基于概念松弛的高效Web服务查询方法  

Efficient Web Service Query Approach Based on Concept Relaxation

作　　者：欧伟杰[1] 曾承[1,2] 项小明[3] 彭智勇[3] 李德毅[1,4] 

机构地区：[1]武汉大学软件工程国家重点实验室,武汉430072 [2]清华大学软件学院,北京100084 [3]武汉大学计算机学院,武汉430072 [4]中国电子工程系统研究所,北京100141

出　　处：《计算机学报》2011年第12期2381-2390,共10页Chinese Journal of Computers

基　　金：国家"九七三"重点基础研究发展规划项目基金(2007CB310806)资助~~

摘　　要：随着云计算技术的发展,面向服务的应用在互联网上呈现快速增长趋势,开放平台中基于云服务的组合服务也如雨后春笋般大量涌现,这给用户快速、精确定位所需服务带来了巨大挑战.尽管传统服务查询方法在查全率和查准率方面已取得较大进步,但仍无法适用于动态的互联网环境下大规模服务发现的要求.文章根据概念之间的语义关系,提出了基于概念松弛的相似性服务查询方法,它通过计算无关概念与服务对查询结果的影响,不仅改善了服务查询的效果,而且满足海量服务查询的高效性要求.经实验证明,文中提出的方法不仅在性能上优于传统方法,且满足服务查询的可扩展性.此外,该方法已经应用于上线的按需服务平台中.Cloud computing and service-oriented applications on the Internet are growing rapidly.On the other hand,open platform speed up the emergence of composited services,which is based on Cloud service.How to discover the desired services for users efficiently has become a significant challenge.Although traditional approaches have made progress in recall rate and precision.But they are not suitable for large-scale service discovery under dynamic environment.In this article,a novel approach for service query based on concept relaxation is proposed,which employs the semantic relation of concepts from hierarchical ontology.The unrelated concepts and services are figured out to improve the efficiency of algorithm.The method has been implemented in a prototype of on-demand service platform.The results of experiments illustrate that the proposed approach not only outperform the traditional ones,but also satisfy the scalability of service discovery.

关 键 词：服务计算 服务发现 语义相似性 二分图匹配 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]