Title :
Training multilayer feedforward neural networks using dynamic programming
Author_Institution :
Dept. of Math., Alabama Univ., Tuscaloosa, AL, USA
fDate :
31 Mar-2 Apr 1996
Abstract :
The dynamic programming method is implemented as an alternative supervised training algorithm for designing multilayer feedforward artificial neural networks. The author demonstrates its feasibility and competitiveness by two examples. It helps to set a stage for further research and applications
Keywords :
dynamic programming; feedforward neural nets; learning (artificial intelligence); multilayer perceptrons; competitiveness; dynamic programming; multilayer feedforward neural networks; supervised training algorithm; Algorithm design and analysis; Artificial neural networks; Computer networks; Dynamic programming; Feedforward neural networks; Multi-layer neural network; Neural networks; Neurons; Optimization methods; Sun;
Conference_Titel :
System Theory, 1996., Proceedings of the Twenty-Eighth Southeastern Symposium on
Conference_Location :
Baton Rouge, LA
Print_ISBN :
0-8186-7352-4
DOI :
10.1109/SSST.1996.493491
