AI-DRIVEN CLOUD RESOURCE OPTIMIZATION FOR SMART CITIES

Abstract

The rapid expansion of smart city ecosystems has significantly increased the demand for scalable and intelligent cloud computing infrastructures capable of processing massive volumes of heterogeneous data generated by Internet of Things (IoT) devices, intelligent transportation systems, healthcare applications, smart grids, environmental monitoring networks, and public utility services. Conventional cloud resource management techniques primarily rely on static provisioning, threshold-based scheduling, and reactive resource allocation strategies, which are often inadequate for handling highly dynamic and unpredictable workloads. These limitations result in inefficient resource utilization, increased latency, excessive energy consumption, higher operational costs, and degraded Quality of Service (QoS). To overcome these challenges, this research proposes an AI-Driven Cloud Resource Optimization Framework for Smart Cities that integrates Machine Learning (ML), Reinforcement Learning (RL), predictive analytics, cloud-edge collaboration, and real-time monitoring into a unified intelligent cloud management platform. The proposed framework continuously collects infrastructure metrics such as CPU utilization, memory consumption, storage usage, network bandwidth, response time, and workload intensity through an intelligent cloud monitoring layer. A Random Forest Regression model is employed to forecast future workload demands using historical and real-time infrastructure data, enabling proactive resource provisioning. Furthermore, a Proximal Policy Optimization (PPO)-based Reinforcement Learning scheduler dynamically determines optimal resource allocation policies by continuously learning from environmental feedback. Cloud-edge optimization mechanisms further reduce latency by intelligently distributing delay-sensitive workloads between centralized cloud servers and edge computing nodes. The framework also incorporates an interactive dashboard for real-time visualization of resource utilization, workload prediction, optimization decisions, and infrastructure performance. Experimental evaluation demonstrates that the proposed AI-driven framework significantly improves resource utilization, reduces response time, minimizes operational costs, lowers energy consumption, and enhances overall system scalability compared with conventional cloud resource management approaches. The proposed architecture provides a robust, adaptive, and intelligent solution capable of supporting next-generation smart city infrastructures while ensuring efficient cloud resource utilization, autonomous decision-making, and sustainable urban digital transformation.