With the popularity of computationally intensive applications, more and more computing resources are required. Mobile edge computing (MEC) is widely applied as an effective method to meet the increasing computing demands. In a relatively stable state, MEC can provide computing services with low latency and energy consumption. However, in special cases such as communication traffic, the unmanned aerial vehicle (UAV), by taking advantage of its mobility and flexibility, can assist the edge server to cope with the challenge of instantaneous computing surge. In this study, the authors consider a UAV-enabled edge computing system. In addition to delay and energy consumption, the authors also consider computing resources costs in the offloading model. Besides, in order to minimise the computing cost of each mobile user (MU), they apply the non-cooperative game method to model the channel and computing resources competition among MUs. Then, the authors prove that the proposed game is an ordinal potential game and the existence of Nash equilibrium in the game. The authors propose the UAV-enabled computation offloading (UECO) algorithm to obtain the equilibrium strategy. Finally, the authors show that the UECO algorithm can quickly converge through iterative experiments, and it can achieve lower computing cost through comparative experiments.

中文翻译：

---

支持无人机的边缘计算中具有成本效益的计算分流

随着计算密集型应用程序的普及，需要越来越多的计算资源。移动边缘计算（MEC）作为满足日益增长的计算需求的一种有效方法而被广泛应用。在相对稳定的状态下，MEC可以为计算服务提供低延迟和低能耗。但是，在通信通信等特殊情况下，无人机（UAV）通过利用其移动性和灵活性可以帮助边缘服务器应对即时计算浪潮的挑战。在这项研究中，作者考虑了支持无人机的边缘计算系统。除了延迟和能耗之外，作者还考虑了分流模型中的计算资源成本。此外，为了最小化每个移动用户（MU）的计算成本，他们应用非合作博弈方法对MU之间的渠道和计算资源竞争进行建模。然后，作者证明了所提出的博弈是有序的潜在博弈，并且博弈中存在纳什均衡。作者提出了支持无人机的计算卸载（UECO）算法来获得均衡策略。最后，作者表明，UECO算法可以通过迭代实验快速收敛，并且可以通过比较实验实现较低的计算成本。