Abstract: In order to effectively locate resourse and search information in P2P community，a hybrid P2P topology is adopted and the document searching in P2P community is divided into three processes：local retrieval，intra-search and inter-search.In local retrieval，the new term weight calculation method is designed to solve the problem about txt search in homogeneous collection.In intra-search and inter-search，a part of peers which are more relevant to the query to conduct the retrieval task is selected.Experimental results show that this method can get better retrieval results with lower cost.

Key words: peer-to-peer community, information retrieval, algorithm design

摘要： 为解决P2P社区的资源定位及信息检索问题，采用混合型P2P网络模型，将社区内的检索划分为本地检索、组内搜索和组间搜索。对于本地检索设计了新的词条权重的计算方法，解决了同构文档集内的文本检索问题。对于组内搜索和组间搜索，通过设计节点选择策略，使一部分与查询相关度高的节点执行查询任务。最后提出结果融合的方法并对特定的实验数据进行测试，实验表明设计的算法在较小的查询开销下，能取得较好的检索效果。

关键词: P2P社区, 信息检索, 算法设计