Optimization Design of Multiple Antenna Arrays Based on PSO-BP

To improve the optimization efficiency of antenna arrays, this paper adopts a novel proxy model approach. The approach combines genetic algorithms with neural network models, constructing a PSO-BP neural network model to enhance the predictive capabilities of the overall model. Additionally, data augmentation and weight zeroing techniques were utilized to optimize the neural network. Taking the decoupling problem of microstrip antenna arrays as a case study, this paper compares the proposed method with other approaches, validating its effectiveness.