A Trajectory Planning and Task Offloading Collaborative Optimization Method for Multi-UAV Assisted MEC

In multi Unmanned Aerial Vehicle (UAV) assisted Mobile Edge Computing (MEC) systems, there are difficulties in balancing user coverage, task delay, and energy consumption for collaborative trajectory planning and task offloading. To address these problems, a collaborative optimization method based on improved Multi-Agent Twin Delayed Deep Deterministic Policy Gradient (MATD3) algorithm is proposed in this research. This method models the system’s complex interactions as a Multi-agent Markov Decision Process (MMDP), and designs an enhanced state representation integrating service association matrices, to enhance the fine-grained perception capability of UAVs towards service topology. Besides, a global reward mechanism based on system-level pheromone signals is constructed to organically integrate multiple objectives such as coverage, delay, energy consumption and flight safety, guiding multi-UAV to efficiently learn collaborative decision-making strategies. Experimental results show that the proposed method can achieve autonomous optimization of flight trajectory and resource allocation strategies for UAVs under dynamic user demands and environmental constraints, effectively improving user coverage, reducing task processing latency, and optimizing system energy consumption.