Let $a_n=1+\frac{1}{2^2}+\dots+\frac{1}{n^2}$. Considering known that $\lim_{n\to\infty}a_n=\frac{\pi^2}{6}$, evaluate $$\lim_{n\to\infty}n\left( a_n-\frac{\pi^2}{6}\right)$$ My attempt: I first proved that $b_n=n\left( a_n-\frac{\pi^2}{6}\right)$ is decreasing , i.e. $b_{n+1}-b_n=a_n+\frac{1}{n+1}-\frac{\pi^2}{6} \leq 0$, which is true since $$\lim_{n\to\infty}\left(a_n+\frac{1}{n+1}\right)=\frac{\pi^2}{6}$$ and $a_n+\frac{1}{n+1}$ is an increasing sequence.

Then I proved that $b_n$ is bounded by $0$ (obviously) and $-1$, the latter being true since it is equivalent to $a_n+\frac{1}{n}\geq \frac{\pi^2}{6}$, which can be proven as above.

So $b_n$ is both decreasing and bounded, which means that it has a limit $l$ and I ended with Stolz-Cesaro: $$l=\lim_{n\to\infty}n\left( a_n-\frac{\pi^2}{6}\right)=\lim_{n\to\infty}\frac{n^2\left( a_n-\frac{\pi^2}{6}\right)}{n}=\lim_{n\to\infty}\left( (n+1)^2\left( a_{n+1}-\frac{\pi^2}{6}\right) -n^2\left( a_n-\frac{\pi^2}{6}\right)\right)=\lim_{n\to\infty}\left( \frac{n^2}{(n+1)^2}+a_{n+1}-\frac{\pi^2}{6}+2n\left( a_{n+1}-\frac{\pi^2}{6}\right)\right)=1+0+2l$$ so $l=-1$.

Can anyone provide a shorter solution, if there is one, please?


Note that




Therefore, by the squeeze theorem, the limit is $-1$.

If one wishes to avoid integrals, creative telescoping may be employed:


I leave the telescoping process to the reader. $\ddot\smile$


You may like this. By the Stolz-Cesaro Theorem, one has \begin{eqnarray} &&\lim_{n\to\infty}n\left( a_n-\frac{\pi^2}{6}\right)\\ &=&\lim_{n\to\infty}\frac{ a_n-\frac{\pi^2}{6}}{\frac1n}\\ &=&\lim_{n\to\infty}\frac{(a_{n+1}-\frac{\pi^2}{6})- (a_n-\frac{\pi^2}{6})}{\frac1{n+1}-\frac1n}\\ &=&\lim_{n\to\infty}\frac{a_{n+1}-a_n}{-\frac1{n(n+1)}}\\ &=&\lim_{n\to\infty}\frac{\frac1{(n+1)^2}}{-\frac1{n(n+1)}}\\ &=&-1. \end{eqnarray}

  • $\begingroup$ I do like this (+1) $\endgroup$ – Simply Beautiful Art Sep 6 '17 at 20:21
  • 1
    $\begingroup$ But in order to apply Stolz-Cesaro, don't we need the sequence on the denomitor to be monotonic and unbounded? $\frac{1}{n}$ is bounded. $\endgroup$ – Shroud Sep 6 '17 at 20:26
  • $\begingroup$ D: Good catch @Shroud $\endgroup$ – Simply Beautiful Art Sep 6 '17 at 20:29
  • 2
    $\begingroup$ I haven't used Stolz-Cesàro many times, so I've learned something new today! =D $\endgroup$ – Simply Beautiful Art Sep 6 '17 at 20:36
  • 2
    $\begingroup$ @SimplyBeautifulArt : the proof for $0/0$ case is not in wiki but luckily MSE has it: math.stackexchange.com/a/631160/72031 $\endgroup$ – Paramanand Singh Sep 6 '17 at 22:52

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.