| Abstract: |
Cloud computing has emerged as the backbone of modern digital infrastructure, providing scalable and flexible computing resources to users worldwide. However, efficient resource allocation remains a key challenge due to fluctuating demand, heterogeneous workloads, and the dynamic nature of cloud environments. Traditional resource allocation strategies often suffer from inefficiencies, leading to resource wastage, increased operational costs, and performance degradation. This paper proposes a hybrid machine learning approach that integrates supervised learning for demand prediction and reinforcement learning for dynamic allocation to optimize resource distribution in cloud environments. The proposed model aims to enhance scalability, adaptability, and cost-efficiency while ensuring Quality of Service (QoS) compliance. Through extensive simulations and real-world cloud workload datasets, we demonstrate that the hybrid model significantly improves resource utilization, task scheduling efficiency, and energy consumption compared to traditional methods. |
