Abstract:
It is known that the values of functions that depend on a large
number of similar variables are almost constant from the point of view
of an observer evaluating their values at random points of the domains of their definition
(the “nonlinear law of large numbers”).
We show that, under certain normalization conditions, the distribution of values of such
functions tends to a normal distribution as the number of variables grows.
Keywords:law of large numbers, normal distribution, measure concentration principle.