边缘计算计算卸载与资源分配联合优化算法

Joint Optimization Algorithm of Edge Computing Computing Offloading and Computing Resource Allocation

针对后勤资产管理系统“本地—云”结构的不足,设计了边缘计算卸载决策与资源分配服务联合优化算法,将原始数据解算任务卸载至边缘端,提供更优卸载决策与资源分配方式.根据任务、计算能力、功率等信息建立时延能耗系统代价模型,基于二分法、拉格朗日乘子法、改进的粒子群算法完成问题的求解,实现多用户多节点有云参与的联合迭代寻优.实验结果表明,该方法有效降低系统总代价,降低Random算法总代价的59.34%,Greedy算法的45.74%,STPSO算法的24.07%.

According to the shortcomings of the structure of"local-cloud"of logistics asset management system,an edge computing offloading decision-making and resource distribution service joint optimization algorithm is designed to address the shortcomings of the"local-cloud"structure of the logistics asset management system.By shifting the initial data computation job to the edge,a better unloading choice and resource distribution strategy are made possible.The time-delay energy consumption system cost model is developed in accordance with the tasks,computing capability,and other data.Dichotomy,the Lagrange multiplier method,and an improved particle swarm optimization algorithm are used to solve the problem,and realize multi-user and multi-node joint iterative optimization with cloud participation.The experimental results demonstrate the effectiveness of the method in lowering system costs overall and reduce the costs of the Random,Greedy,and STPSO algorithms by a combined total of 59.34%,45.74%,and 24.07%.