No Title
Next: Introduction
= 10000
Abstract:
Wolfram [1] proposed a classification of cellular automata
according to their observed behavior. We compare this classification
with a classification of cellular automata according to their mean field approximations.
The mean field classification is a good predictor of Wolfram class
behavior. Membership in a Wolfram class must be decided empirically,
while membership in a mean field class is decided analytically. The
mean field classification provides quantitative estimates of cellular automata
statistical
properties, such as invariant densities. The mean field classification
is parametric: small variation of these parameters typically produces small
variation in the properties of the rules described.
Wed Nov 9 18:54:15 GMT 1994