Abstract: In active service environment, Web services are normally viewed as intelligence agent entities and they can usually aggregate around some requirements to form a magnetic effection phenomenon, because of their ability of seeking and discovering the requests that they can participate in. According to this special aggregation platform driven by requirement for autonomous services, this paper proposes a graph-planning algorithm to build up the composition service requirement model based on the theory of environment ontology. Then, it defines the composition mode concept as a form of logic solution to composition requirement from the perspective of the functional fragment. Besides, it introduces CPEA (combination pattern extraction algorithm) to describe how to extract every suitable composition models from a planning-graph-structure request. Finally, it uses the specific example of travel arrangement to verify the theory referred above.

Key words: active Web service environment, environment ontology, planning graph, composition service requirement, combination pattern

摘要： 在主动服务环境中，Web服务通过发挥智能Agent实体的主动特性，来搜索和发现环境中可参与的服务请求，并最终形成一种需求驱动的“磁石效应”现象。针对这种特定的自主Web服务环境，以环境本体作为基础描述背景，借鉴图规划算法的相关理论，提出了一种基于规划图的组合服务需求模型。在功能层面上，从服务需求的各个可行的逻辑实现方式角度来考虑，定义了组合模式的概念，并给出了相应的组合模式提取算法。最后通过一个经典的旅游安排案例验证了以上理论研究的可行性。

关键词: 自主Web服务环境, 环境本体, 规划图, 组合服务需求, 组合模式