home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!stanford.edu!nntp.Stanford.EDU!ilan
- From: ilan@leland.Stanford.EDU (ilan vardi)
- Subject: Re: Large numbers
- Message-ID: <1992Dec18.205938.22759@leland.Stanford.EDU>
- Sender: news@leland.Stanford.EDU (USENET News System)
- Organization: DSG, Stanford University, CA 94305, USA
- Date: Fri, 18 Dec 92 20:59:38 GMT
- Lines: 48
-
- There is a possibly interesting game consisting of trying to name
- larger numbers than your opponent, where the rules are that you can
- operate on his number and the point is that you use few operations to
- name a bigger number. For example:
-
- A: names 5
-
- B: names 5 + 1
-
- But B could have said 2*5 and done better, or else 2^5. So in general
- you would get
-
- A: names n
-
- B: ``I successor you,'' i.e., n+1
-
- ``I double you'', i.e., 2*n.
-
- ``I exponentiate you,'' i.e., 2^n,
-
- ``I tower you,'' i.e., 2^2^2...^2, n times.
-
- In fact he could have said: ``I Ackerman you,'' which really means
- ``I diagonalize you''.
-
- So in some sense, the game can be viewed as happening on the recursive
- hierarchy with a move taking you from one limit to another. For
- example, after B says ``I diagonalize you'', A responds with ``I
- diagonalize you'', which means ``I diagonalize your diagonalization''.
- E.g., you diagonalize the using the Ackerman function, by
- diagonalizing the sequence
-
- B_1(n, k) = Ackerman(Ackerman...(Ackerman(n))) (k Ackerman's)
-
- B_2(n, k) = B_2(n, B_1(n, k-1))
-
- i.e., the function you end up with is essentially
-
- C(n) = B_n(n, n)
-
- up to some fussing about whether n -> n-1 or not.
-
- Such moves seem to be optimal, in some sense.
-
- -ilan
-
-
-
-