home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!usc!hela.iti.org!cs.widener.edu!iggy.GW.Vitalink.COM!pacbell.com!decwrl!decwrl!atha!aupair.cs.athabascau.ca!burt
- From: burt@aupair.cs.athabascau.ca (Burt Voorhees)
- Newsgroups: comp.theory.cell-automata
- Subject: Re: Undecidability of configurations/progression
- Message-ID: <burt.721344629@aupair.cs.athabascau.ca>
- Date: 9 Nov 92 21:30:29 GMT
- References: <1992Nov8.210220.29731@hellgate.utah.edu> <BxG9GE.4n4@dcs.glasgow.a
- Sender: news@cs.athabascau.ca
- Lines: 44
-
- >fraserc@dcs.glasgow.ac.uk (Campbell Fraser) writes:
-
- >>In article <1992Nov8.210220.29731@hellgate.utah.edu>, tolman%asylum.cs.utah.e
- du@cs.utah.edu (Kenneth Tolman) writes:
- >>> There are some things which must be undecidable for cellular automata...
- >>> see if you agree!
-
- >>> 2) Whether a configuration is garden of eden (as demonstrated by previous
- >>> poster)
-
- >>Surely this is. If the 'world' has length x and height y then there are 2^(x.
- y)
- >>possible states. Try every one and if none lead to the alleged garden of eden
-
- >>then it is a garden of eden. If one does then make it the starting state and
- >>then the alleged garden is reached proving that it isn't.
- >>Of course the problem is probably intractable.
-
-
- >For any given configuration, you can probably determine this. The problem is
- >to find a general algorithm that say yes for this config, and no for that
- >config, that works for any config whatsoever.
-
- >CA worlds are generally unbounded, so the world has length oo and height oo.
-
- For any finite configuration, however, there is a proceedure which will
- say whether or not it is GofE. A very computationally intensive proceedure...
- If I take a state space of half-infinite 0,1 strings then one of these
- will be Gof E if it contains any finite string which has no pre-image,
- that seems to say that it will not be GofE if it contains no such finite
- substrings, although I am no longer entirely certain about that.
- For looking at finite strings and determining whether or not they have
- pre-images, however, there is a graph theoretic based proceedure. I've
- had a paper sitting at Physica D for over a year now which describes this
- (they must be having some kind of editorial problems there, they used
- to give reports rather quickly...). Anybody who is interested send me
- a snail mail address and I'll forward a pre-print.
- Burton Voorhees
- Faculty of Science
- Athabasca University
- Box 10,000
- Athabasca, AB
- CANADA T0G 2R0
- burt@aupair.cs.athabascau.ca
-