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: hurd@math.gatech.EDU (Lyman Hurd)
- Subject: Undecidability of configurations/progression
- Message-ID: <9211110145.AA23882@math.gatech.edu>
- Sender: root@athena.mit.edu (Wizard A. Root)
- Organization: The Internet
- References: <1992Nov8.210220.29731@hellgate.utah.edu>
- Distribution: inet
- Date: Wed, 11 Nov 1992 01:45:00 GMT
- Lines: 85
-
- There are two distinct categories of questions about CA, yielding four
- kinds of decisions.
-
- 1) There is a distinction between finite (quiescent outside a bounded
- region, and infinite configurations.
-
- 2) There is a distinction between 1 and 2 dimensions. I am not aware of
- any problems which get harder as one passes from 2 to 3 but I would be
- interested to hear of any!
-
- To the best of my knowledge here is what is known.
-
- 1) FINITE - 1 or 2 DIMENSIONAL
-
- In one dimension there is a procedure to decide whether a configuration has
- a predecessor. If there exists one, it can be shown that there is a
- predecessor which is eventually periodic. Furthermore you can decide if a
- finite configuration has a finite predecessor. This result is (to my
- knowledge) due to Golze. Proofs are not difficult.
-
- In 2 dimensions you cannot even determine if the all-quiescent (as finite
- as they come) configuration has a predecessor. This follows immediately
- from Berger's theorem which states that there is no general way to decide
- if a given set of Wang tiles (square tiles with a preferred orientation and
- four colored edges---tilings must have touching edges have the same colors)
- admits a plane tiling. This type of construction appears many places,
- including Golze, Kari, Culik, Yu, and me.
-
- 2. Decaying into a cycle. From a finite configuration (in 1-d) it was
- shown by Culik and Yu. The infinite case and in particular if all infinite
- configurations decay into the quiescent state was solved by Kari.
-
- A consequence of this result are what Kari and I call aperiodic CA which
- have the property that all spatially periodic configurations evolve to a
- quiescent state, although NOT in uniform time, but not all configurations
- do so. Also the only temporally periodic state is quiescent. Such rules
- can be made to have highly non-trivial dynamics completely supported off
- the periodic points.
-
- 3. Injectivity and surjectivity. In one dimension there are standard
- constructions for showing if a CA is onto its image (i.e., has NO gardens
- of Eden). If it is injective (F(x)=F(y) implies x=y) then it must be
- reversible (bijective---this is not obvious), and once again in 1-d a
- finite procedure starting with a rule table will answer this question.
-
- In 2D Kari in his thesis showed that both questions are undecidable---i.e.,
- in 2D we cannot even tell if ANY GOE's exist!
-
- One interesting corollary is the following. Given two rules (in 1 or 2
- dimensions) one can decide whether they are inverses. In one dimensions if
- you know the number of states per site and the radius of the rule (speed of
- light) one can get an effective bound on the radius of the inverse and find
- it by exhaustion (I am not suggesting this practically). This bound has
- cryptographic implications. In 2D Kari shows such a recursive bound cannot
- exist. There must be 2D rules whose description in one dimension are
- arbitrarily simpler than their description in the other.
-
- My own prejudice lies in describing sets of configurations of dynamical
- interest (closed and shift-invariant). It turns out that any such set is
- determined by knowing all finite configurations lying within in. In 1-d
- this forms a language.
-
- The distinction between 1 and 2D is again vast. In 1-d one way to specify
- a set of configurations is to exclude a finite set of substrings. The
- resultant sequences form what is known as a subshift of finite type and has
- a regular language description as all bi-infinite walks through a specific
- labeled directed graph. In 2D the undecidability of the tiling problem
- says that given a finite set of excluded blocks, one cannot decide whether
- the set described is non-empty.
-
- In short I want to reiterate:
-
- 1) The questions depend highly on whether one restricts to finite
- configurations.
-
- 2) There is a dramatic difference between 1 and 2 dimensions. Ironically
- the proofs of undecidability are usually much easier in two dimensions.
-
- My own advice for the future is the exploration of reversible 1-D CA rules.
- Almost all of the results I know of are in the irreversible case. For
- example I do not know the answer to the question:
-
- Given a reversible rule F, is it decidable whether there exists N such that
- F^N(x)=x for all (infinite) configurations x.
-
-