Machine-Learning-Based Design Automation for Optimizing Analog/RF Circuit Applications

Abstract

This paper introduces an artificial-intelligence (AI)-based program designed to optimize and automate the circuit design process using the computational capabilities of computers. The program highlights the automation of optimization processes for various representative basic high-frequency (radio-frequency, RF) circuit blocks, which are commonly used in electronic systems. It emphasizes the significant potential for the advancement of circuit design automation through the integration of algorithms, machine learning, and analog circuit design techniques. Multiple algorithms can be employed for circuit design automation, enabling highly efficient identification of the circuit's optimal performance. Moreover, the paper exhibits the use of the Figure of Merit (FoM) as an evaluation metric for circuits, which allows the algorithm to assess circuit performance and determine the direction of learning in a highly effective manner. The overall presentation demonstrates the optimization of various circuit parameters through an automated program utilizing algorithms and FoM.