Regresar

Nonlinear tuning of aircraft controllers using genetic global optimization: A new periodic mutation operator

Abstract:

Despite the excellent capabilities of genetic algorithms (GAs) for solving difficult nonlinear optimization problems, engineers in some practical design engineering domains are reluctant to incorporate GAs into their procedures because of questions about their effectiveness in terms of achieving results close to the global optimum. Regarding the application of GAs for the solution of flight control problems, very little work on important aspects of the effects of the genetic parameters has been reported. The present work seeks to fill the void by presenting a new mutation operator with excellent results in the application of GAs to the optimization of controller gains of flight control systems. By following a sinusoidal rule for changing the value of the probability of mutation, this new operator allows the GA to perform two stages of high and mild exploration over the solution space and one stage of high exploitation.