自组网络的分簇算法

Clustering Algorithms in Ad Hoc Networks

当自组网络(ad hoc)的规模增大时,控制分组在网络中的传输急剧增多,这大大降低了路由和转发的性能.分簇是提高ad hoc网络可扩展性的重要手段.针对已有分簇算法的缺陷以及静态和动态ad hoc网络的特性,分别提出了新的静态分簇算法SACA和动态分簇算法DACA.其中SACA算法将分簇数量减小到已有算法的大约一半,但不增加簇成员数量,提高了分簇效率,减少了分簇时间;DACA算法将簇间的公共节点数目大大降低,从而有效降低了分簇结构的维护开销,提高了网络运行的效率、路由和转发性能.

As the scale of ad hoc network enlarges, control packets will sharply increase, thus routing/forwarding performance will be greatly decreased. Clustering is an efficient way to raise the extensibility of ad hoc networks. Two clustering algorithms, SACA for static ad hoc networks and DACA for dynamic networks, were proposed. SACA reduced the number of clusters by half without increasing the number of cluster members, so clustering efficiency was improved and clustering time was saved. DACA evidently decreased the number of common neighbor clusters, thus effectively lowering maintenance cost and improving routing and transferring performance.