CAJ | 학술논문

针对集中式服务发现方式容易导致单点故障和性能瓶颈的问题，提出一种基于结构化 P2P 网络的语义 Web 服务发现算法。利用本体划分技术，按概念相似度完成本体的划分并生成可偏序排序的组概念，用组概念粗粒度地表示服务的语义信息，并聚类语义相似的服务。在此基础上，设计基于组概念生成向量的服务发布/发现算法，从服务/请求中提取多个散列的关键字，在适当增加服务发布/发现冗余的情况下，实现结构化 P2P 网络中语义 Web 服务的模糊匹配。实验结果显示，与精确匹配算法相比，该算法服务发现的平均召回率由15.2%提高到了37.1%。
침대집중식복무발현방식용역도치단점고장화성능병경적문제，제출일충기우결구화 P2P 망락적어의 Web 복무발현산법。이용본체화분기술，안개념상사도완성본체적화분병생성가편서배서적조개념，용조개념조립도지표시복무적어의신식，병취류어의상사적복무。재차기출상，설계기우조개념생성향량적복무발포/발현산법，종복무/청구중제취다개산렬적관건자，재괄당증가복무발포/발현용여적정황하，실현결구화 P2P 망락중어의 Web 복무적모호필배。실험결과현시，여정학필배산법상비，해산법복무발현적평균소회솔유15.2%제고도료37.1%。
The traditional approach of semantic Web service discovery is centralized which leads to single point failure and performance bottleneck, quickly becomes impractical. Aiming at this problem, this paper presents a semantic Web service discovery algorithm in structured P2P networks. It partitions ontological concepts into different concept groups and defines group concepts to character service coarsely. Then it presents a service publishing/discovery algorithm based on group concepts generating vectors. The algorithm gets multi-keywords from service/query and realizes semantic Web service’s fuzzy match with appropriate redundancy in structured P2P networks. Experimental results show that the recall of this algorithm is increased from 15.2% to 37.1% compared with exact discovery algorithm.