Abstract: With regard to the problems like flow control and throughput of Web service composition, a topic-map-based semantics Web service composition engine was proposed, in which some Web services would be described via ontology and topic map. Web services and topic maps were abstracted by using these description languages which included OWL-S and XTM. Web service topic was employed to express semantics Web service, which aimed to use characteristic of topic maps. And topic maps expediently implemented positioning and navigation of information resource, such as Web services, which made states and goal of Web services definite. Thus, to define UDDI registration(UDDIr) and retrieval capability(UDDIs), grammar checking (GD) and semantics recognition capability (SI), similarity (STS), matching (STM), sufficiency (STF) of Web service topics, facet classification capability of Web service topics characteristics (SFC), access right of Web service topics (STP), facet depth decision (FDD) and interoperability capability (STC) to set up service composition engine approach, and employ mathematical method to describe and define. Finally, Java was employed to develop Web service composition engine. The experimental results show that it is feasible and effective.

Key words: semantic Web service, ontology, Web service composition engine, Web service topic

摘要： 针对Web服务组合流程控制、服务组合吞吐率等问题，提出一种基于主题图的语义Web服务组合引擎方法，通过基于本体的语义Web服务描述，并将Web服务与主题图进行抽象，用一种服务主题去表示语义Web服务，以便使用主题图的特征，而主题图可方便地实现语义服务导航定位，使得Web服务所处状态可定，目标明确。因此该语义Web服务组合引擎是从UDDI注册(UDDIr)和查询能力(UDDIs)，语法检测(GD)和语义识别(SI)能力，服务主题图的相似度(STS)、匹配度(STM)和适应度(STF)，服务主题特征刻面分类能力(SFC)，服务主题权限访问程度(STP)，刻面深度判定(FDD)，服务主题协同能力(STC)角度去建立；并通过数学方法进行定义。最后通过用Java编程实现实验模型表明，该方法可行且有效。

关键词: 语义Web服务, 本体, 服务组合引擎, Web服务主题