中文摘要：

基于节点间的相关性创建网格社区，实现社区资源检索是近年来受到较多关注的网格资源发现方案之一．为了将扁平社区模型扩展为层次结构，并支持资源的多关键字组合检索，提出了基于频繁关键字组合的网格层次社区模型，采用跳表结构组织社区覆盖网络，设计了相应的资源检索方法，可以将查询消息快速路由给目标社区，并支持查询消息向下层社区的受控扩散．仿真实验表明，查询消息到达目标社区的跳数随网格规模的扩大而缓慢增长，使资源发现系统具有良好的伸缩性．

英文摘要：

Organizing grid nodes into communities based on the correlations between them and carrying out resource searching within target communities has become an increasingly promising resource discovery mechanism. In order to expand the flat community model into a hierarchical structure and support resource discovery based on keyword combinations, a grid hierarchical communities model based on frequent keyword combinations was developed and a resource discovery method based on skip list overlap was designed. It can route query messages to target communiies rapidly and diffuse messages to lower communities in a controllable way, so as to obtain a more satisfying result. Simulation results showed that the hops of a query message entering into a target community increased slowly with increases of node numbers, ensuring that the resource discovery system is scalable.