home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!julienas!corton!geocub!plouffe
- From: plouffe@greco-prog.fr (Simon Plouffe [melancon])
- Newsgroups: sci.math
- Subject: A strange recurrence and RESULT ABOUT PI
- Message-ID: <1993Jan26.001256.12160@greco-prog.fr>
- Date: 26 Jan 93 00:12:56 GMT
- Organization: GRECO Programmation du CNRS - Bordeaux,France
- Lines: 60
-
-
- Let a(1)=1/2 and iterate the following recurrence :
-
- a(n)
- a(n + 1) = 2 ---------
- 2
- 1 - a(n)
- This gives the numbers
-
- 336 354144
- 1/2, 4/3, -24/7, ---, ------, ....
- 527 164833
-
- Now define this sign function to be,
-
- sign(x) = 0 if x <= 0
- 1 if x > 0
-
- If you take the sum
-
- infinity
- -----
- \ sign(a(i))
- v = ) ----------
- / i
- ----- 2
- i = 1
-
- This real number appears to be
-
- arctan(1/2)
- v = 1 - -----------
- Pi
-
- The questions are :
-
- 0) Is this known ?
- 1) Can we explain this fact ?
- 2) Are the distribution of binary digits of v random ?
- 3) Can we find another formula that will give the decimal digits of 1/Pi ?
- 4) Can formula be used to calculate v efficiently ?
- Note , for an angle A we have :
-
-
- tan(A)
- 2 ----------- = tan(2 A).
- 2
- 1 - tan(A)
-
-
-
- Any comment, references are welcome.
-
-
- Simon Plouffe.
-
- e-mail: plouffe@lacim.uqam.ca
- presently in Bordeaux at plouffe@geocub.greco-prog.fr
-
- Ps : the real number was found experimentally.
-