Joint Computing Offloading and Trajectory for Multi-UAV Enabled MEC Systems

Abstract

The cooperation of multiple unmanned aerial vehicles (UAVs) is investigated to provide auxiliary computing services for ground users. First, the system cost is defined considering the energy consumption of the UAV, the energy consumption of the user, and the delay of the user at the same time. Take into account the dynamic allocation of bandwidth by the user and the dynamic allocation of computational resources by the UAV, the flight trajectory of UAVs, the offloading object and the offloading ratio of users are jointly optimized to minimize the system cost. Due to the dynamic and long-term feature of the problem, it is described as a Markov decision process. A joint computing offloading and trajectory algorithm is proposed based on the PPO in deep reinforcement learning. Simulation results show the convergence of the proposed algorithm. The proposed algorithm has superior performance compared with the benchmark algorithms.