【作者】 刘志忠；

【导师】 王怀民；

【作者基本信息】 国防科学技术大学 ， 计算机科学与技术， 2007， 博士

【摘要】 随着网络技术,尤其是Internet技术的快速发展,面向服务计算已成为目前广泛采用的分布计算技术之一。并且,不断扩大的应用规模对面向服务计算提出了新的需求,包括服务自动化、服务之间的协作等等。作为领域共享的形式化概念模型,本体是实现服务自动化以及服务在语义层次上互操作的有效手段。但是,在开放的Internet环境下,自主用户将根据自身的需求从不同的角度和抽象层次对相同或相关的领域建立各自的本体。同时,应用也将突破领域的限制,跨领域寻求协作。因此,多本体共存是开放计算环境下应用协作所面临的一个重要问题。多本体环境为面向服务计算,特别是服务发现机制提出了新的挑战。首先,多本体环境下的服务发现需要在本体互操作的基础上进行。其次,多本体环境下的服务发现需要对服务进行良好组织并支持跨本体的服务匹配。同时,服务发现还必须在支持多本体共存的情况下保证服务发现的效率和服务匹配的有效性。针对这些“多本体环境下服务发现”的相关问题,本文的主要工作和创新包括以下几个方面:(1)针对已有本体互操作方法不考虑本体之间内部联系的缺陷,提出了一种层次化的本体模型HOM来刻画本体之间的内部联系。基于HOM,提出了一种基于HOM的本体映射算法HOM-Matchin。该方法利用层次化模型所描述的本体之间的层次化关系来计算本体实体之间的相似性,从而提高本体映射的有效性。实验表明,当本体之间具有共享本体时,算法能够有效地提高本体映射查准率和查全率。(2)为了支持多本体共存,实现服务的良好组织以及跨本体的服务匹配,在现有的面向服务基础之上,提出了一种基于本体社区的服务发现体系结构SSD_OC。该体系结构将服务根据所采用的本体组织成不同的本体社区,并以本体之间的映射关系作为社区之间的桥接,从而支持跨本体的服务匹配。在单个本体社区内,服务注册机制在UDDI基础上进行了扩展,使其能够支持基于本体的服务匹配。该机制能够较好地支持多本体共存并实现跨本体的服务发现。实验表明,该机制通过逻辑匹配能够提高服务匹配的查准率和查全率,并通过“分而治之”的思想可以提高系统的可扩展性。(3)为了缓解集中式注册中心的瓶颈问题,同时提高跨本体的服务发现性能,提出了一种基于双层P2P模型的服务发现机制。该模型以SSD_OC为基础,将P2P结构集成到了虚拟计算环境iVCE中,在本体社区内以及本体社区之间构成分别构建P2P结构。社区内的P2P结构以模块化本体为基础。以此模型为基础,提出了一个两阶段三步骤的服务匹配算法。服务匹配分为社区内匹配和跨社区服务匹配两个阶段。在社区内首先通过简单的语义相似性匹配获取适当的注册服务器后再利用逻辑推理来实现精确匹配。实验表明,通过合理的参数设置,算法可以提高服务匹配的查准率和查全率,并降低服务注册中心的平均负载。另外,通过参数调节,算法可以在查全率和平均响应时间之间达到一个折中。(4)以服务的QoS属性为研究对象,针对QoS属性的特殊性,结合层次化本体模型,提出了一个分层的QoS本体模型QoSHOnt。该模型将QoS本体分为三个部分:QoS核心层的QoS上层本体、QoS属性层的QoS中间层本体和QoS用户层的QoS下层本体。该模型通过分层机制既保证了QoS建模在QoS属性层上的一致,又保证了QoS建模在用户层的多样性。另外,该模型以不同场景下的QoS定义为基础,能够较好地支持服务等级协议SLA。以此模型为基础,结合本体社区与本体转换思想,给出了基于QoSHOnt的服务匹配和选择算法。最后,综合领域本体和QoS本体,给出了一个基于本体的服务发现框架。更多还原

【Abstract】 With the rapid growth of network technologies, Internet especially, service-oriented computing became one of popular distributed computing technology. And application progress raised new challenges to service-oriented computing, such as automatization of service, seamless collaboration of servies, and so on. Ontology, which is the shared formal conceptual model of a domain, is an effective tool to automatize service discovery, composition, monitor and achieve semantic interoperability among services. However, due to the open characteristic of Internet, autonomic users construct their own ontologies of the same or overlapped domain according their own requirements, and the collaboration among applications would break the bound of domain and occur across ontologies. Therefore the fact of multiple ontologies coexisting is one of the most important issues of applications collaboration in open distributed computing environment.Multiple coexistent ontologies challenge service-oriented computing, especially service discovery based on ontology. Firstly, ontology interoperability is the essential basis of service discovery mechanism in multi-ontology environment. Secondly, service disocovery mechanism in multi-ontoltogy environment should support multiple ontologies and service matching across ontologies. And services in multiontology environment should be well organized. At the same time, it must provide an acceptable efficiency and validity of service matching.Aiming to these related technologies of "service discovery in multi-ontology environment", the primary work and contributions of this dissertation include:(1) Because few available approaches of ontology interoperability took relations among ontologies into account, a hierarchical ontology model (HOM) is proposed to formally describe these relationships. Based on HOM, a revised ontology mapping algorithm HOM-Matching is brought forward. HOM-Matching is based HOM, and utilizing the relations between ontologies in calculating semantic similarity between ontology entities. The experimental results shown that while the aligned ontologies share the imported ontologies, HOM-Matching will improve the F-Measure of ontology mapping.(2) To support multiple coexistent ontologies, organize services well and enable service discovery across ontology, a service discovery architecture based on ontology community, named as SSD_OC, is proposed. SSD_OC divides services into different ontology communities according to the ontologies which service referred to, and bridges ontology communities using mappings between ontologies. Further, ontology transformation between ontology communities is used to achieve service matching across ontologies. Within ontology community, it extends UDDI to support service matching based on ontology. SSD_OC can support multiple coexistent ontology and service discovery across ontology. Experiemtal results show that service discovery mechanism in SSD_OC improves the Recall and F-Measure of service matching due to service matching across ontology. And SSD_OC is scalable for it practices a policy of "Divide and Conqure".(3) To ease the bottleneck effect of centralized service registers and improve the performance of service discovery across ontologies, a two layered P2P model for service discovery is proposed in this disssertation. The model is based on SSD_OC and integrates iVCE (Internet-based virtual computing environment) core concepts into a P2P model. Based on this model, a service discovery algorithm composing two stages and three steps is proposed. It matches services across communities as well as within community. Within a community, after locating registers holding service information with a high probability of satisfying a request, algorithm captures semantic matching between service advertisements and service requests by logical reasoning. Service discovery across communities occurs according to some policies. The model is suitable for opening environment with coexistent multiple ontologies. Experimental results show that given an appropriate setting, the model would make a tradeoff between recall and responding time. In addition, the model would release the mean load of registers efficiently while holding recall.(4) Finally, taking QoS of service as research object, and aiming to the particularity of QoS, a hierarchical ontology model for QoS, which is based on HOM and called as QoSHOnt, is proposed in this dissertation. QoSHOnt consists of three components: QoS Upper Ontology in QoS Core Layer, QoS Middle Ontology in QoS Attribute Layer and QoS Lower Ontology in QoS User Layer. With hierarchy, QoSHOnt guarantee the consistence of QoS model in QoS attribute layer and diversity in QoS user layer. In addition, for QoSHOnt models QoS attributes in different context, it would support Service Level Agreement (SLA) well. Based on this model, a QoS-matching and service selection algorithm is proposed. Finally a service discovery framework based on ontology is illustrated. The framework support domain ontology as well as QoS ontology.更多还原