面向多用户的移动边缘计算任务分级处理卸载机制

Multi-User-Oriented MEC Task Hierarchical Processing Offloading Mechanism

在信道资源受限情况下,最小化卸载过程中的时延和能耗是改善基于无线携能通信的多用户移动边缘计算(MEC)网络卸载性能的关键因素之一。通过规划计算任务的卸载比重和链路传输过程中的信道分配,提出一种多任务分级处理机制(MHPM),以实现计算卸载过程中信道资源的合理调度。同时,根据移动终端设备在MEC卸载过程中的平均时间消耗和能量消耗,构建了约束多目标优化问题的数学模型,并结合MHPM和约束非主导的排序遗传算法Ⅱ求解该模型,从而实现了设备时延与能耗之间的有效均衡。仿真实验结果表明,采用MHPM可以降低设备在卸载过程中的平均时间消耗和能量消耗,而利用约束多目标优化算法可以得到目标函数的最优解。

When the channel resourcesis limited,minimizing the delay and energy consumption during offloading is the key to improve the performance of the multi-user mobile edge computing(MEC)network based on simultaneous wireless information and power transfer.By planning the offloading ratio of computing tasks and the channel allocation during the link transmission,a multi-task hierarchical processing mechanism(MHPM)is proposed.It can realize the rational scheduling of channel resources in the process of computing offloading.According to the average time consumption and energy consumption of mobile device in the process of MEC offloading,a mathematical model of constrained multi-objective optimization problem is established.This model is solved by combining MHPM and non-dominated sorting genetic algorithm-I,which effectively balances the relationship between device's delay and energy consumption.Simulation results show that MHPM can reduce the average.time consumption and energy consumption of devices in offloading process,and the optimal solution of objective function can be obtained by using constrained multi-objective optimization algorithm.