WACHM: Weight Based Adaptive Clustering for Large Scale Heterogeneous MANET

Yi Wang,Hairong Chen,Xinyu Yang,Deyun Zhang

DOI: https://doi.org/10.1109/iscit.2007.4392150

2007-01-01

Abstract:Node clustering is an effective technique to mitigate the topology changes in mobile ad hoc networks (MANET) because nodes are dynamic in nature due to the mobility of nodes. Many clustering algorithms, in which MANET is usually assumed to be homogeneous, have been proposed. However, a homogeneous MANET suffers from poor scalability. Recent research has demonstrated its performance bottleneck both theoretically and through simulation experiments and testbed measurement. In this paper, a design methodology is presented, named weight based adaptive clustering for large scale heterogeneous MANET (WACHM), in order to build a large scale hierarchical MANET using heterogeneous nodes to construct different layers. Four critical issues are addressed. First the optimal number of CHs is analyzed in theory. Then the optimal hops via simulations are obtained. Finally the dependency probability and the link stability are defined. In cluster head election, we take into account the degree-difference, battery power, average link stability and dependency probability for electing cluster heads from advanced nodes. At last, four scenarios of cluster maintenance are proposed and solved. Simulation results show that WACHM achieve good scalability and demonstrate the superior performance of WACHM.