IMPROVED K-MEANS AND ENHANCED PARTICLE SWARM OPTIMIZATION FOR ROBUST, RESILIENT, AND ENERGY-EFFICIENT CLUSTERING IN WIRELESS SENSOR NETWORKS

Abstract

 Wireless Sensor Networks(WSNs) are useful for many applications and  their operational life is limited by energy consumption. Clustering is a powerful technique to extend network longevity by reducing communication overhead. A key challenge in WSN clustering is determining the optimal No.of Clusters(K). This paper introduces an Improved K-means and Enhanced Particle Swarm Optimization (PSO) algorithm, designed for robust, resilient, and energy-efficient clustering in WSNs. The study first evaluates Elbow and Silhouette methods for optimal K selection across various network sizes, highlighting their strengths. Building on this, the proposed Improved K-means- Enhanced PSO leverages PSO's optimization capabilities to refine cluster head selection and formation, addressing traditional K-means' sensitivity to initial centroid placement. Extensive simulations validate this enhanced approach against conventional LEACH and standard PSO-based LEACH protocols. Results visualized that Improved K-means- Enhanced PSO significantly surpasses existing methods in energy utilization for improvement of Network Lifetime(NL), and data packet delivery. Specifically, it will leads to fewer dead nodes and higher packet transmission rates to cluster heads. And the sink, substantially boosting the network's overall Energy Efficiency(EE)and operational resilience. This research confirms that strategically integrating an enhanced PSO algorithm with K-means clustering. Which it significantly provides a superior solution for optimizing Wireless Sensor Network(WSN) performance, ensuring robust and adaptive energy management for sustainable sensor networks.