In order to do that I want to find two subsequences $a_n$ and $b_n$ of $\sin (n) $ such that $\lim_{n \to +\infty} a_n $ and $\lim_{n \to +\infty} b_n$ are not the same number. Hints?
-
3Try to think around $k\pi.$ – Balloon Oct 20 '15 at 20:22
-
5http://math.stackexchange.com/questions/238997/prove-the-divergence-of-the-sequence-sinn-n-1-infty – Oct 20 '15 at 20:27
4 Answers
Well if it approached some limit $L$, then $\sin^2 n$ approaches $L^2$ and $\cos^2 n$ approaches $1 - L^2$. For any integer $a$, we use the identity $$\sin(n + a) = \sin n \cos a + \cos n \sin a$$ This gives $$(\sin(n + a) - \sin n \cos a)^2 = \cos^2 n \sin^2 a$$ Now take the limit as $n$ goes to infinity, and you get $$(L - L \cos a)^2 = (1 - L^2) \sin^2 a$$ Equivalently, $${L^2 \over 1 - L^2} = \bigg({\sin a \over 1 - \cos a}\bigg)^2$$ (Take reciprocals of this if $L^2 = 1$).
However, ${\displaystyle \bigg({\sin a \over 1 - \cos a}\bigg)^2}$ is different for $a = 1$ and $2$ for example. Hence the limit cannot exist.

- 44,950
-
As more or less a novice I don't quite follow the the move from step(3) to step(4) (Before and after "Equivalently"). Would you mind spelling it out in a bit more detail? What trig identities are involved? – RTF May 11 '23 at 01:52
Each unit interval center on $\pi/2 + 2k\pi$,
$$ A_k = \left[ \frac{\pi}{2} + 2k\pi - \frac 12, \frac{\pi}{2} + 2k\pi + \frac 12\right] \quad\text{ for integers } k > 0$$
contains a unique positive integer. Call the sequence of such integers $a_k$.
If $\lim_{n\to\infty} \sin(n)$ exists then $\lim_{k\to\infty} \sin(a_k)$ also exists and is equal to that limit, call it $L$. Note that $L$ is positive as
$$L \geq \min_{k,x} \left\{ \sin x \ : \ x \in A_k \right\} = \min_x \left\{ \sin x \ : \ x \in \left[\frac \pi 2- \frac 12,\frac \pi 2+ \frac 12 \right] \right\} = \sin\left(\frac \pi 2- \frac 12\right) > 0$$
Now if we construct another sequence for integers in unit intervals centered on $3\pi/2 + 2k\pi$ we can bound $L < 0$. Contradiction and therefore no such $L$ exists.

- 26,524
A limit such as yours exists if corresponding $\limsup$ and $\liminf$ exist and are equal. For $\sin(x)$ as $x$ goes to infinity, $$ \limsup_{x \to \infty} \sin(x)=1$$ while $$ \liminf_{x \to \infty}\sin(x)=-1 .$$ Since these two limits are not equal, the original limit does not exist. When something is obvious the trick is often to return to definitions.

- 24,478

- 75
-
1See this page for some useful tips on how to write math using mathjax/latex. – Winther Oct 20 '15 at 20:37
-
4The OP appears to be thinking of the sequence ${\sin1,\sin2,\sin3,\ldots}$, with $n$ as an integer variable. – Barry Cipra Oct 20 '15 at 20:42
Hint:
1) $a_n=n\pi$, $n\in\mathbb{Z}$
2) $b_n=(2n+1)\frac{\pi}{2}$, $n\in\mathbb{Z}$
What can you say about the value of $\sin$ evaluated at each term of the sequences $a_n,b_n$?

- 1