<![CDATA[The rapid expansion of cloud computing has increased the complexity of task scheduling and resource management across heterogeneous and dynamic environments. Conventional heuristic methods often suffer from premature convergence, resulting in imbalanced virtual machine (VM) utilization. To address these challenges, this study proposes ECO-FISH, a hybrid Opposition-Based Artificial Fish Swarm Algorithm (AFSA) designed for efficient cloud task scheduling. AFSA is selected for its swarm intelligence behaviors—prey, follow, and swarm—which enable effective local exploration with relatively low computational cost. To enhance global exploration, Opposition-Based Learning (OBL) is incorporated by evaluating opposite task–VM mappings, allowing the algorithm to escape local optima and maintain population diversity. This synergy improves the balance between exploration and exploitation while retaining algorithmic simplicity. The proposed ECO-FISH algorithm is implemented using CloudSim and benchmarked against GA, PSO, and the baseline AFSA using three workload distributions: uniform, normal, and stratified. Experimental results demonstrate that AFSA alone reduces makespan by 28–45%, increases throughput by 34–84.9%, and improves utilization by 44.12–64.59% compared to GA. The OBL enhancement in ECO-FISH provides additional gains of up to 1.6%, showing the most significant improvement under heterogeneous, stratified workloads with high variance. Overall, AFSA performs well on uniform datasets, while ECO-FISH (AFSA with OBL) exhibits superior adaptability and stability in variable cloud environments.]]>
Copyrights © 2025
