True worst-case circuit tolerance analysis using genetic algorithms and affine arithmetic

Abstract

In this paper, a new approach to the calculation of the true worst case in circuit tolerance analysis (TWC-CTA) with parameters characterized by large uncertainties is presented. It is based on the joint use of genetic algorithms (GAs) and affine arithmetic (AA). The GAs are used to minimize the underestimation error which affects stochastic methods while the AA is applied to minimize the overestimation error which affects interval arithmetic methods in TWC-CTA problems. The joint GA-AA approach presented in this paper ensures a reliable TWC evaluation.