Clustering and Routing in Mobile Wireless Networks

In this paper we propose several new clustering algorithms for nodes in a mobile ad hoc network. We propose to combine two known approaches into a single clustering algorithm which considers connectivity as a primary and lower ID as secondary criteria for selecting clusterheads. Several inherently collision-free clustering algorithms are then designed based on a depth first search (DFS) traversal of nodes in the network. These algorithms are initialized by any node and are fully distributed. They create and maintain k-clusters, in which any node is at distance at most k hops from the clusterhead. In the clustering process, each node is either a clusterhead, a covered or undecided node. In the basic k-cluster algorithm, each undecided visited node in DFS declares itself a clusterhead and covers all its k-hop neighbors. In the highest connectivity k-cluster algorithms, each undecided visited node checks all its undecided k-hop neighbors and chooses one with the largest connectivity to be its clusterhead. We also describe algorithms for modifying cluster structure in the presence of topological changes. The efficiency of the clustering algorithms is tested by measuring the average number of created clusters, the number of border nodes, and the cluster size in random unit graphs. In a recently proposed GEDIR (geographic distance routing) algorithm, each node forwards the packet to one of its neighbors that is closest to the destination. We propose A-GEDIR (Avoidance GEDIR) algorithm for clustered networks, in which node currently holding the packet selects the second closest neighbor to destination if the best choice is a clusterhead (and if second choice leads closer to destination). The proposed variation addresses the problem of higher energy consumption of clusterhead nodes. The delivery rates, hop counts and transmission ratios for clusterheads and other nodes in GEDIR and AGEDIR algorithms are then compared.