Suppose I have a sequence of positive integers $\{a_n\}$. Let us denote $b_n=\max_{1\le i\le n} a_i$. Suppose $$\frac{b_n}{\sum\limits_{i=1}^n a_i} \to 0$$ then show that $$\frac{b_n^2}{\sum\limits_{i=1}^n a_i^2} \to 0$$

I am not sure if it is true. But I didn't find any Counterexample. I was trying to get a reasonable lower bound for the denominator. I could not find any. Bounds like $$\sum_{i=1}^n a_i^2 \ge \sum_{i=1}^n a_i$$ won't help though. Note that the converse is true. As you can easily get an upper bound using: $$\sum_{i=1} a_i^2 \le b_n\sum_{i=1} a_i$$

Any help/suggestions?

Edit: Note that $a_n$'s are positive integers, that's why $\sum a_i^2 \ge \sum a_i$ is true.

  • $\begingroup$ Perhaps $a_n=\frac{1}{n}$? $\endgroup$
    – Hypernova
    Nov 28 '19 at 16:42
  • 1
    $\begingroup$ @Hypernova a_n is a sequence of positive integers. $\endgroup$
    – Sayan
    Nov 28 '19 at 16:43
  • $\begingroup$ Interesting. If $a_n$ grows like a polynomial in $n$ then both limits are $0$; if $a_n$ grows exponentially in $n$ then both limits are positive. $\endgroup$ Nov 28 '19 at 16:44

$a_n = \frac1n$, $b_n = 1$, $\frac{b_n}{\sum_{i = 1}^n a_i} \rightarrow 0$, $\frac{b_n^2}{\sum_{i = 1}^n a_i^2} \rightarrow\frac{6}{\pi^2}$.

Above was my answer when I didn't notice the requirement that $a_n$ are positive integers.

Then a counter example which works with this restriction:

if $n = 3^k$ for some integer $k$, then $a_n = 2^k$; otherwise $a_n = 1$.

The verification is kind of straightforword.

  • $\begingroup$ Ah I see! I was too occupied with straight forward polynomial type sequences. It makes sense. Thanks a lot! $\endgroup$
    – Sayan
    Nov 28 '19 at 16:59

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.