摘要
结合事件/订阅本体模型,提出了一种高效、基于语义的事件/订阅匹配算法,称为多维索引匹配计数(MIC)算法.将事件和订阅表示成资源描述框架(RDF)图和RDF图模式.根据RDF订阅图模式的特点,采用多维哈希表和二叉排序树分别对RDF订阅图模式中弧和顶点对的概念类型约束,以及谓词条件约束建立多层索引,以加速订阅图模式中弧和顶点对的匹配.并利用订阅图模式之间的概念类型约束、谓词条件约束的覆盖关系减少重复匹配.实验结果表明,该算法的匹配效率优于已有的图模式匹配算法,适合大规模发布/订阅应用的需求.
An efficient semantic-based matching algorithm named multi-dimension index and counting （MIC） algorithm was proposed for publish/subscribe systems based on the ontology models of publication and subscription. Publication and subscription was represented as resource description framework （RDF） graph and RDF graph patterns. The algorithm respectively uses multi-dimension Hash tables and binary sorted tree to index the concept class constraints of arc and vertex pairs and the predicate condition con- straints of subscription graph patterns according to the characteristics of RDF graph patterns. Then the algorithm can speed up the matching of arc and vertex pairs of subscription graph patterns. The algorithm also exploits the covering relationship of predicate constraints and class concept constraints between subscriptions to eliminate unnecessary redundant matching complexity. Experimental results show that the MIC algorithm outperforms former algorithms and satisfies the requirement of large scale publish/subscribe applications in efficiency.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2009年第1期63-68,共6页
Journal of Zhejiang University：Engineering Science
基金
浙江省科技计划资助项目(2007C13090)
关键词
发布/订阅
匹配算法
语义WEB
资源描述框架
publish/subscribe
matching algorithm
semantic Web
resource description framework （RDF）
