Simulated performance optimization of GaAs MESFET amplifiers

A robust global optimizer that is capable of optimizing linear, small-signal and nonlinear, large-signal figures of merit (FOMs) in the presence of a high density of local minima and noise is presented. The optimizer is linked with a large-signal MESFET model and used to investigate the optimization process and parameter extraction for GaAs MESFET amplifiers given the complex structure of the RF performance FOM hypersurface. It is demonstrated that it is necessary to perform optimization using all available physical parameters simultaneously as opposed to sequential optimization using only a few parameters at a time