Joint Association, Trajectory, Offloading, and Resource Optimization in Air and Ground Cooperative MEC Systems

The air and ground cooperative mobile edge computing (MEC) network provides a new paradigm for the development of the Internet of Things (IoT), which enhances the coverage of the terrestrial base station (TBS) and deploys computing resources near IoT devices. In this paper, we construct a UAV-assisted MEC system for IoT networks and design a data processing procedure. The UAV collects data from devices as a relay and makes decisions to offload some tasks to the central server connected with the TBS, while the onboard edge server in the UAV can perform local computing. Furthermore, we jointly optimize the device association, UAV's trajectories, task offloading, and resource allocation to reduce the energy consumption of the entire system. To solve this problem, we decompose it into three tractable sub-problems and use the block coordinate descent (BCD) method to iteratively optimize each set of control variables. Among them, device association is formulated as a linear programming problem, while UAV's trajectory optimization is transformed into a convex problem by introducing slack variables and using successive convex approximation (SCA). The offloading and resource assignment problem is proved to be convex via theoretical analysis and problem transformation. In addition, we derive the optimal relationship between computation duration and computing energy, which greatly reduces the complexity of problems. Simulation results show that the designed system and the algorithms can significantly reduce the total energy consumption, and the offloading strategies have a decisive impact on computation energy consumption.