Double Star Embedded Clustering Algorithm for Wireless Ad Hoc Networks Using Ranking

Abstract

Clustering of mobile ad hoc network (MANET) is an important concept, which partitions node set into many classes such that each class has some structural property. In literature, there are many papers available, which are due to star topology clusters. In this paper, a new concept of k-hop double star embedded clustering is introduced, in which each cluster has a master, a proxy and two types of slaves. The slaves in the cluster are adjacent to either the master or the proxy or both and themselves. Clearly, this type of partition of the network will induce less number of clusters than the star clustering. Extending the concept of ranking of nodes done by authors in T.N. Janakiraman and J. Janet Lourds Rani, An efficient k-hop weighted domination clustering algorithm for mobile ad hoc networks using ranking, IEEE Proceedings of the International Conference on Computational Intelligence and Multimedia Applications 2007 (ICCIMA ’07) based on the signal strength, in this paper, the edges are also ranked based on reliability of having more number of common nodes and four-con-cyclic neighbours to the end vertices of the edges. Comparing to star clusters, it is clear that this procedure produces lesser number of clusters and also with increasing reliability. Clearly, the double star embedded cluster of the given network must have a sub graph, which is a double star having a master and a proxy as a central pair of nodes.