Antenna Array Beamforming Method Based on Reinforcement Learning Optimization Algorithm

ABSTRACT In current beamforming multi‐objective optimization algorithms, parameter adjustment to satisfy the target pattern given the target pattern is essential. In order to avoid the influence of experience factors on the optimization target weight parameters, we proposed for the first time an intelligent optimization method that combines the optimization algorithm with the RL algorithm to assist in adjust the weight parameters. For different antenna array models, this method can also achieve excellent array beamforming. This article compares the results of PSO with auxiliary adjustment and the results of classic PSO without auxiliary adjustment on three array models and three groups of beam targets, proving the versatility and effectiveness of the algorithm for auxiliary adjustment of intelligent optimization algorithms.