Authors: I.C Emeto, B.P Gbaranwi, A.A. Galadima, A.C Okoloegbo, S. Kwaghbee, E.C Ochuba
Abstract: Cloud computing has emerged as a dominant paradigm for delivering scalable, on-demand computing resources, yet efficient load balancing remains a critical challenge in modern data centers. This paper presents a novel Hybrid Bee Ant Colony (HBAC) Algorithm that synergistically combines Ant Colony Optimization (ACO) and Artificial Bee Colony (ABC) metaheuristics to address the inherent limitations of existing load-balancing approaches. The proposed HBAC algorithm leverages ABC's robust exploration capabilities to identify underutilized virtual machines (VMs) and ACO's pheromone-driven exploitation mechanism to optimize task allocation, thereby achieving superior performance in dynamic cloud environments. Through extensive simulations using CloudSim with Google Cluster Data traces, we demonstrate that HBAC significantly outperforms standalone ACO and ABC algorithms across key performance metrics. Experimental results show 15.7% reduction in makespan, 22.3% improvement in response time, and 18.9% better resource utilization compared to conventional approaches. The hybrid model particularly excels in maintaining balanced VM workloads (degree of imbalance reduced by 27.4%) while demonstrating exceptional scalability under varying workload conditions (from 1,000 to 10,000 tasks). The algorithm's innovative two-phase architecture – where ABC scouts first identify high-potential VMs and ACO ants then optimize task placement – effectively overcomes the slow convergence of pure ACO and the excessive exploration of pure ABC. Energy efficiency analysis reveals 13.2% reduction in power consumption, making HBAC particularly suitable for sustainable cloud operations.
DOI: http://doi.org/10.5281/zenodo.17264412
