Hybrid swarm intelligence-based clustering algorithm for energy management in wireless sensor networks

Swarm intelligence-based algorithms are soft computing techniques, which have already been applied to solve a broad range of optimization problems. Generally, clustering is the most common technique, which, balances the energy consumption among all sensor nodes and minimizes traffic and overhead during data transmission phases of Wireless Sensor Networks. The performance scope of the existing clustering protocols is fixed and hence, cannot adapt to all possible areas of applications. In this paper, a multi-objective swarm intelligence algorithm – which is based on Shuffled Frog-leaping and Firefly Algorithms (SFFA) – is presented as a clustering-based protocol for WSNs. The multi-objective fitness function of SFFA considers different criteria such as cluster heads’ distances from the sink, residual energy of nodes, inter- and intra-cluster distances and finally overlap and load of clusters to select the most proper cluster heads at each round. The parameters of SFFA in clustering phase can be adapted and tuned to achieve the best performance based on the network requirements. The simulation outcomes demonstrated an average lifetime improvement of up to 49.1%, 38.3%, 7.1%, and 11.3% compared to LEACH, ERA, SIF, and FSFLA in different network scenarios, respectively.