1

Consider the following sequence $a_n=\dfrac{\ln n}{n^p}$ where $p>1$. I want to check for the convergence of the series of that sequence i.e. $\displaystyle \sum_{n=1}^{\infty}a_n$.

I intuitively feel that the series is indeed convergent. Here is my idea - Since $p>1$ there must be some $\epsilon$ such that $p>\epsilon>1$. Then, $a_n=\dfrac{\ln n}{n^\epsilon \cdot n^{p-\epsilon}}$ and since polynomial grows faster than logarithm, there must be some $N_0\in \mathbb{N}$ such that $\ln n<n^{p-\epsilon}$ $\forall$ $N_0<n\in\mathbb{N}$. Then for all $n>N_0$, $a_n<\dfrac{1}{n^\epsilon}$, and since the series of the sequence $\dfrac{1}{n^p}$ converges for all $p>1$, $a_n$ must also converge (as $\epsilon>1$).

If the above proof is correct, I would like to know a more formal proof (using comparison tests or limit tests or fundamental properties like cauchy criterion/monotone convergence theorem). If the above proof is incorrect, Please tell me where I went wrong.

Bumblebee
  • 1,209
  • I bet this series has been discussed in detail on this website in the past. I'd encourage you to search for a duplicate. – Gerry Myerson Jan 01 '22 at 03:12
  • Your proof is fine. Note though that usually $\epsilon$ is reserved for a small quantity, so writing $\epsilon > 1$ is a little jarring. Cheers – Jair Taylor Jan 01 '22 at 03:23
  • @GerryMyerson All right, I shall search for it, but I haven't found anything yet, please do let me know if you find any. – PhysicsWizardUd Jan 01 '22 at 04:39
  • @JairTaylor Maybe I should have put it $1+\epsilon$ everywhere in place of $\epsilon$ to avoid confusion? – PhysicsWizardUd Jan 01 '22 at 04:40
  • 1
    @PhysicsWizardUd Sure that's fine. Or you could write $p = p- \epsilon + \epsilon$ with $\epsilon >0$ such that $p-\epsilon > 1$. – Jair Taylor Jan 01 '22 at 04:54
  • 1
    https://math.stackexchange.com/questions/130864/showing-that-sum-frac-log-nnx-converges-for-x1 and https://math.stackexchange.com/questions/1747012/does-sum-infty-n-1-frac-lnnn1-1-converge-or-diverge and https://math.stackexchange.com/questions/1115135/what-is-a-quick-way-to-establish-that-sum-n-1-infty-frac-log-nn3-2 and https://math.stackexchange.com/questions/2652329/series-with-log-and-exponential-sum-n-1-infty-frac-log-nna and https://math.stackexchange.com/questions/1229126/finding-what-p-does-the-series-sum-n-1-infty-frac-ln-nnp-converg and others might be helpful. – Gerry Myerson Jan 01 '22 at 14:26
  • Had a chance to look at those references, Physics? – Gerry Myerson Jan 03 '22 at 19:31
  • Hey, Physics, wake up! Please engage with my comments! – Gerry Myerson Jan 07 '22 at 16:24
  • 1
    I am so sorry @GerryMyerson, I totally forgot about this as I was busy with my exams. Yes, your references were useful, particularly the integral covergence test in the second reference in the comment on January 1. Thank you very much. – PhysicsWizardUd Jan 11 '22 at 16:32

0 Answers0