ROA: Energy and Environment Aware Task Scheduling in Cloud Computing
S Reddy, Morigadi Vishwashanthi, T. Sreenivasulu Reddy, Lavanya Reddy Mendu, Uma Devi Manchala, Santhosh Kumar Medishetti
Abstract
Efficient task scheduling is a critical concern in cloud computing to ensure optimal resource utilization and sustainable operation. This paper introduces a novel Revolution Optimization Algorithm (ROA) for task scheduling in cloud environments, designed to minimize energy consumption, temperature variation, and carbon emissions. Inspired by socio-political revolutions, ROA incorporates a dynamic strategy that balances exploration and exploitation during the task-to-resource mapping process. The proposed approach is validated using the CloudSim simulator under realistic workload conditions based on the CEA-Curie workload, which reflects high-performance computing scenarios. Performance comparisons were conducted against traditional algorithms such as GA, PSO, and ACO. Experimental results demonstrate that ROA significantly reduces average energy consumption and thermal fluctuations while lowering the environmental impact measured in terms of carbon emissions. The adaptive learning mechanism within ROA enables rapid convergence to optimal solutions, even under fluctuating workloads and heterogeneous cloud resources. Compared to baseline models, ROA achieves an average improvement of 18.7% in energy efficiency, 16.3% reduction in temperature variation, and 20.5% decrease in carbon emissions. These findings suggest that ROA is a promising, eco-efficient solution for sustainable task scheduling in modern cloud computing infrastructures.