Task Offloading Strategy in Multi-UAV Mobile Edge Computing via Deep Reinforcement Learning

Unmanned Aerial Vehicles (UAVs) serve as a powerful supplement to Mobile Edge Computing (MEC) systems, offering exceptional flexibility and mobility, which enable rapid deployment and the ability to adapt to rapidly changing demands. However, in this dynamic environment, existing research has not adequately considered the impact of user mobility and edge server load on task offloading performance. To address this gap, this paper proposes a task offloading algorithm based on Deep Deterministic Policy Gradient (DDPG), aiming to minimize the total system energy consumption and delay caused by computation offloading under limited system resources. Furthermore, to solve the load balancing issue among UAVs, this paper introduces a task migration mechanism, which further enhances task offloading efficiency. Simulation results demonstrate that the proposed strategy effectively optimizes overall delay and energy consumption under resource-constrained conditions, while also achieving load balancing.