[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference rusure::math

Title:Mathematics at DEC
Moderator:RUSURE::EDP
Created:Mon Feb 03 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2083
Total number of notes:14613

1759.0. "College Mathematics Journal #505" by RUSURE::EDP (Always mount a scratch monkey.) Wed May 26 1993 19:40

    Proposed by Parviz Khajeh-Khalili, Christopher Newport College, Newport
    News, VA.
    
    Let D[n] = ln(n) - 2 * [ 1/3 + 1/5 + ... + 1/(2n-1)] for n >= 2.
    
    a) Show that D[n] is an increasing sequence.
    b) Find limit(D[n], n->infinity).
T.RTitleUserPersonal
Name
DateLines
1759.1One solution.TROOA::RITCHEFrom the desk of Allen Ritche...Thu Jun 03 1993 11:2866
I am in this notes conference only about twice a year due to lack of time
(not interest).  Note 1759.0 caught my attention.

My answer for (b) is ...



   D ==>   2 - C - ln(4)   =  .036489973...




One solution for .0  ...



(a) Show D[n] is increasing by taking the first derivative of function D of a
real variable n.  

Thus D'[n] = 1/n + 4/[(2n-1)^2] which is positive for all real n>=1.

Therefore D[n] must be an increasing sequence for integer n>=2



(b) To find lim D[n] as n==> infinity...

Let H(n) be the harmonic series 1 + 1/2 + 1/3 + 1/4 + ... + 1/n

It is known that for large n, and specifically in the limit

    H(n) =  ln(n) - C, where C is Euler's constant

[My reference for H and C above is Knuth, Volume I, Fundamental Algorithms]
(rather than C, usual symbol is lower case gamma = .5772156649... )


The series S = [ 1/3 + 1/5 + ...  1/(2n-1) ]  found in .0 can be shown to 
be related to H(n) as follows by manipulation of terms:

    2*S = 2*H(2n-1) - H(n-1) - 2


Thus D[n] = ln(n) - 2H(2n-1) + H(n-1) + 2

          = ln(n) - 2 ln (2n-1) - 2C + ln (n-1) + C + 2

                n(n-1)
          = ln ---------  - C + 2
               (2n-1)^2


lim D[n]  = ln (1/4)      - C + 2

          = 2 - C - ln(4)

          = .036489973...


Interesting result!  As a check this compares within 4-5 decimal places
with a series calculation of approx D[100]=.0364842  and D[1000]=.0364919


Allen


1759.2AUSSIE::GARSONnouveau pauvreThu Jun 03 1993 12:267
    re .1
    
    Neat. You should get in here more often.
    
    If you're married, it is recommended that you get a mistress, or at
    least that's what it says in one of the Maths jokes (probably topic 893
    or 984) as a solution to the problem of insufficient time to do Maths.