Optimization of Small-Signal Model of GaN HEMT by Using Evolutionary Algorithms

In this letter, a new evolutionary algorithm-based method for the optimization of intrinsic elements of small-signal model (SSM) of GaN HEMT devices is presented. The method uses a unique search space exploration strategy for evolutionary algorithms to obtain an optimized compact SSM from the extracted parameter and measured S-parameters. The validity of the method is verified by comparing the measured S-parameter data of a $2 \times 0.1 \times 50 \,\,{\mu \text {m}}^{2}$ GaN/Si HEMT and a $4 \times 0.1 \times 75 \,\,{\mu \text {m}}^{2}$ GaN/SiC HEMT in the frequency range of 1 to 30 GHz. The modeled data and measured data are in good agreement.