【作者】 严旭东；

【导师】 饶若楠；

【作者基本信息】 上海交通大学 ， 软件工程， 2010， 硕士

【摘要】 普适计算技术正随着计算机网络,传感器网络和移动设备的发展得到了越来越多的关注和研究。在普适环境下,设备种类繁多、功能和能力各异、网络通信协议多样、资源状态动态变化,组合服务的可靠性会随着上下文环境变化而发生变化,其可靠性的度量、预测与验证是一个亟待解决的问题。针对上述问题,本文提出一种基于规则推理的方法对组合服务进行可靠性度量、预测和验证。该方法根据普适环境的特点建立基于本体的上下文模型和服务可靠性度量模型,通过上下文感知并使用基于规则的推理方法度量及预测服务的可靠度,然后参照目标需求对服务的可靠性进行验证。在此基础上,本文阐述了原型系统的设计与实现,通过在具体场景中的应用,验证了该方法的可行性与有效性。与传统的方法相比,本文提出的方法具有以下特点:第一,本文提出了一种基于本体的服务可靠性度量模型,同时使用基于规则的推理方法对基本服务及其组合服务的可靠性进行度量及预测;第二,本文提出了两阶段的服务可靠性度量及预测方法,先通过上下文感知及推理计算出基本服务的可靠度,再根据组合服务的控制结构通过推理得出组合服务的可靠度。更多还原

【Abstract】 Pervasive computing is getting more and more spotlights along with the development of computer network, sensor network and mobile technology. In pervasive environment, devices have different type and provide disparate functions while various protocols of networks exist and status of resources change dynamically, which causes the reliability of services varies as the context environment is changed. The measurement, forecast and verification of the service reliability is a main issue.Aiming at the problem, this paper brought forward a rule-based reasoning method to measure, forecast and verify reliability of composite services. According to the characters of pervasive environment, a context model and service reliability measuring model based on ontology technology were established. Services reliability was computed by rule-based reasoning through context awareness, and then a verification method of service reliability referred to the request was proposed. Based on the above method, the design and the realization of the prototype system was discussed in detail. And experiments in the scene also proved that the method is feasible and effective.Compared with the traditional methods, this method in this paper has the following points: Firstly, this paper introduced a service reliability measurement model based on ontology technology and computed the reliability of atomic services and their composite services through rule-based reasoning. Secondly, this paper used two-stage method for service reliability measurement and forecast, which computed the simple service reliability through context awareness and reasoning, and then measured the reliability of composite services according to various of control construct.更多还原