Circuit-Centric Genetic Algorithm for the Optimization of a Radio-Frequency Receiver

Abstract

This paper presents a design automation method for optimizing parameters of radio-frequency front-ends, specifically aiming to maximize the overall performance of a receiver circuit. In this work, the design target includes a reduction in power consumption and noise figure and an increase in conversion gain. The use of an artificial algorithm for the optimization of an RF receiver is investigated, illustrating how to achieve performance goals in a complex design space composed of multiple inter-related circuit parameters. As the basis of the proposed research, the genetic algorithm, a well-known metaheuristic approach, is chosen and utilized in the optimization process. Since the conventional GA has limitations in circuit optimization, including suboptimal performance and slow convergence due to crossover operations, the concept of a circuit-centric genetic algorithm is proposed as a viable approach that primarily focuses on the use of a mutation process. In addition, by preserving high-performing solutions and incorporating a guiding mechanism toward metric-specific best solutions, the proposed method achieves the target performance much faster compared to other optimization approaches. Therefore, it can be utilized in the optimization of circuit parameter sets in RF receiver design.