Oriented Genetic Algorithm for Digital Predistortion Model Structure Optimization

A Power amplifier (PA) is an indispensable component in communication systems, and it is also the main nonlinear device. Digital predistortion (DPD) is the widely used method to linear PA owing to its flexibility and high efficiency. However, before using a DPD model, it is necessary to first determine the model structure. In this paper, a novel oriented genetic algorithm (OGA) is proposed to search an optimal DPD model structure, which has a focused search capability and can search the solution domain in a more efficient way. In addition, ridge regression (RR) is introduced to enhance numerical stability and improve the iteration robustness of the algorithm. Furthermore, orthogonal matching pursuit (OMP) is used to further reduce the parameters. The experimental results of a 100MHz signal confirm that the model searched by the proposed OGA can produce excellent performance with relatively lower complexity.