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!

Homework Help: Proof of Inequality Between Lower and Upper Bounds

  1. Jul 10, 2014 #1
    Convergence of Divergent Series Whose Sequence Has a Limit

    1. The problem statement, all variables and given/known data
    Suppose [itex]∑a_{n}[/itex] is a series with lim [itex]a_{n}[/itex] = L ≠ 0. Obviously this diverges since L ≠ 0. Suppose we make the new series, [itex]∑(a_{n} - L)[/itex]. My question is this: is there some sufficient condition we could put solely on [itex]a_{n}[/itex] (or maybe its sum) without calculating the limit or referencing it other than the existence of a limit so that the new series converges?

    2. Relevant equations
    I can't think of anything here; I could list all of the convergence tests from Ch. 3 of Baby Rudin but doesn't seem wholly relevant with the way I phrased the question. Feels like this is probably something trivial from some higher level analysis course or analytic number theory course that I haven't had the chance to take yet.

    There's one thing that seems possibly relevant (from Rudin):
    Theorem Suppose [itex]a_{1}≥a_{2}≥a_{3}≥...≥0.[/itex] Then the series [itex]∑a_{n}[/itex] converges if and only if the series [itex]∑2^{k}a_{2^{k}}[/itex] converges.

    3. The attempt at a solution
    I really don't know how to approach this without introducing the L somehow into the proof which isn't what I really wish to explore.

    I thought to make some simple conditions first to try and reduce the nature of the problem, like making the sequence monotonic and non-negative. With the theorem, I was thinking of maybe introducing some sort of scaling maybe with the [itex]2^{k}[/itex] factor? I don't know.

    I think the root test gives convergence for 0 < L < 1, but what about for L > 1? But this sort of references the limit.
    Last edited: Jul 10, 2014
  2. jcsd
  3. Jul 10, 2014 #2
    That's a pretty vague question. Is this an actual book problem or a question you invented yourself? I don't see any reason why there would be such a simple condition. Well, obviously you can try all the series tests you've seen and adapt it to this case. But I don't really know what kind of answer you want here.
  4. Jul 10, 2014 #3
    Invented. Just curious if there exists any such condition to give convergence. I don't know anything about techniques to deal with divergent series, so I'd be satisfied with any specific reference material in lieu of some answer.
  5. Jul 10, 2014 #4
  6. Jul 10, 2014 #5
    Alright, thanks. Messed up the title of this thread but think I corrected it. Guess I can't fix the thread title, ah well.

    Criterion of Abel looks promising maybe:
    The criterion of Abel
    Let ∑+∞n=0an be a (real or complex) convergent series . Let (vn)n be a bounded sequence of real numbers which is either nondecreasing or nonincreasing. Then the sequence ∑+∞n=0vnan converges.

    [itex]v_{n}[/itex] could be the [itex]a_{n}[/itex] and would have to figure out some convergent sequence so that the product equals (L - [itex]a_{n}[/itex]).
    Last edited: Jul 10, 2014
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted