Optimizing an amplifier by a many-objective algorithm based on R2 indicator

Author

de la Fraga, Luis Gerardo ; Tlelo-Cuautle, Esteban

Author_Institution

Comput. Sci. Dept., Cinvestav, Mexico City, Mexico

fYear

2015

fDate

24-27 May 2015

Firstpage

265

Lastpage

268

Abstract

It is well known that Pareto based algorithms do not scale when the number of objectives is greater than three. That way, it is an open problem in integrated circuit design because optimizing amplifiers by multi-objective evolutionary algorithms requires evaluating many objectives with many constraints. In this manner, we show the usefulness of MOMBI, an algorithm based on R2 indicator, for improving the optimal sizing of amplifiers designed with MOSFETs. It is highlighted that such kind of algorithms could lead to solve problems in analog integrated circuit optimization when considering many objectives. In addition, a handling constraint scheme is incorporated to MOMBI. The optimization algorithm based on R2 indicator is applied herein to an amplifier that is used to implement a universal active filter showing good SPICE simulations results.

Keywords

Pareto optimisation; active filters; amplifiers; analogue integrated circuits; constraint handling; evolutionary computation; integrated circuit design; MOMBI; MOSFET; Pareto algorithm; SPICE simulation; amplifier optimization; analog integrated circuit optimization; handling constraint scheme; integrated circuit design; many-objective metaheuristic based on the R2 indicator; multiobjective evolutionary algorithm; universal active filter; Algorithm design and analysis; Approximation algorithms; Evolutionary computation; Optimization; Sociology; Statistics; Transistors;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems (ISCAS), 2015 IEEE International Symposium on

Conference_Location

Lisbon

Type

conf

DOI

10.1109/ISCAS.2015.7168621

Filename

7168621