ON THE APPLICATION OF A NEURAL-NETWORK IN THE DESIGN OF CASCADED GRATINGS

This article presents a study of the ARTMAP neural network in designin g cascaded gratings. A neural network can be trained to keep changing the dimensions of the metallic strips, their distance of separation, t heir width, the number of layers required in a multilayer structure, a nd the angle of wave incidence until the frequency response of the str ucture matches the desired one. In the past, the back-propagation (bac k-prop) learning algorithm was used in conjunction with an inversion a lgorithm for the design of frequency-selective surfaces. Unfortunately , both the back-prop algorithm and the inversion procedure are slow to converge. In this work the Fuzzy ARTMAP neural network is utilized. T he fuzzy ARTMAP is faster to train than the back-prop, and it does not require an inversion algorithm. Several results (frequency responses) from cascaded gratings for various angles of wave incidence, layer se paration, width strips, and interstrip separation are presented and di scussed. (C) 1995 John Wiley & Sons, Inc.