Proof of Divergence of Series with Non-Negative Real Numbers

  • Thread starter Thread starter zeebo17
  • Start date Start date
  • Tags Tags
    Divergence Series
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
12 replies · 2K views
zeebo17
Messages
40
Reaction score
0

Homework Statement


Suppose [tex](a_n)[/tex] is a sequence of non-negative real numbers such that the series [tex]{\sum_{n=1}}^\infty a_n[/tex] diverges. Prove that the series [tex]{\sum_{n=1}}^\infty \frac{a_n}{1+a_n}[/tex] must also diverge.

Homework Equations


The Attempt at a Solution



I was thinking about looking at [tex]l=limsup(a_n)[/tex] and perhaps the requirements on it in the root test in order to see if that could tell me something about the [tex]l=limsup \left( \frac{a_n}{1+a_n} \right)[/tex], but I haven't had much luck.

Any suggestions?
Thanks!
 
Physics news on Phys.org
Might it help to write

[tex]\frac{a_n}{a_n + 1} = \frac{a_n + 1 - 1}{a_n + 1} = 1 - \frac{1}{a_n + 1}[/tex]
?
 
It's kind of a backdoor route, but I'd prove the contrapositive and use a simple comparison test.
 
Since [tex]a_n[/tex] is a non negative sequence can we assume that it diverges to [tex]+ \infty[/tex] so then [tex]1 - \frac{1}{a_n + 1}[/tex] diverges to [tex]- \infty[/tex]?
 
You're not correct with the limit of the fraction part, but that part isn't really important. What is [itex]\sum_{n = 1}^\infty 1[/itex]?
 
Oo oops! It's the lim (1 - 0) = 1. But then wouldn't the sequence then converge to 1 rather than diverging?
 
It doesn't converge to 1.
[tex]\sum_{n = 1}^\infty 1 = ~?[/tex]

What do you get when you add an infinite number of 1's?
 
Wow, tonight is not my night. haha

[tex] \sum_{n = 1}^\infty 1 = + \infty[/tex]

Great! Thanks!
 
[tex]\sum_{n=1}^\infty \left(1 - \frac{1}{a_n + 1}\right) = \sum_{n=1}^\infty 1 - \sum_{n=1}^\infty \frac{1}{a_n + 1}[/tex] should diverge whether or not an goes to infinity, no? If it doesn't go to infinity, shouldn't that make the series diverge "more"?
 
Except that if an is small, both series are wildly divergent and you can't make that split in the first place
 
hmmm. Is there another way I should approach this problem?
 
You can assume that [tex]\frac{a_n}{a_n + 1}[/tex] goes to 0 for n going to infinity, because if it didn't then the series would diverge anyway. [tex]\frac{a_n}{1 + a_n} = 1 - \frac{1}{1 + a_n}[/tex]. If [tex]\lim_{n \rightarrow \infty} 1 - \frac{1}{1 + a_n} = 0[/tex] then we see that [tex]a_n \rightarrow 0[/tex] as well.

For large enough n, [tex](a_n)^2 < a_n \Rightarrow (a_n)^2 + a_n < 2a_n[/tex].

Play around with that inequality to get a lower bound for [tex]\frac{a_n}{1 + a_n}[/tex] so that the series for that lower bound diverges.