Ad Hoc Network Clustering Algorithm Research

In recent years Mobile Ad Hoc Wireless Network attracted a lot of attention from researchers. As a complimentarily for traditional fixed network facilities, Ad Hoc wireless network play an irreplaceable role in some applications. When the Ad Hoc Network size increase, a large number of control information for maintaining dynamic routing is needed and sometimes the costs increase exponentially. This phenomenon can induce consumption of the entire network bandwidth. To solve this problem, the large Ad Hoc wireless network need to be clustered effectively.Current network clustering algorithms research for Ad Hoc network mainly concentrated on improving the adaptability and effectiveness of sub-clusters. In order to improve the double index, it is important to estimate the network topology changes effectively.This paper presents a point which using the similarity of different time node neighbor's matrix to judge the changes of network. In classic clustering ,algorithms adaptive research are focuses on optimizing or integrating the various network elements to void that existing networks are limited by some network elements. Classical clustering algorithms generally do not consider the specific network topology on the impact of clustering. In order to make better use of special network topology, this paper presents a new clustering algorithms(BMOC, Block Mobility On-demand Clustering Algorithm). The complexity of BMOC algorithm is analyzed and simulation modeling is established. From simulation BMOC algorithm has certain self-adaptability in network topology compared with similar complexity of classic algorithms, and under some certain conditions it can significantly improve the effectiveness of network clustering. At the same time BMOC algorithm and the traditional algorithms have no conflict in adaptive, so we can combination of the merits of two algorithms in our next research.