Multi-agent deep reinforcement learning in the path planning problem

This work is devoted to multi-agent deep reinforcement learning in the path planning problem. The use of UAV swarms in precision agriculture is justified. It is shown that for the use of drone swarms it is necessary to apply artificial intelligence, in particular reinforcement learning. The task of path planning in the presence of poor quality or absence of GPS navigation is set. The use of the Multi-Agent Proximal Policy Optimization method is proposed. The results obtained showed high quality of path planning in the presence of obstacles and poor quality or absence of GPS navigation.Recieved2025-12-27Recieved2026-02-02Accepted2026-02-11