GENETIC ALGORITHMS FOR RECONFIGURATION

This paper describes the use of genetic algorithms for the reconfigura tion of the control system of a contemporary fly-by-wire aircraft. Con trol system reconfiguration is required in case of failures or damages in the aircraft actuator system (elevator, aileron, rudder, thrust,.. .). The control system has to be adapted in order to have intact actua tors take over the task of the impaired ones. Besides knowledge based reconfiguration for that part of the fault that can be quantitatively sensed, automatic control system parameter adaptation by means of robu st generic algorithms can take care of those fault impacts that have n ot been numerically determined. Genetic algorithms utilize the same th ree procedures (survival of the fittest, recombination, and mutation) mother nature uses to adapt and optimize her individuals. Such algorit hms have been proven very effective during flight tests to adapt the f eedforward controller of a model following control system of a modern variable stability aircraft after a 50% elevator effectiveness reducti on. The benefits of genetic algorithms for the adaptation of model fol lowing control systems will be addressed as well as advanced operators and techniques in genetic search.