1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Name of a mathematical series

  1. Jul 25, 2011 #1
    Is there a name for this series:

    1/2 + 2/3 + 3/4 + 4/5 + 5/6 + 6/7 + 7/8 + 8/9 +....+ n/(n+1)

  2. jcsd
  3. Jul 25, 2011 #2
    I don't think so. Since the series diverges, I don't think many will find it interesting.
  4. Jul 25, 2011 #3
    It's equal to n + 1 - (1 + 1/2 + 1/3 + 1/4 + ... + 1/n+1). The thing in parens is the http://en.wikipedia.org/wiki/Harmonic_series_%28mathematics%29" [Broken]. Umm, actually, I think "harmonic series" refers to the infinite series of which this is the first n+1 terms.
    Last edited by a moderator: May 5, 2017
  5. Jul 25, 2011 #4
    The harmonic series diverges. I think you hurt its feelings :(

    okay I'm just feeling silly.....
  6. Jul 25, 2011 #5
    Just barely, though...
  7. Jul 25, 2011 #6
    What does barely mean? I can deleted infinitely many terms from the harmonic series and it will still diverge... I can make the terms much smaller and it will still diverge.
    Last edited: Jul 25, 2011
  8. Jul 25, 2011 #7
    You can divide it in half and it still diverges. I know that's not as cool as taking infinitely many terms from it, but then again, it kind of is the same thing...

    My prof started telling me about how you can take the terms with "9, 99, 999" or maybe with that as an exponent, or something?... and make it converge. He didn't really lay it out though, just kind of said something in passing. Was this a baseless rumor or is there something like that?
  9. Jul 25, 2011 #8
    Check http://en.wikipedia.org/wiki/Small_set_(combinatorics) :smile:
  10. Jul 25, 2011 #9
  11. Jul 25, 2011 #10
    It diverges very, very slowly!
  12. Jul 25, 2011 #11
    It divergence is in the order of log(n). While this is extremely slow for all applications, I can still easily find sequences that diverge 100000 times slower. I just want to make clear that "slow" is relative :smile:
  13. Jul 25, 2011 #12


    User Avatar
    Homework Helper

    If I increase the exponent on the series just "barely", though, it converges!

    [tex]\sum_{n=1}^\infty \frac{1}{n^{1+\epsilon}} < \infty[/tex]
    for any [itex]\epsilon > 0[/itex]! ;) (the exclamation point denotes excitement, not a factorial! =P)
  14. Jul 25, 2011 #13
    Maybe if you discover cool enough properties for the series, you can get to name it yourself. ;)
  15. Jul 26, 2011 #14
    Sure. In fact, I could find a series that diverges infinitely more slowly, and then I could find another that diverges infinitely more slowly than that, and so, on, ad infinitum:

    \sum _{k=n}^{\infty } 1 \\
    \sum _{k=n}^{\infty } \frac{1}{k} \\
    \sum _{k=n}^{\infty } \frac{1}{k \log (k)} \\
    \sum _{k=n}^{\infty } \frac{1}{k \log (k) \log (\log (k))} \\
    \sum _{k=n}^{\infty } \frac{1}{k \log (k) \log (\log (k)) \log (\log (\log (k)))} \\

    But those would be contrived series, made up just for the purpose of diverging slowly. The harmonic series is about as slowly diverging a series as you're likely to bump into, unless you go hunting for slowly diverging series.

    I also had in mind the point Mute made: considering just series with terms of the form ip, p=-1 is the edge case.
  16. Jul 26, 2011 #15
    And yet


    also diverges. So I can increase the exponent a bit, and it will still diverge!
  17. Jul 26, 2011 #16
    Those series are not contrived. I've seen them popping up in probability theory. Fine, they're useless, but they do pop up from time to time :smile:
    Last edited: Jul 26, 2011
  18. Jul 26, 2011 #17
    Really! :-)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook