A hybrid grey wolf and improved particle swarm optimization technique with Pareto knee selection for energy efficient load balancing in cloud computing

In the realm of Cloud Computing, multi-objective task scheduling aims to map tasks to virtual machines while minimizing execution time, balancing load, and reducing cost. Achieving this requires intelligent algorithms that optimize competing objectives for efficient resource use and user satisfaction. While various optimization methods have been explored over the past decade, the growing demand for cloud-based applications and energy-efficient computing calls for more adaptive strategies. This study proposes a Hybrid Grey Wolf and Improved Particle Swarm Optimization with Pareto-Knee Selection (HGWIPSO-PKS) Load Balancing scheme for energy-efficient, performance-optimized resource allocation in cloud environments. The method combines GWO’s strong global exploration with PSO’s rapid convergence and focuses on Pareto “knee” solutions to effectively balance trade-offs. It dynamically adapts to workload fluctuations, optimizing energy use and performance. Extensive CloudSim experiments across diverse virtual machine configurations and task volumes show that HGWIPSO-PKS outperforms standalone PSO, GWO, and other state-of-the-art hybrids, reducing makespan by 14.96%, imbalance by 20.0%, energy consumption by 10.88%, and execution time by 10.87%. Response, waiting, and turnaround times drop to about 7%, while throughput and resource utilization improve by 19.36% and 17.32%, respectively. Moreover, HGWIPSO-PKS achieves an average reduction of about 13% in computational cost across varying workloads, while maintaining scalability comparable to hybrid schedulers. These results highlight HGWIPSO-PKS as an effective, green-computing scheduler for dynamic cloud environments.