A gradient optimization methodology for use with nonlinear envelope simulat
ion is presented. Emphasis is placed on efficient evaluation of cost functi
on gradients. To this end, an envelope sensitivity equation is derived and
methods for its efficient solution are proposed. In the case of circuits of
moderate size, the solution of the sensitivity equation is shown to be sim
ple and inexpensive. The more troublesome case of larger circuits is treate
d by a novel application of a recently developed iterative linear equation
solver. The result is a general-purpose, rigorous optimization methodology
for the fine tuning of gain, adjacent channel power, power efficiency, and
related performance measures in radio frequency (RF) amplifiers for digital
communications. As an illustrative example, the optimization of a feedforw
ard-linearized power amplifier is presented. (C) 2000 John Wiley & Sons, In
c.