Title:
ENERGY-EFFICIENT CLOUD COMPUTING SOLUTIONS FOR HEALTHCARE SYSTEMS USING GENETIC ALGORITHMS

Authors:
Vijaykumar Mamidala and Aiswarya RS, India

Abstract:
Cloud resource management in healthcare systems is essential for optimizing computational efficiency and minimizing energy consumption while handling large-scale data. Despite existing efforts, challenges such as poor task-to-machine assignment and lack of dynamic resource allocation continue to impact system performance and energy efficiency. This work proposes an energy-aware scheduling solution using Genetic Algorithms (GA) to address these challenges, focusing on optimizing task assignments and dynamically allocating resources based on workload changes. The workflow begins with data collection from healthcare sources like Electronic Health Records and, followed by preprocessing using mode imputation for missing data and Min-Max scaling for standardization. The system model is constructed using a system graph for machine representation and a program graph for task assignment. The Genetic Algorithms optimizes the task-to-machine assignments by evaluating energy consumption, applying selection, crossover, and mutation techniques through multiple generations. Finally, energy consumption is minimized, and task completion time is optimized by adjusting machine power states during active and idle phases. The results demonstrate a significant reduction in energy consumption, from 240 kWh to 140 kWh as tasks increased from 50 to 100, while throughput improved from 90 tasks/time to 125 tasks/time over 15 generations. The proposed method improves the energy efficiency, task completion time, and scalability in cloud-based healthcare systems, offering a sustainable approach to resource management.