移动边缘计算中的联合优化迁移决策和资源分配

Joint Optimization of Offloading Decision and Resource Allocation in Mobile Edge Computing System

移动边缘迁移计算中,边缘服务器之间的协作能为用户提供更高效的服务。本文对正交频分复用上行无线通信系统,基于移动边缘计算技术的任务迁移的子载波选择、用户发射功率和迁移量的联合优化问题进行了研究。在公平性原则下,本文考虑最小化迁移计算的最大时延问题,并提出了一种非凸问题的拉格朗日对偶法解决方案。首先将min-max问题转化为最小化问题,再用泰勒级数将其近似为一个凸问题,最后用拉格朗日对偶法求解。本文还给出特殊情况下的简便算法,适用于低信噪比的通信环境下。仿真结果证实了本算法的收敛性和实用性。

The joint optimization of subcarriers assignment,transmit power and offloading data size in Orthogonal Frequency Division Multiplexing uplink offloading computing was discussed.The optimization goal is to minimize the maximum computing delay of a task,where the objective considers the fairness of collaborative computing.It is a non-convex problem and is transformed into a maximization problem with Dinkelbach’s method,firstly.Then Taylor expansion and Lagrangian dual theorem is adopted to jointly optimize subcarriers assignment,transmit power and offloading data size.Finally,the simulation results confirm the convergence and practicability of the algorithm.