TY  - JOUR
T1  - Deep Reinforcement Learning Based Weapon-Target Assignment to Support Military Decision-Making
AU  - Lee, Jaehwi 
AU  - Eom, Chanin 
AU  - Kim, Kyeongsoo 
AU  - Kang, Hyunsu 
AU  - Kwon, Minhae 
JO  - The Journal of Korean Institute of Communications and Information Sciences
PY  - 2025
DA  - 2025/1/1
DO  - 10.7840/kics.2025.50.6.884
KW  - Weapon-target assignment
KW  - Deep reinforcement learning
KW  - Markov decision process (MDP)
KW  - Deep deterministic policy gradient (DDPG)
KW  - Twin delayed DDPG (TD3)
AB  - Recently, there has been significant research into technologies to assist commanders in military decision-making through the weapon-target assignment (WTA) problem. As the modern battlefield has evolved, WTA problems consider realistic and complex environments, where deep reinforcement learning methods can be utilized for dynamic decision-making. This paper proposes a Markov decision process (MDP) model for efficient decision-making in WTA problems, leveraging deep reinforcement learning to optimize the actions of friendly units. As a result, we confirmed that the reinforcement learning model based on the proposed MDP improved the commander's objective achievement by 27.17% and ammo efficiency by 38.61%, while reducing ammo usage cost by 11.98% compared to the heuristic approaches.