A later and more rigorous classification scheme for one-dimensional cellular automata, see [7], was developed by Robert
Gilman. Here a probabilistic/measure theoretic classification scheme was developed based on the probability of choosing aconfiguration that will stay arbitrarily close to a given initial configuration under forward iteration (evolution).2 Gilman’s classification
partitions the cellular automata rules into three classes. Class A is the class of equicontinuous automata, whereby there
is an open disk of configurations that stay arbitrarily close to the given initial configuration. Automata in class B conform to a
stochastic analog of equicontinuity. Automata in this class have the property that the probability is positive that one can find
(at random) another configuration that can stay arbitrary close to an initial under forward evolution. Automata in class C have
the property that the probability of finding another configuration that stays arbitrarily close to the initial under forward iterations
is 0. Owing to the fact that the lens of measure theory does not distinguish between countably infinite and uncountably
infinite (in the Cantor sense) sets, it is noted that automata in all the classes have some indistinguishable dynamic similarities.
For instance, in both Gilman classes A and B there are an infinite amount of configurations that can stay arbitrarily close to a given
initial configuration under forward evolution. In the classification presented herein, using the Infinite Unit Axiom of Sergeyev, see
[15–19], this similarity is overcome by actually having a numeric representation for the number of configurations in each class
that equal (or match) an initial configuration under forward evolution. Thus making the classes more distinguishable.
A later and more rigorous classification scheme for one-dimensional cellular automata, see [7], was developed by RobertGilman. Here a probabilistic/measure theoretic classification scheme was developed based on the probability of choosing aconfiguration that will stay arbitrarily close to a given initial configuration under forward iteration (evolution).2 Gilman’s classificationpartitions the cellular automata rules into three classes. Class A is the class of equicontinuous automata, whereby thereis an open disk of configurations that stay arbitrarily close to the given initial configuration. Automata in class B conform to astochastic analog of equicontinuity. Automata in this class have the property that the probability is positive that one can find(at random) another configuration that can stay arbitrary close to an initial under forward evolution. Automata in class C havethe property that the probability of finding another configuration that stays arbitrarily close to the initial under forward iterationsis 0. Owing to the fact that the lens of measure theory does not distinguish between countably infinite and uncountablyinfinite (in the Cantor sense) sets, it is noted that automata in all the classes have some indistinguishable dynamic similarities.For instance, in both Gilman classes A and B there are an infinite amount of configurations that can stay arbitrarily close to a giveninitial configuration under forward evolution. In the classification presented herein, using the Infinite Unit Axiom of Sergeyev, see[15–19], this similarity is overcome by actually having a numeric representation for the number of configurations in each classthat equal (or match) an initial configuration under forward evolution. Thus making the classes more distinguishable.
การแปล กรุณารอสักครู่..