A Design Methodology of MMIC Power Amplifiers Using AI-driven Design Techniques

Abstract

Designing a monolithic microwave integrated circuit (MMIC) power amplifier (PA) is challenging due to the involvement of multiple stages with tens of parameters and several types of simulations. To address this challenge, artificial intelligence (AI) techniques have gained significant attention. This paper presents an AI-driven design methodology for MMIC PAs, which incorporates a surrogate model-assisted global optimization algorithm, data-flow interface, and simulators. The proposed methodology is verified on a practical 3-stage PA design case that operates in 27–31 GHz and contains 30 variables. Notably, The case is successfully synthesized without an initial solution and exhibits good in-band performance consistency. The obtained results demonstrate the potential of AI-driven PA design for future applications.