CBA2018-1483 Teoria de Controle e Aplicações

Comparison of Several Genetic Algorithm Strategies on a nonlinear GAPID Controller Optimization Applied to a Buck Converter

Mauricio dos Santos Kaster¹; Fábio Galvão Borges¹; Marco A. Itaborahy Filho¹; Hugo V. Siqueira¹; Fernanda C. Corrêa¹

¹ Universidade Tecnológica Federal do Paraná

Baixar PDF doi:10.20906/CPS/CBA2018-1483

Resumo

Abstract This work presents a comparison of six versions of the Genetic Algorithm to optimize the parameters of the nonlinear GAPID controller (Gaussian Adaptive Proportional, Integral and Derivative), elaborated to control a step-down DC-DC converter. This task comprises 8 free parameters and there is no analytic solution to solve it. Also, the design of the controller is hard to determine because there can exist several near-optimal solutions with different values for the parameters, which defines this problem as multimodal. In this sense, different optimization strategies can lead to different solutions. This paper analyzes the behavior of six distinct Genetic Algorithm strategies and compares the obtained results.

Palavras-chave: Genetic Algorithm; Optimization; Gaussian Adaptive PID Control

Como citar

Mauricio dos Santos Kaster; Fábio Galvão Borges; Marco A. Itaborahy Filho; Hugo V. Siqueira; Fernanda C. Corrêa. “Comparison of Several Genetic Algorithm Strategies on a nonlinear GAPID Controller Optimization Applied to a Buck Converter”. XXII Congresso Brasileiro de Automática. CBA2018. 2018. DOI: 10.20906/CPS/CBA2018-1483