ABSTRACT
Feedforward N N are the most used in the present-day intelligent control o f
nonlinear dynamic systems. Besides, feedforward N N are mathematically very
close, and sometimes even equal, to the fuzzy logic models. Both neural and fuzzy
logic approximation techniques could be given graphical representation which
could be called a neural network or fuzzy logic model. Wi th such an interpretation
o f N N or F L M tools we did not add anything new to the approximation theory but,
from the point o f view o f implementation (primarily in the sense o f parallel and
massive computing), this graphical representation is a more desirable property.