AI-driven optimization in cloud computing: a systematic review of cost, resource management, and security.

Area of Science:

• Cloud Computing
• Artificial Intelligence
• Machine Learning

Background:

• Cloud environments face persistent challenges in cost control, resource allocation, and security.
• Traditional infrastructure approaches are inadequate for addressing these dynamic cloud computing issues.
• Artificial intelligence (AI) and machine learning (ML) offer potential solutions for cloud optimization.

Purpose of the Study:

• To systematically review empirical evidence on AI/ML applications in cloud computing.
• To synthesize findings on AI/ML for cost optimisation, resource management, and security enhancement.
• To identify current research gaps and future directions in AI for cloud infrastructure.

Main Methods:

• Systematic literature review following PRISMA 2020 and Kitchenham-Charters methodology.
• Searched major digital libraries (IEEE Xplore, Web of Science, ScienceDirect, ACM Digital Library) for studies from 2020-2025.
• Selected 18 primary studies based on the PICOC framework and quality assessment (Cohen's κ = 0.86).

Main Results:

• Predictive provisioning and intelligent load balancing reduce operational costs by up to 85%.
• Metaheuristic algorithms improve energy efficiency (30%-40%) and resource utilization (up to 80%).
• Deep learning intrusion detection systems achieve >92% accuracy; ensemble methods reach >97%.

Conclusions:

• AI/ML models are effective structural mechanisms for enhancing cloud economic efficiency, operational resilience, and sustainability.
• Significant cost savings (30%-40%) and efficiency gains are achievable with AI-driven approaches.
• Gaps exist in standardized benchmarking and validation in production, hybrid, and multicloud environments.