Abstract: Combined with genetic clustering and K-means algorithm, a mixed node clustering algorithm is presented, which clusters the distance between nodes for unstructured P2P network. It optimizes the overlay network, and improves the routing efficiency. Experimental results show this algorithm has good reliability and expansibility. It can have an effective nodes clustering for unstructured P2P network. After clustering, the average latencies of nodes obviously decrease.

Key words: unstructured P2P network, network coordinate, clustering algorithm

摘要： 结合遗传聚类算法及K均值算法，提出一种混合的节点聚类算法，在无结构P2P网络中根据节点之间的距离进行聚类，可以优化其覆盖网络，提高系统的路由效率。实验结果表明，该算法具有良好的可靠性及可扩展性，能对节点进行有效聚类，且聚类后的节点平均延迟明显降低。

关键词: 无结构P2P网络, 网络坐标, 聚类算法

CLC Number: 

• TP311.52