Dynamic load balancing optimization model based on bidirectional network edge detection algorithm in cloud computing environment

Purpose This study aims to enhance the efficiency of cloud computing resource utilization by addressing the challenge of unbalanced workloads through an intelligent, adaptive load-balancing framework. Design/methodology/approach The proposed solution is a three-stage adaptive load-balancing model grounded in bidirectional network edge detection. First, a bidirectional network-based resource sensing method is developed for real-time online detection of edge device capacities. Second, a dynamic load recognition technique utilizing stable diffusion is employed to monitor device load conditions. Lastly, a mutation particle-based scheduling mechanism is designed to optimize task allocation and resource dispatching within the cloud environment. Findings Experimental evaluations demonstrate the effectiveness of the proposed model, achieving an accuracy of 0.816 and a mean average precision of 0.825. Furthermore, in terms of operational performance, the model significantly reduces average task completion time, bottom-line violation rate and average scheduling overhead, outperforming existing benchmarks. Originality/value This work introduces an innovative integration of bidirectional edge detection, stable diffusion for dynamic load analysis and MP-based scheduling within a unified cloud computing framework. The proposed approach provides a novel perspective on real-time resource awareness and adaptive scheduling in complex cloud environments.