Performance measures in the genetic design of digital model-following flight-control systems

Genetic algorithms can be used to provide a convenient means of tuning the digital PID controllers incorporated in a class of digital model-following systems. However, in the genetic tuning of such PID controllers, various performance measures are available to the designer (such as the difference between non-asymptotically optimal and asymptotically optimal eigenvalue locations, and the maximum multivariable generalised tracking error). In this paper, a detailed comparison is made of the genetic design procedure when these two performance measures are minimised in the design of digital model-following flight-control systems for the F-16 aircraft. It is shown that both performance measures constitute equally powerful tools for the solution of complex problems in the design of flight-control systems