2

What's the range of $a$ when $\displaystyle \sum_{n=1}^\infty \frac{(\log_e n)^{2012}}{n^a} $ is convergent?

Ratio test doesn't work. By root test I think a>0 is enough, but the answer is a>1.

Mark Viola
  • 179,405
Hank
  • 668

1 Answers1

2

PRIMER:

In THIS ANSWER, I showed using only the limit definition of the exponential function and Bernoulli's Inequality that the logarithm function satisfies the inequalities

$$\bbox[5px,border:2px solid #C0A000]{\frac{x-1}{x}\le \log(x)\le x-1<x}\tag{P1}$$

for $x>0$. Now, using $\log(x^b)=b\log(x)$, then we see from $(P1)$ that for $b>0$,

$$\log(x)\le \frac{x^b-1}{b}<\frac{x^b}{b}\tag {P2}$$


Note from $(P2)$, with $x=n$, that $\log(n)\le \frac{n^b-1}{b}<\frac{n^b}{b}$ for all $b>0$. Therefore,

$$\frac{\log^{2012}(n)}{n^a}<\frac{n^{2012b}}{bn^a} \tag 1$$

The inequality in $(1)$ is valid for all $b>0$ including those values of $b$ such that $a-2012b>1$. Therefore, taking $b<\frac{a-1}{2012}$, we see that the series converges by the "$p$ test" for $a>1$.


Just as a numerical example, suppose $a=1.001$. Then, we can take $b<\frac{1}{2,012,000}$, say $b=\frac{1}{4,012,000}$, and thus

$$\frac{\log^{2012}(n)}{n^{1.001}}<\left(\frac{4,012,000}{n^{1.0005}}\right)$$

And the series $\sum_{n=1}^\infty\left(\frac{4,012,000}{n^{1.0005}}\right)$ converges.


Now, if $a=1$ the series diverges. We can use the integral test. Simply write

$$\int_1^L \frac{\log^{2012}(x)}{x}\,dx=\frac{\log^{2013}(L)}{2013}\to \infty$$

The integral test shows that the series diverges for $a=1$

If $a<1$, then the series diverges by comparison with the harmonic series since for $a<1$

$$\frac{\log^{2012}(n)}{n^a}\ge \frac{1}{n}$$


Putting it all together, we see that the series $\sum_{n=1}^\infty \frac{\log^{2012}(n)}{n^a}$ converges for all $a>1$ and diverges for all $a\le 1$.

Mark Viola
  • 179,405
  • It may be worth noting that in essence, the behaviour of $\frac{\log^b(n)}{n^a}$ is the same for every $b$ i.e. it depends only on $a$. Only if $a = 1$ then $b$ may make the sum converge if $b \le -1$. – Ant Mar 31 '17 at 15:37
  • @Ant The series $\sum_{n=2}^\infty \frac{\log^b(n)}{n}$ converges for $b<-1$, but not for $b=-1$. – Mark Viola Mar 31 '17 at 16:15
  • Well, thanks for your answer. But $\displaystyle \log(n)\le \frac{n^b-1}{b}<\frac{n^b}{b}$ took me some time to prove and I don't think I can think of this inequality if this problem shows up on a timed exam. Is it a fact that I should remember? I still wonder why the root test doesn't work. By root test we can get $\displaystyle \lim_{n \to \infty } \frac{2012}{a} \times \frac {1}{n^{\frac{a}{2012}}} =0$ for a>0 – Hank Apr 01 '17 at 01:49
  • Hank, I've added a primer that explains the origin of the inequality in question. I would try and remember $\log(x)\le x-1$ and $\log(x^b)=b\log(x)$ since these are quite important and often used. -Mark – Mark Viola Apr 01 '17 at 03:30
  • The root test doesn't work here since $$\limsup_{n\to \infty}|a_n|^{1/n}=\limsup_{n\to \infty}\left|\frac{\log^{2012}(n)}{n^a}\right|^{1/n}=\limsup_{n\to \infty}\left|\frac{\log^{2012/n}(n)}{n^{a/n}}\right|=1$$ – Mark Viola Apr 01 '17 at 03:33
  • Sorry, I made a mistake in using the root test. Thanks a lot! It's very kind of you to give such a thorough answer. – Hank Apr 02 '17 at 02:13
  • @hank Hank, you're welcome. It's my pleasure! – Mark Viola Apr 02 '17 at 03:00