DocumentCode :
2406951
Title :
Estimation of the basins of attraction of recurrent type analog neural networks
Author :
Ciliz, M. Ice emal ; Harova, Arif
Author_Institution :
Dept. of Electr. & Electron., Bogazici Univ., Istanbul, Turkey
fYear :
1992
fDate :
1992
Firstpage :
3496
Abstract :
The basins of attraction (stability regions) of the stable equilibria of the Hopfield type recurrent neural networks are investigated. For estimating the dimensions of these regions, a constructive method is utilized. It is shown that, with increasing number of iterations, larger estimates of equilibrium point basins are generated. The method is tested with an example, and the estimated basins are shown to be within the exact stability regions
Keywords :
Hopfield neural nets; pattern recognition; stability; Hopfield type recurrent neural networks; basins of attraction estimation; equilibrium point basins; recurrent type analog neural networks; stability regions; stable equilibria; Artificial neural networks; Associative memory; Computer networks; Hopfield neural networks; Lyapunov method; Neural networks; Neurofeedback; Recurrent neural networks; Stability; Testing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Decision and Control, 1992., Proceedings of the 31st IEEE Conference on
Conference_Location :
Tucson, AZ
Print_ISBN :
0-7803-0872-7
Type :
conf
DOI :
10.1109/CDC.1992.371201
Filename :
371201
Link To Document :
بازگشت