子任务调度和时延联合优化的MEC卸载方案

移动边缘计算(mobile edge computing, MEC)为5G超低时延业务提供了解决方案。如何设计低时延、高效率的任务卸载方案，是MEC面临的主要难题之一。为此，针对端-边协同MEC服务场景，研究了大型计算任务的低时延、低能耗部分卸载方案，通过将用户任务划分为多个有顺序依赖关系的子任务并构建子任务的有向无环关系图，设计了能够最小化卸载时延的子任务调度方案，提出了基于任务复制的最早卸载执行算法，解决了能耗受限下的任务最小时延卸载计算。仿真结果表明，提出的MEC卸载方案能够有效减少任务处理时延，降低系统能耗。

MEC offloading scheme based on joint optimization of subtaskscheduling and delay

Mobile edge computing (MEC) provides solutions for 5G ultra-low latency services. How to design a low-latency and high-efficiency task offloading program is one of the main problems faced by MEC. To this end, the low-latency, low-energy partial offloading scheme of large-scale computing tasks is studied for the end-side collaborative MEC service scenario, and the user task is divided into multiple subtasks with sequential dependencies and directed acyclic graph for subtasks is constructed. A subtask scheduling scheme that can minimize the offload delay is designed, and the earliest offload execution algorithm based on task replication is proposed, which solves the task of minimum delay offload calculation under energy constraints. The simulation results show that the proposed MEC offloading scheme can effectively reduce the task processing delay and reduce system energy consumption.