面向QoS与负载均衡的边缘计算能耗感知资源调度算法OA

To address the resource bottleneck problem caused by the surge of terminals and tasks in edge com-puting,as well as the deficiency of existing scheduling algorithms neglecting the multi-objective collaborative op-timization of quality of service(QoS),load,and energy consumption,a QoS-oriented energy-aware load-balanc-ing resource scheduling(QEALBRS)algorithm is proposed.Firstly,a QoS prediction model integrating spa-tial-temporal and energy consumption features is constructed,and precise resource demand prediction is achieved by combining graph neural networks and collaborative filtering.Secondly,a multi-objective scheduling model with the joint goals of minimizing QoS loss,balancing load and optimizing energy consumption is established.Fi-nally,a double-depth Q-network(DDQN)-beetle antennae search algorithm is designed for efficient solution.Ex-perimental results show that under the premise of ensuring QoS,the QEALBRS algorithm significantly improves resource utilization and load balancing,with system energy consumption and completion time reduced.