home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!tcsi.com!iat.holonet.net!news.cerf.net!usc!howland.reston.ans.net!news.ans.net!newsgate.watson.ibm.com!yktnews!admin!yktnews!victor
- From: victor@watson.ibm.com (Victor Miller)
- Subject: Re: A strange recurrence and RESULT ABOUT PI
- Sender: news@watson.ibm.com (NNTP News Poster)
- Message-ID: <VICTOR.93Jan26142615@terse.watson.ibm.com>
- In-Reply-To: plouffe@greco-prog.fr's message of 26 Jan 93 00:12:56 GMT
- Date: Tue, 26 Jan 1993 19:26:15 GMT
- Lines: 79
- Reply-To: victor@watson.ibm.com
- Disclaimer: This posting represents the poster's views, not necessarily those of IBM
- References: <1993Jan26.001256.12160@greco-prog.fr>
- Nntp-Posting-Host: terse.watson.ibm.com
- Organization: IBM, T.J. Watson Research Center
-
- >>>>> On 26 Jan 93 00:12:56 GMT, plouffe@greco-prog.fr (Simon Plouffe [melancon]) said:
-
-
- Simon> Let a(1)=1/2 and iterate the following recurrence :
-
- Simon> a(n)
- Simon> a(n + 1) = 2 ---------
- Simon> 2
- Simon> 1 - a(n)
- Simon> This gives the numbers
-
- Simon> 336 354144
- Simon> 1/2, 4/3, -24/7, ---, ------, ....
- Simon> 527 164833
-
- Simon> Now define this sign function to be,
-
- Simon> sign(x) = 0 if x <= 0
- Simon> 1 if x > 0
-
- Simon> If you take the sum
-
- Simon> infinity
- Simon> -----
- Simon> \ sign(a(i))
- Simon> v = ) ----------
- Simon> / i
- Simon> ----- 2
- Simon> i = 1
-
- Simon> This real number appears to be
-
- Simon> arctan(1/2)
- Simon> v = 1 - -----------
- Simon> Pi
-
-
- Simon> The questions are :
-
- Simon> 0) Is this known ?
- Simon> 1) Can we explain this fact ?
-
- This isn't that hard to explain: Suppose \theta is real, then tan
- \theta is >=0 if and only if {2 \theta/\pi} is >=0 and < 1/2, where
- {x} denotes the fractional part of x. This condition occurs exactly
- when the expansion of the fractional part of 2\theta/\pi in binary
- starts .0.... Thus, you see, that sign(a(i)) is 1-d(i+1) where d(i)
- is 0 or 1 and 2 \theta/\pi = \sum_i d(i) 2^{-i} (assuming that a(i) is
- not zero). Here \theta = arctan(1/2).
- is irrational (this makes matters slightly easier).
- Simon> 2) Are the distribution of binary digits of v random ?
- By Weyl's criteria, the binary digits of should be random (uniform).
-
- Simon> 3) Can we find another formula that will give the decimal digits of 1/Pi ?
- Simon> 4) Can formula be used to calculate v efficiently ?
- Simon> Note , for an angle A we have :
-
-
- Simon> tan(A)
- Simon> 2 ----------- = tan(2 A).
- Simon> 2
- Simon> 1 - tan(A)
-
-
-
- Simon> Any comment, references are welcome.
-
-
- Simon> Simon Plouffe.
-
- Simon> e-mail: plouffe@lacim.uqam.ca
- Simon> presently in Bordeaux at plouffe@geocub.greco-prog.fr
-
- Simon> Ps : the real number was found experimentally.
- --
- Victor S. Miller
- Vnet and Bitnet: VICTOR at WATSON
- Internet: victor@watson.ibm.com
- IBM, TJ Watson Research Center
-