A random walk based clustering with local recomputations for mobile ad hoc networks

In this paper, we present a clustering algorithm for MANETs. A cluster is divided in two parts: core nodes, that are recruited by a randomly moving token, and ordinary nodes, that are adjacent to core nodes. The clustering built by this algorithm is adaptive to topological changes. When a topological change occurs, nodes that are not in the cluster in which it occurs, or in adjacent clusters, are not affected. This property allows the scalability of this algorithm, by ensuring that only a bounded number of nodes have to recompute their cluster.