7
$\begingroup$

I think it doesn't exists but I can't find out how to prove it; note that $n$ is integer.

I know that there are infinite many integers $a_n$ s.t. $\sin(a_n)>0$ and then limsup is $+ \infty$.

I want to prove that liminf $\ne +\infty$ but I don't know how.

$\endgroup$
5
  • $\begingroup$ How efficiently can you approximate $\pi$ with rationals? $\endgroup$
    – anomaly
    May 25, 2017 at 18:31
  • $\begingroup$ I know that there are sequences of integers s.t. $|p_m - \pi q_m| < 1/q_m$ $\endgroup$
    – Bremen000
    May 25, 2017 at 18:44
  • $\begingroup$ You want to show the $\lim\sup$ is different than $\lim\inf$. Think about subsequences of the naturals where $\sin$ is very close to $1$, and where $\sin$ is very close to $-1$. $\endgroup$
    – Steve D
    May 25, 2017 at 19:18
  • $\begingroup$ @SteveD he knows that, the actual proof is the difficult part. $\endgroup$
    – Mickey
    May 25, 2017 at 19:21
  • $\begingroup$ @Mickey: The fact "he knows that" was not evident from the question at the time of my comment. $\endgroup$
    – Steve D
    May 25, 2017 at 19:43

2 Answers 2

7
$\begingroup$

The limit doesn't exist. We are going to prove it the hard way.

  1. $\limsup\limits_{n\to\infty}n(\sin n +1 ) = +\infty$

    If one make a plot of $\max( \sin(x),\sin(x+3) )$ over $[0,2\pi]$, one will notice it is bounded from below by $-2\cos(1.5) \approx -0.0707372$. For any $n$, at least one of $\sin n + 1$ or $\sin(n+3) + 1$ will be greater than $0.9$. As a result $$\limsup_{n\to\infty} n(\sin n+1) \ge \limsup_{n\to\infty} 0.9n = \infty$$

  2. $\liminf_\limits{n\to\infty} n(\sin n+1) = 0$.

    We will use a theorem by Minkowski,

    Given any $\theta \notin \mathbb{Q}$ and $\alpha \notin \mathbb{Z}$ such that $x - \theta y - \alpha = 0$ has no integer solutions.
    For any $\epsilon > 0$, there exists infinitely many pairs of integers $p,q$ such that $$|q(p - \theta q - \alpha)| < \frac14\quad\text{ and }\quad |p - \theta q - \alpha| < \epsilon$$

    Apply this to $\theta = \frac{1}{2\pi}$ and $\alpha = -\frac34$, there are infinitely many pairs of non-zero $p,q$ such that $$\left|q - 2\pi\left(p+\frac34\right)\right| < \frac{\pi}{2|q|}$$ If $q$ is positive, let $n = q, m = p$. If $q$ is negative, let $n = -3q$ and $m = -3(p+1)$.
    Above inequality reduces to $$\left|n - 2\pi\left(m + \frac34\right)\right| < \frac{\pi}{2n} \times \begin{cases}1, & q > 0\\ 9, & q < 0\end{cases}$$ Together with another easy to verify inequality, $$0 \le \sin\left(\frac{3\pi}{2}+\theta\right) + 1 \le \frac{\theta^2}{2}$$ we obtain infinitely many positive $n$ such that $$0 \le n(\sin(n) + 1) \le \frac{81\pi^2}{8n}$$ This implies $$\liminf_{n\to\infty} n(\sin n+1) \le \liminf_{n\to\infty} \frac{81\pi^2}{8n} = 0$$ Since that limit inferior has to be non-negative, it has to be zero.

Combine $1)$ and $2)$, we can conclude $\;\lim_{n\to\infty}n(\sin n + 1)\;$ doesn't exist.

Notes

About this particular theorem (Minkowski has tons of theorems), Minkowski first proved it around $1901$ using some geometrical methods he invented. Nowadays, those methods are known as Geometry of numbers. Look at any book under this subject, you should find a proof of that.

I learned this stuff from a book The Geometry of Numbers by Olds, Lax and Davidoff's. It is easy to read and has a proof at Chapter 10. Another proof can be found at Chapter 2 of Ivan Niven's classic Diophantine Approximations.

$\endgroup$
12
  • $\begingroup$ Thank you! Do you know where I can find a proof of Minkowsky theorem? $\endgroup$
    – Bremen000
    May 25, 2017 at 20:36
  • $\begingroup$ @Bremen000 I updated my answer with some refs. $\endgroup$ May 25, 2017 at 21:15
  • $\begingroup$ @achillehui Can you take a look at my answer and see whether it is true or not ? $\endgroup$
    – S.H.W
    May 25, 2017 at 21:17
  • 1
    $\begingroup$ The sequence $1,2,3,\dots$ lands in each of $(0,\pi), (2\pi,3\pi), (4\pi,5\pi),\dots,$ which gives $\sin n > 0$ for infinitely many $n.$ $\endgroup$
    – zhw.
    May 26, 2017 at 15:21
  • 1
    $\begingroup$ @Bremen000 No, we can't. We don't know what is the value of $\mu(\pi)$ ( the irrationality measure for $\pi$ ). We just know $2 \le \mu(\pi) < 7.6063$. If we know the value of $\mu(\pi)$, depends on whether $p$ is less than or greater than $2(\mu(\pi)-1)$, it might be possible to show $\liminf_{n\to\infty} n^p(1+\sin(n)) = 0$ or $+\infty$ using a similar technique.... $\endgroup$ May 30, 2017 at 19:47
0
$\begingroup$

consider $$u_n=v_n.w_n $$ with $\lim v_n=+\infty $ and $\lim w_n $ doesn't exist, so to compute $\lim u_n $, we get an ideterminate form.

$$\{\sin (n),n\in\mathbb N\} $$ is dense in $[-1,1] $ .

$\endgroup$
3
  • $\begingroup$ You haven't disproved that the limit is $\infty.$ $\endgroup$
    – zhw.
    May 25, 2017 at 18:55
  • $\begingroup$ @zhw. Ok. i will edit it now. $\endgroup$ May 25, 2017 at 18:56
  • $\begingroup$ I already knew that $\lim_{n \to \infty} \sin(n)$ doesn't exists but i do not understand how "we conclude..." $\endgroup$
    – Bremen000
    May 25, 2017 at 19:11

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.