prove or disprove $$\lim_{n\to\infty}\dfrac{\displaystyle\sum_{k=1}^{n}|\sin{(x+k)}|}{n}=\dfrac{1}{\pi}\int_{0}^{\pi}|\sin{x}|dx$$

if this is right,and How prove it? Thank you

My idea maybe this can use Uniform distribution? and I don't prove it.

  • $\begingroup$ I have tried,and I can't true this right,and this problem is find limit $\lim_{n\to\infty}\dfrac{\sum_{k=1}^{n}|\sin{(x+k)}}{n}$! $\endgroup$ – math110 Jun 4 '14 at 3:16
  • $\begingroup$ I believe you need a factor of $ \frac{1}{\pi}$ on the RHS. $\endgroup$ – Calvin Lin Jun 4 '14 at 3:24

Lemma 1. Let $\theta$ be a real number which is not a rational multiple of $\pi$. For $m\in\mathbb{Z}$, we define $$I_n(m)=\frac{1}{n}\sum_{k=1}^ne^{ikm\theta}$$ Then, $$\lim_{n\to\infty}I_n(m)=\left\{\matrix{0&\hbox{if}&m\ne0\cr1&\hbox{if}&m=0}\right. $$

Proof. Indeed, for $m\ne0$ we have $$\left\vert I_n(m)\right\vert=\frac{1}{n}\left\vert\frac{e^{im(n+1)\theta}-e^{im\theta}}{e^{im\theta}-1}\right\vert\leq\frac{1}{n\vert \sin(m\theta/2)\vert} $$ and the conclusion follows since $\theta\notin\pi\mathbb{Q}$.$\qquad\square$

Lemma 2. Let $\theta$ be a real number with $\theta\notin\pi\mathbb{Q}$. Let $P$ be a trigonometric polynomial, $P(t)=\sum_{m=-r}^r c_me^{imt}$, and define $$J_n(P,t)=\frac{1}{n}\sum_{k=1}^nP(t+k\theta)$$ Then, $$\lim_{n\to\infty}J_n(P,t)=c_0=\frac{1}{2\pi}\int_0^{2\pi}P(x)dx $$

Proof. In fact $$ J_n(P,t)=\sum_{m=-r}^{r}c_me^{imt}\left(\frac{1}{n}\sum_{k=1}^ne^{imk\theta}\right) =\sum_{m=-r}^{r}c_me^{imt}I_n(m) $$ and the conclusion follows by letting $n$ tend to infinity. $\qquad\square$

Proposition 3. Let $\theta$ be a real number with $\theta\notin\pi\mathbb{Q}$. Let $f$ be a continuous $2\pi$-periodic function, and define $$J_n(f,t)=\frac{1}{n}\sum_{k=1}^nf(t+k\theta)$$ Then, $$\lim_{n\to\infty}J_n(f,t)= \frac{1}{2\pi}\int_0^{2\pi}f(x)dx $$

Proof. Indeed, for $\epsilon>0$ there is a trigonometric polynomial $P_\epsilon$ such that $$ \Vert f-P_\epsilon\Vert_\infty=\sup_{\mathbb{R}}|f(t)-P_\epsilon(t)|<\frac{\epsilon}{3} $$ It follows that $$ \vert J_n(f,t)-J_n(P_\epsilon,t)\vert\leq\Vert f-P_\epsilon\Vert_\infty<\frac{\epsilon}{3}\tag{1} $$ and also $$ \left\vert \frac{1}{2\pi}\int_0^{2\pi}f(x)dx-\frac{1}{2\pi}\int_0^{2\pi}P_\epsilon(x)dx\right\vert\leq\Vert f-P_\epsilon\Vert_\infty<\frac{\epsilon}{3}\tag{2} $$ Finally, using Lemma 2. there is $N_\epsilon$ such that for $n>N_\epsilon$ we have $$ \left\vert J_n(P_\epsilon,t)-\frac{1}{2\pi}\int_0^{2\pi}P_\epsilon(x)dx\right\vert<\frac{\epsilon}{3}\tag{3} $$ Combining $(1)$, $(2)$ and $(3)$ we obtain $$ \forall\,n>N_\epsilon,\quad\left\vert J_n(f_\epsilon,t)-\frac{1}{2\pi}\int_0^{2\pi}f(x)dx\right\vert<\epsilon. $$ and the Proposition follows.$\qquad\square$

Application. Taking $f(x)=|\sin(x)|$, and $\theta=1$ we see that $$ \lim_{n\to\infty}\frac{1}{n}\sum_{k=1}^n|\sin(t+k)|= \frac{1}{2\pi}\int_0^{2\pi}|\sin(x)|dx.$$

  • $\begingroup$ It's Nice! Thank you very much!+1 $\endgroup$ – math110 Jun 4 '14 at 11:03
  • $\begingroup$ Now this is an answer. =) +1 $\endgroup$ – Pedro Tamaroff Jun 4 '14 at 21:59
  • $\begingroup$ It's curious the result is $x$-independent. Is that a kind of Ergodic Theorem ?. $\endgroup$ – Felix Marin Aug 9 '14 at 21:57

Claim: If the continuous function $f(x)$ is periodic with period $T$ which is irrational, then

$$ \lim_{n\rightarrow \infty} \frac{ \sum_{k=1}^{n} f(t + k)}{n} = \frac{1}{T} \int_0^T f(x) \, dx$$

Hint: The RHS calculates the average value of $f(x)$ (over 1 period). Show that the LHS is equivalent to (approximating) the average value of $f(x)$.

Hint: This is an application of the pigeonhole argument which shows that for any $ \epsilon > 0$, there exists an integer $n$ such that $n T$ is within $ \epsilon$ of an integer.

  • $\begingroup$ Why the down vote? $\endgroup$ – Calvin Lin Jun 4 '14 at 3:17
  • $\begingroup$ Hello,I think you hint is not usefull? because I have think it this hint,and I can't prove it? can you post your full solution? $\endgroup$ – math110 Jun 4 '14 at 3:18
  • 3
    $\begingroup$ You'd have to prove the LHS is independent of the choice of $t$, I guess. $\endgroup$ – Pedro Tamaroff Jun 4 '14 at 3:27
  • $\begingroup$ @PedroTamaroff Yes, that is true, and that's why I intentionally chose the variable $t$ instead of $x$. $\endgroup$ – Calvin Lin Jun 4 '14 at 3:32
  • 5
    $\begingroup$ @math110 I'm not hard up for up votes, nor points. Thank you very much. $\endgroup$ – Calvin Lin Jun 4 '14 at 4:04

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.