An ML-Assisted Multi-Objective Butterfly Optimization Framework for Adaptive Energy-Efficient Clustering in Wireless Sensor Networks

This paper presents a machine-learning based multi-objective butterfly optimization algorithm for dynamic cluster head selection in a wireless sensor network. This method integrates machine learning to predict Pareto-optimal solutions, thus reducing the computational time by reusing previously generated Pareto-optimal front. The multi-objective optimization integration allows to find out the best trade-off solutions of CHs. Simulation results demonstrate that the proposed framework significantly enhanced energy consumption, network lifetime, and throughput and reduced delay as compared to baseline algorithms LEACH and butterfly optimization algorithm. The proposed method achieves 40–50% higher throughput and prolonged residual energy retention. The lifetime is increased up to 5× as compared to baseline butterfly optimization algorithm and 6× compared to LEACH.