Deployment Method of Emergency Communication Hub Based on Deep Reinforcement Learning

An emergency communication hub contains multiple communication elements. The deployment locations of elements directly determine the communication efficiency of the entire hub. However, manual deployment often requires additional knowledge and longer planning time. This often results in commanders being unable to plan a high-quality deployment scheme immediately. To address this issue, this paper proposes a new algorithm based on deep Q-learning, which uses a fully convolutional network as its value network and is trained with the data accumulated in the example replay buffer. Compared to manual method, the experimental results show that the proposed approach can accomplish the task better and improve the deployment efficiency significantly.