AI-Based Resource Allocation, Security, and Task Scheduling in Cloud Computing Using Hybrid Split-Attention Networks

Abstract

Cloud computing has revolutionized modern computing by enabling scalable, on-demand access to computational resources. However, efficient resource allocation, secure data processing, and optimal task scheduling remain critical challenges due to the dynamic and heterogeneous nature of cloud environments. Recent advancements in Artificial Intelligence (AI), particularly deep learning and hybrid optimization techniques, have introduced intelligent frameworks for addressing these challenges. This paper presents a comprehensive review of AI-driven techniques for joint resource allocation, security, and task scheduling, emphasizing hybrid pyramidal convolution and split-attention network architectures. The study explores recent developments, highlighting the integration of convolutional neural networks, reinforcement learning, and optimization algorithms. A systematic literature review is conducted to analyse performance improvements in terms of resource utilization, latency reduction, energy efficiency, and security enhancement. Furthermore, trends and challenges such as scalability, data privacy, model complexity, and real-time adaptability are discussed. The paper concludes by identifying future research directions for intelligent cloud management systems.