【作者】 赵德新；

【导师】 冯志勇；

【作者基本信息】 天津大学 ， 计算机应用技术， 2008， 博士

【摘要】 随着网络技术的发展,计算越来越呈现出移动和普适的特点,在动态、异构的网络环境中发现合适的服务是实现信息共享、复用的重要前提。传统的语法层次上的服务发现协议在查全率和查准率上都无法满足人们对智能化、个性化服务的需求。解决上述问题,需要计算实体能够理解和处理网络中的信息,即需要“语义”层次上服务发现的支持。目前语义服务发现的研究还处于探索阶段,特别是对于具有不确定信息的服务发现还没有成熟的理论和技术。本文在模糊逻辑理论基础上,对带有模糊知识的语义服务发现问题进行了深入研究,重点关注服务描述和服务匹配这两个关键问题,取得了下列一些研究成果:1.给出了描述逻辑SHOIN的模糊扩展SHOIN-f,将其作为知识表示的形式化基础。将模糊合成关系、模糊数量、模糊修饰词以及模糊语言值引入到描述逻辑SHOIN中,增加了描述逻辑对模糊信息的表达和推理能力。与其它模糊描述逻辑相比,SHOIN-f不仅对描述逻辑的语义进行了模糊扩展,而且还在语法上进行了扩展,增加了描述逻辑的语法构造子,为服务描述语言奠定了逻辑基础。2.提出了基于本体的服务描述语言FEOWL。由于本体语言具有强大的演绎推理能力,目前采用本体来表示服务的语义信息是一种最佳的解决方案。本体语言FEOWL以模糊描述逻辑SHOIN-f作为其逻辑基础,是对OWL DL的模糊扩展,使OWL DL在类、属性、个体、公理等方面具有了表达和处理模糊信息的能力。3.提出了服务发现体系和服务描述模型。服务发现体系中通过服务本体将服务进行两种不同程度的抽象,分别对应抽象服务层和具体服务层,这种设计有利于服务的分类和组织,提高服务发现的效率。基于该体系,提出了一种建立在本体语言FEOWL基础上,能描述模糊信息的轻量级服务描述模型,该模型可以对服务的功能信息和模糊信息进行描述。4.提出了“二级+二层”服务匹配策略和实现方法。在服务模型基础上,将服务分两级过滤(服务分类过滤级和服务地域过滤级),分两层(服务功能匹配层和模糊匹配层)匹配。服务匹配结果是返回给用户用自然语言值表示的匹配度。这种弹性的、带过滤机制的多层次服务匹配策略有助于用户做出判断和选择,能有效提高服务发现的效率以及查全率和查准率。本文的研究工作,将模糊理论引入到服务发现中,使服务发现结果更符合客观实际,对于基于逻辑的服务描述和知识推理的研究有一定的理论和实践意义。更多还原

【Abstract】 With the development of network technology, computing presents its mobile and pervasive characters. In the dynamic and Theterogeneous T network environment, finding the appropriate service is an important premise for information share and reuse. The traditional service discovery protocol on the syntax level cannot satisfy people’s intelligent and individual demands in recall and precision aspects. To solve these problems, computing entities need to understand and process information in the network, which is the support of service discovery on the“semantic”level. Currently, the research of semantic service discovery is in the exploration phase, especially the uncertain information service discovery with no mature theory and technology. This thesis performs deep research in the semantic service discovery with vague knowledge based on fuzzy logic theory, which focuses on the two key problems: semantic service description and semantic service matchmaking, and gets some results below:1. SHOIN-f is proposed to describe the fuzzy extension of description logic SHOIN and is treated as the formal basis of knowledge representation. Fuzzy composition relationship, fuzzy quantity, fuzzy modifier and fuzzy linguistic value are introduced into SHOIN to increase description logic’s representing and deducing ability of fuzzy information. Compared with other fuzzy description logics, SHOIN-f not only extends the semantics, but also extends the syntax of the description logic. The syntax extension enriches constructors of the describe logic, and founds logical basis for service description.2. The service description language FEOWL is proposed based on ontology. Because ontology languages possess strong deduction and ratiocination ability, currently, adopting ontology to represent services’semantic information is one of the best solutions. Ontology language FEOWL is a fuzzy extension of OWL DL based on SHOIN-f and allows OWL DL to possess representing and processing ability in class, property, individual, and axiom aspects.3. The service discovery architecture and service description model are proposed. The architecture regards the services as two different degrees of abstract through service ontology, corresponding to abstract service level and concrete service level. This design is good to category and organization of service, and increases the service discovery efficiency. Based on this architecture, a service description model based on ontology language FEOWL is proposed to describe the two aspects of services, those are service capability information and fuzzy information.4.“2-stage plus 2-level”service matchmaking strategy and method are proposed. Based on service model, the services are divided to two stage filterings (service category filtering and service area filtering), two level matchmakings (service capability matchmaking and fuzzy matchmaking). The results from service matchmaking are natural languages, which presents the matchmaking degree. This elastic multi level service matchmaking with filtering mechanism helps users to make choice and decision, increases the service discovery’s efficiency, recall and precision.The research introduces the fuzzy theory into service discovery, allows the result of service discovery tally with the practical reality and presents theory and practical meaning of researches in logical service description and knowledge ratiocination.更多还原