基于用户兴趣的P2P资源搜索算法设计与分析  

Search Algorithm Based on User Interest-proximity Measurement

作　　者：何颂颂[1,2] 顾乃杰[1,2] 田舟贤[1,2] 谢静[1,2] 

机构地区：[1]中国科学技术大学计算机科学与技术系,安徽合肥230027 [2]安徽省计算与通信软件重点实验室,安徽合肥230027

出　　处：《小型微型计算机系统》2008年第11期2027-2030,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金重点支持项目(60533020)资助;博士点基金部分支持项目(20060358011)资助

摘　　要：P2P搜索效率跟P2P用户的特性相关,本文将基于P2P用户的群聚特性,在纯P2P网络之上,构建基于用户兴趣的覆盖网络,并设计相应的资源搜索策略.本文通过计算节点间的相似度来构建覆盖网络,根据查询请求和节点的相关度来决定搜索策略并从理论上证明了本文提出的搜索算法的性能优于随机漫步和简单洪泛.实验结果显示,本文提出的搜索算法的平均搜索时间和消息复杂度均优于随机漫步搜索算法.Efficiency of P2P search has strong relation with the habit of users. Based on the clustering of users in P2P overlay network, this paper proposed a new network model and corresponding search algorithm. This paper computed the proximity to form out overlay network, and adjust this search algorithm according to the proximity. This paper also proved its search algo- rithm had better efficiency than random walk search algorithm and simple flooding algorithm theoretically. Experiment result shows that, its search algorithm have better performance than simple random walk search algorithm in the aspect of average search time as well as message complexity.

关 键 词：P2P 兴趣网 资源搜索 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]