A Hybrid Optimization Algorithm for Large-Scale Combinatorial Problems in Cloud Computing Environments

Combinatorial problems, such as task scheduling and resource allocation, present significant challenges in cloud computing due to the exponential growth of solution spaces. Conventional optimization algorithms often prove inadequate in efficiently handling large-scale problems, resulting in suboptimal resource utilization and increased operational costs. To address these limitations, this study proposes a hybrid optimization algorithm that combines the exploration capabilities of metaheuristic methods with the precision of exact optimization techniques. The proposed approach integrates genetic algorithms (GA) and particle swarm optimization (PSO) with branch-and-bound, facilitating efficient search and refinement of solutions. Furthermore, the algorithm incorporates domain-specific enhancements, including task prioritization heuristics and resource clustering, to reduce computational complexity. Learning-based adaptations, such as reinforcement learning and predictive modeling, are employed to dynamically adjust algorithm parameters and enhance adaptability to varying cloud loads. This ensures real-time responsiveness while maintaining cost-effectiveness. The performance of the hybrid algorithm is evaluated on benchmark datasets and compared with state-of-the-art optimization methods. Experimental results demonstrate significant improvements in scalability, solution quality, and execution time. The algorithm also exhibits robust adaptability to dynamic cloud environments, rendering it a practical and efficient tool for addressing large-scale combinatorial problems. This study highlights the potential of hybrid and adaptive optimization techniques to overcome the challenges of resource management in cloud computing, offering a scalable and cost-effective solution for modern cloud infrastructures.