Design and Optimization of a Multi-Octave Power Amplifier Using an Improved PSO Algorithm with Nonlinear Adaptive Hyper-Parameters

Due to the rapid development of modern wireless systems, higher performance requirements are imposed on PA, resulting in PA structures becoming more complex, making PA optimization extremely difficult. This paper presents an improved particle swarm optimization technique with nonlinear adaptive hyperparameters (PSO-NAH) for designing multi-octave PAs. A high-performance multi-octave PA with 10-W GaN HEMTs transistors has been designed and implemented to validate the proposed strategy. In comparison with the PSO algorithm with fixed hyperparameters (PSO-FH) and the PSO algorithm with linear adaptive hyperparameters (PSO-LAH), the proposed technique shows superior performance. As a result, the fabricated PA can achieve an output power exceeding 40.4 dBm, a drain efficiency greater than 60.1%, and a gain over 10.4 dB between 0.5 GHz and 3.5 GHz.