LEARNING NOVEL FIGHTER COMBAT MANEUVER RULES VIA GENETIC ALGORITHMS

This paper reports on a project where a genetics-based machine learnin g system acquired rules for novel fighter combat maneuvers through sim ulation. In this project, a genetic machine learning system was implem ented to generate high angle-of-attack air combat tactics for the NASA X-31 research aircraft. The Genetic Learning System (GLS), based on a learning classifier system approach, employed a digital simulation mo del of one on one air combat and a Genetic Algorithm to develop effect ive tactics for the X-31. The resulting maneuvers allowed the X-31 to successfully exploit its post-stall capabilities against a conventiona l fighter opponent, demonstrating the ability of the GLS to discover n ovel tactics in a dynamic air combat environment. Moreover, the projec t demonstrates how genetic machine learning can acquire rules that imp lement novel approaches to unforeseen problems via experience with sim ulations.