Application Of Design Equations And Particle Swarm Optimization Method To Automatic CMOS Multi-Stage Amplifier Sizing

As the size of process characteristics continues to shrink,the interest in multi-stage operational amplifier(opamp)in analog integrated circuit design is increasing.The multi-stage opamp can achieve higher gain,but its compensation design also becomes difficult.This paper presents a study on using analytical design equations in the particle swarm optimization(PSO)for the automatic sizing of multi-stage opamps.Different from the existing researches,this paper attempts to introduce the hand-derived design equations into the heuristic search program based on simulation,which can effectively reduce the search dimension and improve the quality of results of PSO algorithm.Besides,some other algorithms such as genetic algorithm(GA),artificial neural network(ANN)and non-dominated sorted genetic algorithm-II(NSGA-II)based search methods are implemented and tested for comparison.In order to reduce the number of PSO iterations,embedding the relaxation iteration(RI)into PSO is introduced.Besides,to further reduce the number of PSO iterations,using targets to screen the unqualified particles is helpful.However,this method is for reference only because it leads to a decrease in accuracy.The automatic sizing method which combines both design equations and PSO is useful in assisting designers.It provides a feasible way to solve the automatic design of analog circuits.Besides,it provides initial experimental verification for embedding the design equation automatically generated by symbolic calculation into the optimization algorithm.In a word,the research in this paper is meaningful.