sci.math
[Top] [All Lists]

Re: A formula of Pi; how is it derived?

Subject: Re: A formula of Pi; how is it derived?
From: Rob Johnson
Date: Sun, 30 Apr 2006 17:18:31 GMT
Newsgroups: sci.math
In article <e32geu$jak$1@xxxxxxxxxxxxxxxxxxxxxxxx>,
"Tsukamoto Tenma" <1@xxx> wrote:
>I have read somewhere that
>Pi/2=3/2*5/6*7/6*11/10*13/14*17/18*...
>where the nominators are prime numbers, with corresponding denominator its
>closest integer in the form 4n+2. I'm looking for a proof of this, any help
>appreciated.

First prove a similar infinite product where the numerators are the
odd prime numbers and the denominators are the closest integers that
are 0 mod 4:

           3 5 7 11 13 17 19
    pi/4 = - - - -- -- -- -- ...                             [1]
           4 4 8 12 12 16 20

Define a multiplicative function a(n) on the integers by defining its
values on the primes.  Let a(2) = 0.  Let a(p) = 1 for primes which
are 1 mod 4 and a(p) = -1 for primes which are 3 mod 4.

The product in [1] is then equal to

    ---     1
    | | ----------
     p  1 - a(p)/p

      --- a(n)
    = >   ----                                               [2]
      ---  n
       n

It can pretty easily be shown that a(2n) = 0 and a(2n+1) = (-1)^n.
Therefore, the sum in [2] is equal to the Gregory series for pi/4:

    1   1   1   1   1
    - - - + - - - + - - ...                                  [3]
    1   3   5   7   9

This proves [1].  The product you cite above is equal to

    ---     1
    | | ----------                                           [4]
     p  1 + a(p)/p

The product of [2] and [4] is then

    ---     1
    | | ---------
    p>2 1 - 1/p^2

      3 ---     1
    = - | | ---------
      4  p  1 - 1/p^2

      3 ---  1
    = - >   ---
      4 --- n^2
         n

    = pi^2/8                                                 [5]

Thus, [4] must be (pi^2/8)/(pi/4) = pi/2

Rob Johnson <rob@xxxxxxxxxxxxxx>
take out the trash before replying
to view any ASCII art, display article in a monospaced font

<Prev in Thread] Current Thread [Next in Thread>
Privacy Policy