Combined genetic algorithm and neural network technique for transistor modeling

A genetic algorithm neural network (GANN) technique for transistor modeling is presented. The proposed method is based on genetic optimization to extract the extrinsic elements of the equivalent circuit model; while the intrinsic part is modelled using a combined genetic neural network approach. The developed procedure has been applied to 1-mm gate width GaN high electron mobility transistor (HEMT) and validated by small- and large-signal measurements.