Efficient Cloud Resource Allocation Using a Hybrid Approach of Gradient Descent and Genetic Bee Colony Optimization(HGD-GBCO)

Efficient resource allocation in cloud computing environments is essential for improving system performance and reducing operational costs. This study presents a hybrid optimization framework HGD-GBCO that combines Gradient Descent (GD) and Genetic Bee Colony Optimization (GBCO) to leverage both fast local convergence and broad global search capabilities. Unlike existing hybrid methods, HGD-GBCO integrates a stagnation-aware coordination mechanism that dynamically switches between exploration and exploitation phases, promoting solution diversity and reducing the likelihood of premature convergence.Extensive experiments were conducted using benchmark algorithms, including Simulated Annealing (SA), Genetic Algorithm (GA), and Tabu Search (TS). The proposed HGD-GBCO achieved the lowest average VM response time ($\text{3 6 1. 1 1 ~ m s}$), minimal storage latency ($\text{0. 9 ~ m s}$), and reduced memory utilization (75 %), demonstrating superior performance across key cloud metrics. Additionally, tests involving 500 tasks across 20 virtual machines confirmed that parallel execution with HGD-GBCO consistently outperforms sequential approaches in terms of resource efficiency. These results highlight the robustness, scalability, and practical effectiveness of HGD-GBCO for dynamic and large-scale cloud resource optimization.