Energy Efficient Load Balanced Clustering Algorithm Based on Learning Automata for Wireless Sensor Networks

Abstract

Aiming at the random distribution of nodes in wireless sensor networks (WSNs), and based on ICLA protocol adopting the learning automata (LA), an energy balanced unequal clustering algorithm with considering the node density is proposed and evaluated in this paper. The approach considers the residual energy and the node density in cluster head election and adopts LA for information exchange with the surrounding environment, so it can choose relatively better cluster heads. Meanwhile, according to the distance between cluster heads and the base station and node density, a series of unequal clusters are formed to balance the energy load of intra- and inter-clusters in different positions and node density degrees of networks. The approach also adopts an evaluation function to choose optimal relay cluster heads and form multi-hop routing, which achieves a tradeoff between the energy of cluster heads, node density in cluster and distances from cluster heads to the base station. Therefore, it can achieve the goal of optimizing cluster heads selection and balancing energy load among all sensor nodes in the network. Simulation results show that the protocol can choose relatively more reasonable cluster heads, efficiently balance the energy load among nodes and significantly prolong the network lifetime.