2

Prove that $$\sum_{n=2}^\infty \frac{1}{(n(\log n)^p )}$$ converges if and only if $p>1$.

I know that $\sum_{n=2}^\infty f(n)$ converges if and only if $\int_2^\infty f(x)$ converges by the integral test, provided that $f$ is a positive decreasing function on $[2,\infty)$. However, for some values of $p$, $f(x)=\frac{1}{x(\log x)^p}$ is decreasing, as shows the derivative $f(x)=-\frac{1+px}{x^2(\log x)^{p-1}}$. Can we apply the integral test whatsoever or do we need to account for changes? Can we apply the integral test at all?

I have to use the integral test.

Guy Fsone
  • 25,237

3 Answers3

3

For $p>0$, you can apply the integral test since $f$ is decreasing from a certain moment because the first terms of a sum do not change its nature. And I think you know $\sum \frac{1}{n}$ diverges so you can conclude for all $p$.

2

You can use the Cauchy condensation test

$$ 0 \ \leq\ \sum_{n=1}^{\infty} f(n)\ \leq\ \sum_{n=0}^{\infty} 2^{n}f(2^{n})\ \leq\ 2\sum_{n=1}^{\infty} f(n)$$

user
  • 162,563
1

Hint: By Cauchy condensation test your series is equivalent to

$$\sum_{n=2}^{\infty} \frac{2^n}{2^n(\ln (2^n))^p}=\frac{1}{\ln^p 2}\sum_{n=2}^{\infty} \frac{1}{n^p}$$ can you conclude?

Guy Fsone
  • 25,237