Genetic design of unconstrained digital PID controllers

Abstract

In previous genetic design procedures, the equations for the digital PID controllers were incorporated into the genetic algorithm in order to obtain optimally tuned values of various controller parameters for finite sampling frequencies. The performance of PID controllers constrained by such design equations may be sub-optimal and so this paper illustrates the use of genetic algorithms in selecting controller matrices for PID controllers without using controller design equations. This unconstrained genetic design methodology is illustrated in this paper by the design of model-following flight-control systems for the F-16 aircraft.