Multi-level clustering protocol for load-balanced and scalable clustering in large-scale wireless sensor networks

Abstract

The advent of wireless sensor networks (WSNs) has revolutionized the field of smart applications. In order to improve the performance of WSNs, refinement of clustering and routing protocols can make a vast difference. Existing classical and evolutionary optimization technique-based protocols have high computational complexity since clustering and routing problems are solved separately. Moreover, these protocols suffer from hot-spot problem due to uneven formation of clusters. In this paper, we propose a multi-level clustering protocol (MLCP) for energy-efficient data gathering in large-scale WSNs. Additionally, a hierarchical clustering architecture is designed in MLCP to jointly solve the problems of clustering and routing. Further, for the purpose of cluster head selection, a hybrid dragonfly algorithm-based particle swarm optimization technique is proposed which combines the exploration and exploitation capabilities of dragonfly algorithm and particle swarm optimization, respectively. MLCP considers intra-cluster distance, node degree and inter-cluster distance for the formation of scalable, load-balanced and energy-efficient clusters. To demonstrate the full potential of MLCP, network simulations have been carried out in diverse network conditions. MLCP has shown up to 90% increase in the network lifetime and an improvement of 19.36% in conservation of energy in comparison with the competent protocols. The comparison of obtained results with state-of-the-art clustering protocols clearly establishes the superiority of MLCP in achieving load-balanced, scalable and energy-efficient clustering.