Hybrid small-signal modeling of GaN HEMTs based on improved genetic algorithm

—In this paper, an improved genetic algorithm-based hybrid direct-optimal extraction method for small-signal model of GaN HEMT devices is proposed. Simulated annealing algorithm and fuzzy adaptive strategy are incorporated into genetic algorithm for model parameter optimization, to overcome the shortcoming of conventional genetic algorithm that tend to fall into local optimum solutions. The validity of proposed algorithm is verified by using a 20-element GaN HEMT small-signal equivalent circuit model under different operating conditions. The results show that the modeled S-parameters agree well with the measured S-parameters in the frequency range of 0.5–20.5 GHz.