Enhanced multi-level K-means clustering and cluster head selection using a modernized pufferfish optimization algorithm for lifetime maximization in wireless sensor networks

• Optimize the performance of Wireless sensor networks using a novel meta-heuristic optimization algorithm; the modernized pufferfish optimization algorithm • Optimization algorithm mimics the unique defensive mechanisms of pufferfish in nature • Enhanced multi-level K-means algorithm; cluster the nodes dynamically in response to evolving network conditions • Proposed optimization algorithm used for selecting the most suitable Cluster Head by introducing a new random parameter to solve premature convergence and to obtain the global optimum solution • Comparative analysis with other models shows that the proposed technique performs better Strategic sensor node deployment is crucial for optimizing the performance of Wireless Sensor Networks (WSN). This optimization enhances the network coverage, data acquisition efficiency, and energy conservation, ultimately prolonging the network lifetime and improving overall system reliability. Therefore, researchers are paying more attention to improving the network lifetime. However, existing methods often fall short of achieving optimal network coverage due to inadequate sensor deployment. Additionally, the absence of clustering in certain models compromises the network capacity. To tackle these challenges, a new multi-level clustering technique with a heuristic optimization algorithm is proposed in this research work. A novel heuristic, the Modernized Pufferfish Optimization Algorithm (MPOA), is introduced to optimize WSN performance, drawing inspiration from the pufferfish's natural defense strategies. Initially, an enhanced multi-level K-means algorithm (MKA) is employed to cluster the nodes dynamically in response to evolving network conditions. Finally, an MPOA is suggested for selecting the most suitable Cluster Head (CH) by introducing a new random parameter to solve premature convergence and to obtain the global optimum solution. The test results show that the proposed strategy significantly improves the CH selection, increases the network lifetime by 12.5% to 62.5%, decreases the energy consumption by 16.5% to 28.7%, decreases the path loss by 16% to 46%, increases the number of alive nodes by 32.5% to 55%, increases the residual energy by 3.3% to 77.7%, decreases the cost function by 1.76% to 4.42%, and increases the normalized energy, compared to several benchmark techniques based on varying number of nodes and rounds.