home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory.cell-automata
- Path: sparky!uunet!caen!hellgate.utah.edu!asylum.cs.utah.edu!tolman
- From: tolman%asylum.cs.utah.edu@cs.utah.edu (Kenneth Tolman)
- Subject: Undecidability of configurations/progression
- Date: 8 Nov 92 21:02:19 MST
- Message-ID: <1992Nov8.210220.29731@hellgate.utah.edu>
- Organization: University of Utah, CompSci Dept
- Lines: 23
-
- There are some things which must be undecidable for cellular automata...
- see if you agree!
-
- 1) Decaying into a cycle. For any CA capable of supporting universal
- computation, it could be set up to perform some computation. If that
- computation fell into a cycle, it would be non halting, and thus we
- would have a decision procedure for halting or non halting for Turing
- machines. Thus, for at least a certain class of CA, there can be no
- decision procedure determining whether they go into a cycle or not.
- (perhaps for weaker CA as well?)
-
- 2) Whether a configuration is garden of eden (as demonstrated by previous
- poster)
-
- 3) Whether given state A will evolve to given state B... as before, this
- could not be decidable for we would have a decision procedure for
- halting or non halting.
-
-
- What other things are not decidable for CA?
-
-
-
-