home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!cs.utexas.edu!sdd.hp.com!ux1.cso.uiuc.edu!news.cso.uiuc.edu!levine
- From: levine@symcom.math.uiuc.edu (Lenore Levine)
- Subject: More About the Collatz Function
- Message-ID: <BuH7Mt.IC@news.cso.uiuc.edu>
- Sender: usenet@news.cso.uiuc.edu (Net Noise owner)
- Organization: University of Illinois at Urbana
- Date: Sat, 12 Sep 1992 17:37:39 GMT
- Lines: 40
-
- I was messing around with the Collatz function this weekend.
-
- My idea is that you can prove interesting things about it if
- you can prove something else about another function.
-
- I call it the smallest nontrivial backward iterate (SNBI) function.
-
- You notice, in the Collatz function, every odd number that is not
- divisible by three has nontrivial odd preimages. (The preimages of
- an odd number divisible by three are just an infinite path).
-
- What I want to do is get from an odd number not divisible by three,
- via the smallest path possible to another odd number not divisible
- by three.
-
- That is, if n is an odd number not divisible by three, the following
- algorithm obtains SNBI(n):
-
- (Written in Pascal.)
-
- repeat
- n := n*2;
- until (n mod 3 = 1) and (n mod 9 <> 1);
- snbi := (n - 1) div 3;
-
- From SNBI(n) I get another function, r(n). r(n) is a real number
- between 0 and 1, determined as follows:
-
- The first digit after the decimal is SNBI(n) mod 9. The second digit
- after the decimal is the second iterate of SNBI of n mod 9 (that is,
- SNBI(SNBI(n)) mod 9). Continuing, the kth digit after the decimal is
- the kth iterate of SNBI of n mod 9 (SNBI^k(n) mod 9). Note all
- digits of r(n) will be 1, 2, 4, 5, 7 or 8.
-
- My idea is that if you can show r(n) is not rational for all
- n for which SNBI is defined, you may have more insight into the
- Collatz function.
-
- How to proceed: Maybe showing that *any* sequence of the digits used
- will pop up if SNBI is repeated on n enough times...
-