Good evening everyone,

I'd like to discuss with you the following exercise :

$\sum\limits_{n=1}^{\infty} (-1)^{n} \frac{n^{2} +3n - \sin(n)}{n^{4}-\arctan(n^{2})}$

I can prove that $\lim\limits_{x \to \infty} a_{n} = 0$ , where $a_{n} = \frac{n^{2} +3n - \sin(n)}{n^{4}-\arctan(n^{2})}$

But I can't still proove its convergence, I'd have used Leibnitz alternating series test (due to $(-1)^{n}$), but I was unable to say $a_{n+1} \leq a_{n}$.

Maybe I could study the Absolute convergence and then by Comparison test find that converges ?

Any help would be appreciated,

Thanks anyway.

  • 1
    $\begingroup$ $-\pi\leq \arctan(x)\leq \pi$ and $-1\leq \sin(x)\leq 1$ for all $x$. Compare how your series acts then to $\sum\dfrac{n^2}{n^4}$ $\endgroup$ – JMoravitz Jun 29 '18 at 20:05
  • $\begingroup$ @JMoravitz this works ? $\endgroup$ – jacopoburelli Jun 29 '18 at 20:07
  • $\begingroup$ With an appropriately phrased argument, yes. $\endgroup$ – JMoravitz Jun 29 '18 at 20:08

No need for Leibniz: For $n>1,$

$$\left |(-1)^{n} \frac{n^{2} +3n - \sin n}{n^{4}-\arctan(n^{2})}\right| \le \frac{n^{2} +3n +1}{n^{4}-\pi/2}.$$

This looks good, yes? It appears that the terms on the right are like $1/n^2.$ Verify that the terms on the right divided by $1/n^2$ approach $1.$ Since $\sum 1/n^2<\infty,$ the limit comparsion test shows the sum of the terms on the right converges. Hence our series converges absolutely, and therefore converges. (Moral: Don't blindly assume Leibniz is the way to go just because it's got the $(-1)^n$ thing happening.)


Hint/Walkthrough: Since $\sin(n)$ and $\arctan(n^2)$ are both easily bounded, while $n^2+3n$ and $n^4$ are not, one can prove that your expression $a_n$ is definitely monotonically decreasing after a certain point. Then you can use the Alternating Series test.


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.