Research on collaborative adversarial strategies for drone swarms based on deep reinforcement learning

With the continuous advancement of remote-controlled aircraft technology and the growing demand for intelligence, multi-robot collaborative drone swarms are playing an increasingly important role in military confrontations. Conventional multi-agent systems (MAS) often struggle with inefficient data utilization and limited adaptability in decision-making when responding to dynamic environments, making it challenging to meet the diverse demands of battlefield settings. To address these limitations, this paper proposes a UAV Swarm Dynamic Strategy Gradient Algorithm (UD-SGA) based on multi-agent reinforcement learning (MARL). This algorithm combines the Deep Deterministic Policy Gradient (DDPG) and Multi-Agent Deep Deterministic Policy Gradient (MADDPG) methods to improve training stability and accelerate policy updates. Empirical testing in complex, variable, multi-dimensional adversarial scenarios demonstrates that the proposed strategy effectively enhances the adaptability and robustness of drone swarms, enabling them to execute multiple tasks and address multiple targets, thereby significantly improving combat effectiveness in confrontational settings.