Performance Enhancement of Cloud Resource Management by Using Optimized Reinforcement Learning Approaches

Abstract

Cloud computing provides on-demand access to a shared pool of specially configured computing resources. Scheduling tasks in these dynamic environments is challenging due to fluctuating workload demands, varying resource availability, and differing task priorities. Traditional optimization algorithms often fail to adapt effectively to these conditions. In this paper, we introduce the Q-Whale algorithm (QWA) and SARSA-Whale algorithm (SWA), a novel hybrid approach that integrates the Whale Optimization Algorithm (WOA) with Reinforcement learning techniques to address these challenges. The Q-Whale algorithm combines WOA's exploration capabilities with Q-learning's adaptive decision-making to optimize task scheduling in real time. Our experiments in dynamic computing environments show that the hybrid algorithm enhances resource utilization, reduces makespan, and meets task deadlines more effectively than traditional methods.