Performance Study of Bio-Inspired Approach to Clustering in Wireless Sensor Networks

Wireless sensor networks experience an urgent need to deliver maximum efficiency in order to sustain their massive deployment. Clustering has been regarded as an effective technique to enhance efficiency of sensor nodes by suitably distributing them over the network region. Bio-inspired algorithms, with their ability to optimize complex problems with no well-defined solution, have been considered as perfect tool to aid this form of centralized clustering. In this paper we perform a study on the recently developed biologically inspired algorithms namely, firefly algorithm, symbiotic organism search algorithm and the bat algorithm and analyze the same against the meta-heuristic harmony search algorithm and the deterministic k-means algorithm for clustering the sensor nodes. Here, the intra-cluster communication cost has been considered as the optimization function orchestrating the clustering process. It is found that firefly algorithm displays a superior performance in terms of optimization owing to its unique ability to explore and exploit the solution set appropriately.