Comparison of Training Algorithms for Optimal Neural Controllers

J. B. Galván (Spain)


Learning algorithms, optimal control, recursive algorithms


In this work a study of training algorithms for optimal neural controllers is made. Three methods of first and second order in two different versions, recursive and non recursive, have been studied. An adecuate strategy that combines the use of both versions of a second order train ing algorithm, Levenberg Marquardt, is proposed.

Important Links:

Go Back