基于兴趣聚类的网格资源发现算法*

为了更有效地解决网格资源的搜索和定位问题,提出一种以P2P形式实现的、基于兴趣聚类的非集中式网格资源发现算法.算法采用被动学习方式,通过用户的访问历史抽取节点的兴趣属性,将节点按照兴趣属性划分为多个簇,资源发现请求在簇内朋友节点之间传播,查找失败后,将请求路由到与其兴趣最相似的其他簇内.仿真测试表明,算法稳定高效,相比传统算法在低开销情况下性能有显著的提高.

Interest clustering based on grid resource discovery algorithm

This paper proposed a grid resource discovery algorithm: peer to peer and decentralized interest clustering based algorithm to address the search and location of issues.The algorithm learned passively interest attributes between nodes from history search results, divided nodes into interest clusters. Search request was propagated between nodes with similar interest within interest clusters. Simulation results show that, compared to the traditional algorithm,this algorithm improves query efficiency notably without a significant increase in load.