A Joint Optimization Scheme in Heterogeneous UAV-Assisted MEC

Mobile Edge Computing (MEC) is considered as a promising technology to meet the high-quality service requirements of emerging applications in mobile intelligent terminals. It can effectively handle computation-intensive and latency-sensitive tasks in the Internet of Things (IoT). However, location-fixed edge servers in MEC cannot efficiently handle time-varying tasks in the hot-spot area. Therefore, it is necessary to utilize the Unmanned Aerial Vehicle (UAV) with communication and computation resources making on-demand network deployment for handling the time-varying tasks above. In this paper, we build a novel MEC system based on heterogeneous multi-UAV, in which we take both the UAV scheduling problem and the task allocation problem into consideration. What's more, in order to minimize the system energy consumption, we propose a joint optimization method, named JoSA, for the two problems mentioned above. To be specific, we first regard the UAV scheduling problem as a knapsack problem. Based on this, we then divide the tasks in the hot-spot area according to geographic location and allocate them in different situations. Finally, compared with the other two benchmarks, the simulation experiments show that our method demonstrates good generalization ability and makes better performance with a reduction of 8% and 11% in system energy consumption, and 3% and 4% in system time cost, respectively.