物联网边缘计算卸载和资源分配李氏优化

Li’s optimization of edge computing offloading and resource allocation in internet of things

基于物联网终端设备计算资源有限,为减少其运行时延和能耗,提出基于李氏优化的一定时延约束下的能耗最小化的动态计算算法。建立物联网边缘计算的层次模型;建立计算卸载和资源分配的代价模型,得到时延约束时最小化系统能耗的优化函数模型;利用李雅普诺夫优化理论将优化函数分解,在本地执行得到最优计算频率,在边缘服务器端执行得到最优传输功率;利用李雅普诺夫优化中的漂移加罚函数,建立联合计算卸载和资源分配算法,通过调整卸载决策、计算频率和传输功率,得到队列稳定时延有限时,系统的最小化能耗。仿真实验验证了所提算法的有效性。

Based on the limited computing resources of internet of things terminal devices,to reduce its runtime delay and energy consumption,a dynamic algorithm based on Li’s optimization was proposed to minimize energy consumption under certain delay constraints.The hierarchical mode of edge computing in internet of things was established,the cost model of computing offloa-ding and resource allocation was established,and the function optimization model for minimizing system energy consumption at the time of delay constraints was obtained.Lyapunov optimization theory was used to decompose the optimization function,the optimal computation frequency was obtained by local execution,and the optimal transmission power was obtained by edge server execution.The shift penalty function optimized by Lyapunov was used,a joint algorithm of computing offloading and resource allocation was established.By adjusting the offloading decision,computing frequency and transmission power,the minimum energy consumption of the system was obtained when the queue stability delay was limited.Simulation results show the effectiveness of the proposed algorithm.