Distributed graph clustering for application in wireless networks

Abstract

We consider distributed clustering of weighted graphs. Each node in the graph is represented by an agent, with agents independent of each other. The target is to maximize the sum weight of intra-cluster edges with cluster size constrained by an upper limit. To avoid getting stuck in not-too-good local optima, we approach this problem by allowing bad decision-making with a small probability that is dependent on the depth of local optima. We evaluate performance in a setting inspired by self-organizing coordination area formation for coordinated transmission in wireless networks. The results show that our distributed clustering algorithm cab perform better than a distributed greedy local search. © 2011 Springer-Verlag.