I recently read an article which contains the following facts :---

Let $\{a_n\}$ be a sequence of positive number such that $\sum_{n=1}^{\infty} a_n$ diverges, so we must have $a_n\sim \frac{1}{n^p}$ where $p\leq 1$. Here $\sim$ means nearby, for instance $\frac{k}{n^p}$ for some constant $k$.

Another fact is that , let $\sum_{n=1}^{\infty} b_n$ be a series of positive terms and $\sum_{n=1}^{\infty} b_n$ is convergent, that means $a_n=O(\frac{1}{n^p})$ for some $p>1$, where $O$ means order.

I cannot understand how do I prove these facts rigoursly, even I don't know whether they are true or not.


Either you are missing some hypotheses and/or context, or the article is blatantly wrong.

Take for instance $a_n=2^{-n}$ if $n\neq 2^{p^2}$ for any integer $p$ and $a_{2^{p^2}}=\frac{1}{p}$. The sum of $a_n$ is divergent, but there is no real number $a>0$ such that $a_n=O(n^{-a})$ or $n^{-a} = O(a_n)$ (however $a_n$ does go to $0$).

Take now the same sequence, except that $a_{2^{p^2}}=p^{-2}$. Then there is no real number $a>0$ such that $a_n=O(n^{-a})$ but the sum of the $a_n$ converges.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.