Intelligent Cloud Resource Optimization System Using Machine Learning

Abstract

Cloud computing has significantly transformed the delivery of computing resources by providing scalability, flexibility, and cost efficiency. However, dynamic workloads and unpredictable user demands continue to pose challenges in efficient resource allocation. Traditional methods often result in either underutilization or over-provisioning of resources. In this paper, an enhanced intelligent cloud resource optimization system is proposed using Machine Learning techniques. Initially, multiple models including Linear Regression, Decision Tree, and Random Forest were trained and evaluated, where Random Forest achieved the highest prediction accuracy and was selected for further implementation. Building upon this, the system is extended to a real-time environment using Streamlit, where live system parameters such as CPU usage, memory utilization, storage, and workload are continuously monitored. The trained model predicts CPU utilization dynamically, and based on the prediction, an automated resource scaling mechanism is implemented to allocate virtual machines efficiently. The proposed system not only improves prediction accuracy but also optimizes resource utilization and reduces operational costs through intelligent decision-making. Experimental observations demonstrate that the system provides better adaptability and efficiency compared to traditional static allocation approaches.