Genetic slow-mode/fast-mode optimisation of digital PID controllers

Abstract

The optimal parameter settings for digital PID controllers incorporated in model-following systems are complicated, unknown functions of the underlying sampling frequency. In order to simplify the design process, genetic algorithms have previously been used to tune the restricted set of controller parameters affecting the fast modes of the closed-loop system. However, it is important to investigate the effects of both slow and fast modes on model-following behaviour by using an enlarged set of controller parameters in the genetic optimisation procedure. This use of genetic algorithms is illustrated in this paper by the design of model-following flight-control systems for the F-16 aircraft.