Multi-Cloud Security Policy Optimization

Cloud computing has revolutionized the deployment and management of large-scale applications. However, with the advantages of rapid scalability and flexibility come significant security challenges. Optimizing security policies in a multi-cloud environment is a multi-dimensional, complex problem, often involving trade-offs among encryption overhead, open port configurations, privilege levels, and subscription costs for advanced threat detection services. In this paper, we present a detailed comparative study of these metaheuristics for security policy optimization in the cloud. We implement an enhanced objective function that incorporates baseline risk, encryption overhead, open port penalties, and advanced analytics subscription costs. We benchmark the algorithms over 30 independent runs, each with a maximum of 200 iterations, and measure their performance according to best, worst, mean, median, standard deviation (Std), interquartile range (IQR), runtime, and an average rank metric. Our results reveal distinct advantages and trade-offs among the metaheuristics, demonstrating that no single algorithm dominates every scenario. Grey Wolf Optimizer (GWO) and Marine Predators Algorithm (MPA) emerge as strong contenders, achieving low objective values and relatively consistent performance. We discuss these findings in depth, and we provide recommendations on the suitability of each method for large-scale cloud security challenges.