The paper proposed a deep Q-learning method to solve the weapon target assignment problem with dynamic damage rates and target values.The DQN method adopted a double network structure.A straightforward method is proposed for modeling the state and reward function of the DQN,which is designed based on the objective function of the weapon target assignment problem.The proposed DQN model was tested by using several simulated scenarios.Results showed the model can converge fast and effectively solve the weapon target assignment problem with dynamic damage rates and target values.The proposed method achieved a better total damage rate and less computation time than the particle swarm-based method.The proposed method can be applied to weapon target assignment problem under the scenario of combat mission planning and combat simulation.
关键词
武器目标分配/深度强化学习/动态毁伤/动态目标价值
Key words
weapon target assignment/deep reinforcement learning/dynamic damage rate/dynamic target value
维普
万方数据