Effective method for diagnosing fault of avionics system

Our method is very effective because it combines the advantages inherent in GA (genetic algorithm) and rough set theory. We discretize the measured data by using GA and diagnose the system fault by means of knowledge discovered by rough set theory. We present two algorithms to discover knowledge: one is to discover knowledge and the other is to update the discovered knowledge. We took a numerical example for the avionics system of a certain fighter aircraft. There were many faults of this avionics system that were difficult to diagnose by existing methods. We applied our method to 500 such difficult-to-diagnose faults and diagnosed 493 faults correctly. This shows preliminarily that our method is very effective and the correct diagnosis rate is as high as 98.6%.