【作者】 肖芬；

【导师】 刘发贵；

【作者基本信息】 华南理工大学 ， 计算机软件与理论， 2013， 硕士

【摘要】 分布式计算一直是计算机科学领域的一个重点，Web服务自从问世以来就成为分布式计算领域的宠儿。如今随着Web服务技术的不断成熟，它在各个领域得到了广泛的应用。越来越多的Web服务被开发并发布，有效的服务发现技术成为Web服务继续推广的重要动力。语义Web服务为服务描述添加了语义支持以期提高服务的发现能力。本文以语义Web服务发现为中心，分别从服务功能发现与服务QoS选择两个方面提高服务发现的质量，并最终实现了一个具有QoS感知的语义Web服务发现系统。服务的功能发现以本体概念相似度的计算为基础。针对现有研究成果存在的不足，本文提出了一种基于直觉模糊集的概念相似度计算方法，它利用本体中的关系将概念转换成一个直觉模糊集，并选用一个合适的直觉模糊集相似度计算方法来计算概念的相似度。该模型调动了概念的所有语义信息进行相似度计算，它不仅适用于简单的单继承体系的本体，更考虑了复杂本体中多继承、区别于父概念的属性概念对概念相似度的影响。在服务QoS选择方面，本文提出了一种融合专家意见与用户偏好的服务QoS选择方法：使用直觉模糊语言来作为评价语言，以多位Web服务专家对服务的多个QoS指标的主观评价为基础，形成一个多指标群决策问题；将用户的偏好作用于专家的意见上从而影响最终的决策结果；最后结合用户的反馈来得到服务QoS的综合排名。方法使得QoS选择既有专业意见的支持，又允许用户根据需求提出特殊要求，从而使得选择结果同时具备共性与个性。基于上述研究成果，本文实现了一个具有QoS感知的语义Web服务发现系统。以物流运输为背景，本文最后通过构建领域本体与设计相关语义Web服务，对系统的性能进行了测试与分析。实验结果表明系统在查准率、查全率及响应时间方面均具有较好的表现。更多还原

【Abstract】 Distributed Computing is a research focus for Computer Scientists, and Web Servicebecame a star in this field the moment it was introduced. With the increasing development ofthe related techniques, nowadays Web Service has been widely used in various fields. Moreand more Web Services are being developed and published, demanding more effective servicediscovery techniques for its continuing momentum of development. Semantic Web Service isproviding semantic supports for web service description in order to improve discoverycapabilities. Concentrated on semantic web service discovery, in this paper, we try to improvethe result of which from both aspects of functionality discovery and QoS selection, and a QoSaware semantic web service discovery system is developed as a byproduct.Functionality discovery is based on concept similarity. Witnessing the shortness ofrelated achievement state of the art, in this paper, an intuitionistic fuzzy set (IFS) model isproposed to calculate the concept similarity. The model first transforms the concepts to IFSsusing relations from the ontology, and then chooses an effective IFS similarity measure tocalculate concept similarity. The model takes into account all available semantic relations ofconcept for similarity calculation; it is designed to handle not only simple ontologies as aperfect tree structure but also complicated ones where multiple inheritance and differentiatedproperties are presented.On service QoS selection, we propose a method that combines expert opinions with userpreference in QoS selection. Using IFS linguistic variable, a group of expert are asked to rateon multiple QoS criterion of services, which forms into a multi-criterion group decisionmaking model; then a specific user performs his own preference on experts opinion and thusmaking a sound on the result of group decision making; finally users feedback is consideredwith the DM result to yield an ordered list. This method ensure that QoS selection hasprofessional opinion as back-up and at the same time allows users to declaim their owndemands, thus the result is of common and yet of individuality.With the above achievement in hand, we develop a QoS aware semantic web servicediscovery system to inspect our theories. A field-ontology of logistic transport is designed andrelated semantic web services are set up to test our system. The result proves goodperformance of our system in precision, recall and response time.更多还原