A Game-theory Based Clustering Approach for Wireless Sensor Networks

Clustering has long been regarded as an efficient technique for data collection in wireless sensor networks. The approach divides network into clusters and makes cluster heads (CHs) responsible for data aggregation. However, due to the constraints on available resources, a sensor node is likely to be selfish and refuse to declare itself as a cluster head. In this paper, we propose a game-theory based clustering approach for wireless sensor networks. A game-theoretic model is built for CH selection. This paper adopts data replication to reduce possible network disconnection. The selection of a candidate CH is discussed under a second price sealed auction. Simulation results show that the throughput of the sink can still be guaranteed if any CH fails to work.