Title :
A/I supervisor for F-4 weapon coefficient optimization
Author :
Allred, Lloyd G.
Author_Institution :
Ogden Air Logistics Center, Hill AFB, UT, USA
Abstract :
A hybrid technique of nonlinear regression analysis coupled with gradient descent was developed for F-4 weapon coefficient optimization, reducing computation time to two or three minutes. An additional artificially intelligent supervisor was used to detect algorithm instabilities, make corrections, and restart the process. The supervisor uses the first law of good electrical engineering (when the thing goes unstable, turn down the gain). The new algorithm not only produces better solutions, but works without operator assistance. It is concluded that the result will be more accurate weapon delivery
Keywords :
artificial intelligence; military computing; optimisation; weapons; A/I supervisor; Air Force; F-4 weapon coefficient optimization; artificially intelligent supervisor; ballistics; energy function; gradient descent; hybrid technique; nonlinear regression analysis; supervised learning; weapon delivery; Application software; Artificial intelligence; Artificial neural networks; Logistics; Military computing; Nonlinear equations; Predictive models; Programming; Regression analysis; Weapons;
Conference_Titel :
Aerospace and Electronics Conference, 1990. NAECON 1990., Proceedings of the IEEE 1990 National
Conference_Location :
Dayton, OH
DOI :
10.1109/NAECON.1990.112797