home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.theory.cell-automata
- Path: sparky!uunet!snorkelwacker.mit.edu!bloom-beacon!INTERNET!dont-send-mail-to-path-lines
- From: MCINTOSH@UNAMVM1.BITNET ("Harold V. McIntosh")
- Subject: Basins of attraction.
- Message-ID: <9211181017.AA17646@Early-Bird.Think.COM>
- X-Unparsable-Date: Tue, 17 Nov 92 00: 30:46 MEX
- Sender: root@athena.mit.edu (Wizard A. Root)
- Organization: The Internet
- Distribution: inet
- Date: Wed, 18 Nov 1992 10:18:24 GMT
- Lines: 50
-
- Andrew Wuensche <100020.2727(at)COMPUSERVE.COM> (Mon, 16 Nov 1992 08:23:16
- EST) remarks:
- >
- > For finite CA with periodic boundary conditions there is a proceedure.
- > Checking if a state is a G of E follows from computing its pre-images,
- > if any. This can be done without exhaustive testing. The proceedure is
- > described for 1-D CA (and the software included) in my recently
- > published book, (see below). The G of E density in state space of a
- > given CA rule is approx predictable by the rule's Z parameter. For most
- > of rule-space with very few exceptions (the limited pre-image rules),
- > most states in state-space are G of E states.
- > If anyone out there gets around to reading the book, or
- > experimenting with the software, I would welcome comments or
- > discussion.
- >
- Well, I have ordered a copy of the book; delivery has been promised for
- January or February. In the meantime, it is only possible to conjecture
- the contents of the book, and the relationship of the procedures to those
- of E. F. Moore, G. A. Hedlund, M. Nasu, Erica Jen, or even myself. It is
- well established that for one dimensional cellular automata, Gardens of
- Eden are absolutely computable; difficulties lie in two dimensions and
- beyond. In fact, there is no need to restrict to periodic boundary
- conditions; the procedures serve as well for quiescence at infinity or
- for boundaryless (that is, unrestrictedly infinite) automata.
- -
- It is necessary to commiserate with a remark made by Burton Voorhees here
- recently about his article on this topic having disappeared into a black
- hole. Two articles of my own have suffered the same fate. Is this traffic
- entirely one-way?
- -
- In the meantime, here are a couple of questions in the way of making a
- discussion. 1) Unless the counterimages are balanced, there must be a
- Garden-of-Eden. Presumably this coincides with your 'limited pre-image
- rules' remark above, and immediately foredooms most automata to have
- Gardens of Eden.
- -
- 2) In my early calculations of the variance in the number of ancestors,
- there appeared to be a gap between the rules of zero-variance and the
- others? Have you observed such an effect? Is it real (I later decided
- that it probably wasn't)?
- -
- In any event, I look forward to seeing the book. To formulate our ideas
- about automata well, it is convenient to have compilations of data to look
- at. I have probably turned to the appendices in Wolfram's book more often
- than I have consulted any of the other parts.
- -
- Harold V. McIntosh |Depto. de Aplicaci'on de Microcomputadoras
- MCINTOSH@UNAMVM1.BITNET |Instituto de Ciencias/UAP
- mcintosh@unamvm1.dgsca.unam.mx |Apdo. Postal 461
- (+52+22)43-6330 |72000 Puebla, Pue., MEXICO
-