network drives a large number of connected and controllable device nodes into the network, encapsulates the service capability of these device nodes and manages them on a unified platform. Will generate a user-oriented personal service cloud. SOA provides a canonical mechanism for generating complex business through simple services, Context computing can be based on context information to guide applications according to policy or the combination of business execution SOA and context computing makes dynamic service composition possible. The context computing of dynamic service composition for personal service cloud has a broad application prospect. However, the traditional context computing has the following problems: (1) the definition of contextual scenario reasoning model is not perfect, which is not conducive to the updating of the situation and the operation of the user; (3) the strategy machinery guiding business logic is not precise, flexible, and can not adapt to the evolution of user preferences, and update the strategy through autonomous learning. In order to solve the above problems, a context computing framework for dynamic service composition for personal service cloud is presented in this paper, and CBR is introduced into the context computing system. Secondly, the scenario model and scenario reasoning mechanism based on CBR are designed, the scenario model is mapped to the case model in CBR, and the scenario management is realized through the existing case management mechanism of CBR. The scenario inference is mapped to the retrieval link in CBR, and the likelihood degree of the scenario is obtained by the similarity calculation. Then, this paper designs the service information model, the corresponding service evaluation mechanism and the feedback processing mechanism, taking the above quantitative scenario likelihood degree and the preference in the service information model as the condition to complete the accurate service evaluation and the flexible service choice. Self-learning and updating of service preference is completed on the basis of user feedback on service evaluation results. After that, this paper verifies the above system function through a business process in a smart home environment. Finally, this paper summarizes the main work of the paper, and clarifies the main research direction in the future.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
本文编号：2208838
[Abstract]:The development of ubiquitous network drives a large number of connected and controllable device nodes into the network, encapsulates the service capability of these device nodes and manages them on a unified platform. Will generate a user-oriented personal service cloud. SOA provides a canonical mechanism for generating complex business through simple services, Context computing can be based on context information to guide applications according to policy or the combination of business execution SOA and context computing makes dynamic service composition possible. The context computing of dynamic service composition for personal service cloud has a broad application prospect. However, the traditional context computing has the following problems: (1) the definition of contextual scenario reasoning model is not perfect, which is not conducive to the updating of the situation and the operation of the user; (3) the strategy machinery guiding business logic is not precise, flexible, and can not adapt to the evolution of user preferences, and update the strategy through autonomous learning. In order to solve the above problems, a context computing framework for dynamic service composition for personal service cloud is presented in this paper, and CBR is introduced into the context computing system. Secondly, the scenario model and scenario reasoning mechanism based on CBR are designed, the scenario model is mapped to the case model in CBR, and the scenario management is realized through the existing case management mechanism of CBR. The scenario inference is mapped to the retrieval link in CBR, and the likelihood degree of the scenario is obtained by the similarity calculation. Then, this paper designs the service information model, the corresponding service evaluation mechanism and the feedback processing mechanism, taking the above quantitative scenario likelihood degree and the preference in the service information model as the condition to complete the accurate service evaluation and the flexible service choice. Self-learning and updating of service preference is completed on the basis of user feedback on service evaluation results. After that, this paper verifies the above system function through a business process in a smart home environment. Finally, this paper summarizes the main work of the paper, and clarifies the main research direction in the future.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
【参考文献】
相关期刊论文 前1条
1 高俊杰;邓贵仕;;基于本体的范例推理系统研究综述[J];计算机应用研究;2009年02期
，本文编号：2208838
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2208838.html
