Series Convergence: Is 1/k(ln(k)) Converging or Diverging?

  • Thread starter Thread starter ndnbolla
  • Start date Start date
  • Tags Tags
    Convergence
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
1 reply · 4K views
ndnbolla
Messages
19
Reaction score
0
The series is from k=2 to infinity.

the term is 1/k(ln(k)).

What I did was change the term to an integral... int(1/k(ln(k)) and solved it.

I came up with ln(ln(k)).

If I plug in infinty, then ln(ln(k)) goes to infinity so that means it diverges right.

From the original graph of 1/k(ln(k)), it seems like it converges but it doesn't.

Am I doing something wrong in my math? The original question asks if the series converges absolutely, conditionally, or diverges.

The positive term (to check for absolute convergence) is already the whole term itself right?
 
Physics news on Phys.org
First an obvious point, 1/kln(k) is positive for all k>1 so if the series converges, it converges absolutely, not conditionally.

[itex]\Sum \frac{1}{n}[/itex] diverges, [itex]\Sum \frac{1}{k^2}[/itex] converges and [itex]\frac{1}{k^2} \le \frac{1}{k ln(k)} \le \frac{1}{k}[/itex] so that (or looking at the graph) doesn't tell us anything!

To integrate [itex]\int \frac{1}{x ln(x)}dx[/itex] let u= ln(x) (as I presume you did). Then [itex]du= \frac{1}{x}dx[/itex] so the integral becomes [itex]\int \frac{1}{u}du[/itex]= ln(u)= ln(ln(x)) as you say.
That does not converge as x goes to infinity and so the series does not converge.