Intelligent Clustering-Based Virtual Machine Load Balancing in Cloud Computing Using K-Means Clustering with Particle Swarm Optimization

Efficient Virtual Machine (VM) load balancing remains a fundamental challenge in modern cloud computing environments characterized by heterogeneous workloads and dynamic resource constraints. This research presents a novel intelligent clustering-based approach—K-Means Clustering with Particle Swarm Optimization (KMC-PSO)—that combines data clustering techniques with swarm intelligence to achieve superior VM distribution and resource allocation. Through comprehensive empirical evaluation and comparative analysis, KMC-PSO demonstrates significant improvements over state-of-the-art algorithms including Hybrid Grasshopper Optimization Algorithm-Genetic Algorithm (HGOA-GA), Ant Colony Optimization (ACO), Genetic Algorithm (GA), Markov Decision Processes (MDP), and traditional Round Robin scheduling. The proposed KMC-PSO approach achieves 12.61% improvement over HGOA-GA and 18.96% reduction in total completion time compared to Round Robin baseline. Additionally, KMC-PSO reduces operational costs by 15.89%, minimizes average waiting time by 13.62%, improves makespan by 16.86%, reduces energy consumption by 15.57%, and enhances resource utilization to 84.3%. This research demonstrates that intelligent clustering-based approaches combined with swarm intelligence optimization provide superior solutions for multi-objective cloud resource management, with significant implications for cost reduction, energy efficiency, and service quality in cloud data centers.

Keywords: Virtual Machine Load Balancing, K-Means Clustering, Particle Swarm Optimization, Swarm Intelligence, Cloud Computing, Resource Optimization, Energy Efficiency, Multi-Objective Optimization, Clustering-Based Scheduling