Genetic Algorithm Optimized Back Propagation Neural Networks in Behavioral Modeling of Power Amplifiers Excited by 5G NR Signal

In this paper, a back propagation (BP) neural network (NN) based on genetic algorithm (GA) optimization is presented to characterize the physical properties of wideband radio frequency (RF) power amplifiers (PAs) with 5G new radio (NR) test. Taking the weight and bias matrices of neurons inside the BP NN as hyper-parameters, the initial parameters are optimized by iterative calculations of genetic algorithm to solve the problem that BP neural network is prone to fall into local optimal solutions, significantly improving the probability to achieve the best modeling accuracy of RF PAs. The experimental results show that the proposed model achieves higher modeling performance compared with the existing one.