1
$\begingroup$

How to prove that $\sin\left(\dfrac{\pi}{n}\right)$ is decreasing, $\forall n\in\mathbb{N}$.

My original question was to determine the convergence of $\sum (-1)^{n+1}\sin\left(\dfrac{\pi}{n}\right).$

I showed that the absolute value does not converge, so it does not converge absolutely. I now need to check for conditional convergence.

I want to solve the series using the alternating series test.

I already showed that $b_n\to\ 0$, as $n\to\infty$.

Now I need to show $b_n$ decreasing. I found the derivative, which is $-\dfrac{\pi \cos \left(\frac{\pi }{n}\right)}{n^2}.$

The problem is that $\cos$ is sometimes negative, and I have a negative sign in front of the derivative, which means that the derivative is sometimes positive. So it is not decreasing for all $n\in\mathbb{N}$, but the answer says converges conditionally? How?

$\endgroup$
  • 1
    $\begingroup$ you can use the fact that $cos$ is positive in $[0, \frac \pi 2]$, and therefore for any $n \ge 2$ the derivative is indeed negative and the sequence is decreasing $\endgroup$ – AsafHaas Apr 15 '17 at 20:36
  • $\begingroup$ You have the issue that $\sin\left(\frac\pi{1}\right)=0 \not \gt 1 = \sin\left(\frac\pi{2}\right)$ $\endgroup$ – Henry Apr 15 '17 at 20:43
  • $\begingroup$ @Henry That's not an issue though. Finitely many exceptional terms do not matter for convergence. $\endgroup$ – Hagen von Eitzen Apr 15 '17 at 20:55
  • $\begingroup$ @HagenvonEitzen. You are correct: it does not matter for convergence. But it does mean that "$\sin\left(\dfrac{\pi}{n}\right)$ is decreasing, $\forall n\in\mathbb{N}$" is in fact false $\endgroup$ – Henry Apr 15 '17 at 20:57
2
$\begingroup$

One may observe that $$ x \mapsto \sin x \quad \text{is increasing over} \quad \left[0,\frac \pi2\right] $$ and that $$ x \mapsto \frac \pi x \quad \text{is decreasing over} \quad \left[1,\infty\right) $$ giving that $ \sin \circ \:\frac \pi x$ is decreasing over $\left(2,\infty\right)$.

$\endgroup$
  • $\begingroup$ What about the case $x=1$. $\endgroup$ – hamam_Abdallah Apr 15 '17 at 20:50
  • $\begingroup$ When $x=1$ we are positive, but I get what the question is answering $\endgroup$ – K Split X Apr 15 '17 at 21:08
  • $\begingroup$ $x>2$, we are decreasing* $\endgroup$ – K Split X Apr 15 '17 at 21:09
0
$\begingroup$

We have

$$(\forall n\geq 2)\; \;\; \; \frac {\pi}{n}\in (0,\frac {\pi}{2}] $$

on the other hand,

$f:x\mapsto \sin (x) $ is increasing at $(0,\frac{\pi}{2}] $.

and

$g:n\mapsto \frac{\pi}{n}$ is decreasing from $n=2$.

thus

$$n\mapsto f(g (n))=\sin(\frac {\pi}{n})$$ is decreasing from $n=2$.

$\endgroup$
  • 1
    $\begingroup$ @K Split X, For proving convergence, all one needs is to prove that $\sin(\pi/n)$ is decreasing for $n >N$ for some $N$. Thus, it's incorrect to rephrase the problem to require that $\sin(\pi/n)$ be decreasing for all $n > 0$. $\endgroup$ – Hoc Ngo Apr 15 '17 at 22:18
  • $\begingroup$ But $x\mapsto \sin (x) $ is not increasing over $(0,\pi] $ so, you need $n>1$. $\endgroup$ – hamam_Abdallah Apr 16 '17 at 11:29
  • $\begingroup$ What I meant was that K Split X stated wrongly: How to prove that $\sin\left(\frac{\pi}{n}\right)$ is decreasing, $\forall n\in\mathbb{N}$, which means to prove for $n = 1,2,3,\ldots$. As you have shown, $n>1$ is what you need. In general, you don't need to pick $N=1$. You can pick $N=10^5$. $\endgroup$ – Hoc Ngo Apr 16 '17 at 12:06

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.