Live demonstration: High-level optimization of ΣΔ modulators using multi-objetive evolutionary algorithms

Abstract

This demo shows how to use multi-objective evolutionary algorithms for the optimum high-level design of ΣΔ analog-to-digital converters. The methodology illustrated in the demo is based on the combination of SIMSIDES, a SIMULINK-based time-domain behavioral simulator for ΣΔ modulators, with multi-objective optimization techniques. The proposed methodology allows designers to explore the design space in an efficient and intuitive way in order to fulfill a number of different design objectives simultaneously, by finding out the best sets of target specifications - defined as Pareto-optimal fronts. The presented approach can be extended to several kinds of optimizers implemented in MATLAB, and diverse examples are illustrated so that visitors will learn how to apply it to their own designs and projects. Although the demo is focused on ΣΔ ADCs, the tools shown in the demo can be used for the optimization of any other analog integrated circuits and systems.