Nine-Dimensional Gradient Optimization for Automated RF Power Amplifier Design

Designing RF amplifier circuits can be a tedious and time-consuming process, often involving numerous iterations of various types of simulation. Instead of using repeated back-and-forth source-pull, load-pull, power sweep, and bias sweep simulations, this work presents a fast optimization method that simultaneously optimizes amplifier design parameters to meet user-defined design criteria. Gradient maximization optimization is applied over the multi-dimensional hyper-space of the available design variables. This can serve usefully in the development of an automated amplifier power-amplifier circuit computer-aided design tool, minimizing the required human expertise, time, and back-and-forth effort.