CayDHT 是一种基于Cayley 图的常数度结构化对等覆盖网络，对于精确的单关键字搜索效率非常高，但不支持不限定搜索形式的复杂搜索．通过分析CayDHT 的拓扑性质，提出了一种基于虚拟搜索树的复杂搜索算法VTCS．该算法无需维护额外的树结构，根据消息参数就可以获得下一跳节点的地址．理论分析表明该算法可以在O( logN) 的时间复杂度内完成无冗余消息的搜索．仿真实验比较了Flooding、RW 和VTCS，结果表明VTCS是较为合理的方案．

CayDHT is a constant-degree structured peer-to-peer overlay network based on Cayley graph,and it is very efficient in accurate single keyword search but does not support complex searches with unrestricted form. In this paper,by analyzing the topological properties of CayDHT,a virtual tree-based complex search ( VTCS) algorithm is proposed,which needs no maintenance of the extra tree structure and can obtain the address of the next hop only according to the parameters inside the messages. Theoretical analyses indicate that the proposed algorithm can complete the search without redundant messages within O( logN) hops. Simulation results show that VTCS is more reasonable than Flooding and RW.