home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory.cell-automata
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!cs.utexas.edu!hellgate.utah.edu!asylum.cs.utah.edu!tolman
- From: tolman%asylum.cs.utah.edu@cs.utah.edu (Kenneth Tolman)
- Subject: What is the information content of cellular automata?
- Date: 20 Nov 92 16:20:17 MST
- Message-ID: <1992Nov20.162017.28766@hellgate.utah.edu>
- Organization: University of Utah, CompSci Dept
- Lines: 44
-
-
- What is the information content of a particular configuration for a CA?
- What are references which would discuss this? (see below)
-
-
- One could define the information content for a cellular automata to be
- -log(probability of state) Is this correct?
-
- In S. Amoroso he suggests that the information content
- does not change for a reversible deterministic CA, which seems agreeable.
- This would also correspond to the above metric, for any later state would
- have the same probability as its predecessor. (1)
-
- In T. Toffoli, he suggests that the information content depends
- on the initial state for irreversible deterministic CA. If one has enough
- detail to construct the initial state, then one has preserved the information
- but if there is not enough to reconstruct then one has lost information. (2)
-
- Wolfram suggests that the "entropy" of a CA can be given by the logarithm
- of the "average number of possible states of a system", or as
-
- SUM p(i) log (p(i))
-
- Then he shows a graph with the entropy decreasing for an irreversible
- CA, which is based on the probabilities for each of the possible states the
- system can be in..... this is not entirely clear to me, for the INITIAL
- state would appear to have an equal distribution, and any later states would
- have an unequal probability distribution.. so is he calculating probabilities
- on the infinite limit of that state or what? (the alternative is of course
- that he calculates the probability for a state from all possible initial
- states for a particular time interval) It appears here that the information
- content of a particular state then WOULD be the -log(probability of state) (3)
-
- So what IS the information content of a particular state in a cellular
- automata?
-
-
- 1 S. Amoroso, Journal Comp. Sys. Sciences, Vol 6, pp. 448-464.
-
- 2 T. Toffoli, Journal Comp. Sys. Sciences, Vol. 15, pp. 213-231
-
- 3 S. Wolfram, Rev. Mod. Phys. Vol. 55, No. 3, July 1983, pp. 601-644
-
-
-