Automated design method for parameters optimization of CMOS analog circuits based on adaptive genetic algorithm

Abstract

A new method for optimizing the parameters of CMOS analog circuits based on adaptive EGA (elitist genetic algorithm) is proposed in this paper. In the method the Hspice simulation tool is called to evaluate the fitness of every circuit repeatedly in a generation. According to the results of the evaluation better circuits can be reserved. By adjusting the parameters of transistors through EGA the evolution can find the circuit which will satisfy our specifications. The outcome of the experiment for a two-stage operational amplifier shows that this is an accurate and promising way in determining the device sizes in an analog circuit.