A Fusion Deep Q-Learning and Particle Swarm Optimization Algorithm for Adaptive Resource Allocation in Cloud Computing Circumstances

Efficient resource allocation in cloud computing remains a critical challenge due to the dynamic nature of workloads, increasing service-level expectations, and the need to balance multiple optimization goals such as execution time, cost, and energy consumption. Traditional heuristic-based scheduling methods often fail to adapt to unpredictable workload patterns, leading to suboptimal performance and SLA violations. To address this issue, this study proposes a novel hybrid algorithm that integrates Deep Q-Learning (DQL) with Particle Swarm Optimization (PSO) to enable intelligent and adaptive decision-making in cloud environments. The DQL component learns optimal resource allocation policies through continuous interaction with the environment, while the PSO module enhances action selection by performing global optimization and accelerating convergence. The proposed framework was evaluated using real-world workload datasets and benchmark simulators such as CloudSim. Experimental results demonstrate that the hybrid model outperforms existing standalone and hybrid algorithms by achieving up to a 35% reduction in task execution time and a 40% improvement in resource utilization. Additionally, the method shows strong statistical significance and robustness under varying system conditions. These findings highlight the effectiveness of combining reinforcement learning with swarm intelligence to address complex, multi-objective resource scheduling problems in cloud computing.