Deep reinforcement learning is one kind of machine learning algorithms which uses the maximum cumulative reward to learn the optimal strategy. The difficulty is how to ensure the fast convergence of the model and generate a large number of sample data to promote the model optimization. Using the deep reinforcement learning framework of the AlphaZero algorithm, the deployment problem of wireless nodes in wireless ad hoc networks is equivalent to the game of Go. A deployment model of mobile nodes in wireless ad hoc networks based on the AlphaZero algorithm is designed. Because the application scenario of wireless ad hoc network does not have the characteristics of chessboard symmetry and invariability, it cannot expand the data sample set by rotating and changing the chessboard orientation. The strategy of dynamic updating learning rate and the method of selecting the latest model to generate sample data are used to solve the problem of fast model convergence.

The construction of mobile ad‐hoc networks on the battlefield is mainly planned by staff or automatically planned with the help of network topology planning models in the network planning software. Most of these algorithms are actually more or less based on human knowledge or thinking ways to model network entities, environments, and rules, and the accuracy and rationality of models are not strong enough to match the changed battlefield. The AlphaZero algorithm generalised in chess and other games provides a new intelligent method to solve complex problems in the military field. Based on the AlphaZero algorithm, this study proposes a method for intelligent deployment of mobile ad‐hoc networks with tactical communication node vehicles. Making an analogy between deploying tactical communication node vehicles and playing Go, the authors construct a deep reinforcement learning model for deployment of communication node vehicles. Starting from random play, and giving no domain knowledge, only setting the judgment of the network structure, with training the designing strategy value deep neural network by self‐play reinforcement learning, they successfully deployed communication node vehicles on tabula rasa map and constructed battlefield mobile ad‐hoc networks with deep reinforcement learning and Monte–Carlo tree search.