中文摘要：

接到请求时，如何快速准确地查找满足用户需求的服务是服务发现的目的所在．现有的服务发现方法大都通过对服务功能属性的匹配计算来进行服务查找，对服务的过程模型并没有进行相应地匹配，导致服务的查找效率较低．本文提出在服务功能属性的相似度计算时，将服务的输入输出参数根据其依赖关系进行配对，同时将服务过程模型转换成对应的有向树表示，通过对树同构的判断和有向树中结点及边的相似度计算得到对应服务过程之间的相似度，并在此基础上设计了服务发现框架．实验表明该方法有较高的查准率和较快的查询响应时间．

英文摘要：

The discovery of suitable web services for a given requirement is one of the critical operations in service architectures. Most of the exiting discovery algorithms focus on the service functional property calculation, and the match of service process model has not been taken into account in usual. But this leads to the low efficiency and low precision. In this research, the input and output parame- ters of services are divided into matching pairs according to their dependencies in the similarity calculation of service functional proper- ties. The process model of service is transformed into corresponding directed tree at the same time, and the similarity between proces- ses is obtained by the isomorphism judgment of the directed tree and the similarity calculation of tree＇ s nodes and edges. And the service discovery framework is also designed in this paper. Experiments show that the efficiency of web service discovery in precision and time consumed is improved significantly by the new service discovery algorithm.