In infrastructure-less scenarios such as rural environments, wild emergency response, military applications and disaster relief, unmanned aerial vehicles (UAVs) are capable of providing enhanced mobile edge computing (MEC) services for ground users. Although small latency is the most important advantage of MEC system, how to provide delay aware scheduling in UAV-enabled MEC system still remains unsolved. In this study, the authors investigate the delay aware scheduling problem in UAV-enabled orthogonal frequency division multiple access (OFDMA) MEC system and formulate two non-convex optimisation problems. Moreover, they consider uplink and downlink architecture with characteristics in different UAV-ground links and traffic load. Furthermore, they propose two novel multi-stages resource allocation algorithms, i.e. the JSPA-T and JSPA-F algorithms with respect to downlink transmit power allocation and sub-carrier assignment. The mathematical frameworks with duality theory based alternative search optimisation and successive approximation method are proposed. The simulation results validate the performance improvement of the proposed solutions as well as the fast converge behaviour and small computational complexity.

中文翻译：

---

启用无人机的OFDMA移动边缘计算系统中的延迟感知调度

在农村环境，野外应急响应，军事应用和救灾等缺乏基础设施的场景中，无人驾驶飞机（UAV）能够为地面用户提供增强的移动边缘计算（MEC）服务。尽管小的等待时间是MEC系统的最重要优势，但是如何在支持UAV的MEC系统中提供延迟感知调度仍然悬而未决。在这项研究中，作者研究了启用无人机的正交频分多址（OFDMA）MEC系统中的延迟感知调度问题，并提出了两个非凸优化问题。此外，他们考虑了具有不同UAV地面链路和流量负载特征的上行链路和下行链路架构。此外，他们提出了两种新颖的多阶段资源分配算法，即 关于下行链路发射功率分配和子载波分配的JSPA-T和JSPA-F算法。提出了以对偶理论为基础的交替搜索优化和逐次逼近的数学框架。仿真结果验证了所提出解决方案的性能改进以及快速收敛行为和较小的计算复杂性。