19
$\begingroup$

How do I prove $\sin x$ is uniformly continuous on $\mathbb R$ with delta and epsilon?

I proved geometrically that $\sin x<x$ and thus, $$|f(x_1)-f(x_2)|=|\sin x_1 - \sin x_2|\le|\sin x_1|+|\sin x_2|<|x_1|+|x_2|$$

But this doesn't help me much finding a delta...

Thanks for any help!

P.S. I'm only at the beginning of calculus so I can't use many theorems and derivation (because they haven't been regorously proven).

$\endgroup$
  • 1
    $\begingroup$ I changed $sinx_1$, etc., to $\sin x_1$. That is standard TeX usage. $\endgroup$ – Michael Hardy Jan 1 '13 at 19:35
  • $\begingroup$ This is a particular case of (at least) two more general results. First: a periodic continuous function on $\mathbb{R}$ is uniformly continuous on $\mathbb{R}$. Second: a Lipschitz function (mathworld.wolfram.com/LipschitzFunction.html) is uniformly continuous. To prove that $\sin$ is Lipschitz, you can use a trigonometric identity like Nameless did in his answer, or you can claim that its derivative is bounded by $1$. $\endgroup$ – Julien Jan 1 '13 at 19:41
23
$\begingroup$

Let $\epsilon>0$ and $x,y\in \mathbb{R}$. We want $$\left|f(x)-f(y)\right|<\epsilon\implies \left|\sin x-\sin y\right|<\epsilon\implies \left|2\cos\frac{x+y}2\sin\frac{x-y}2\right|$$ Because $$\left|2\cos\frac{x+y}2\sin\frac{x-y}2\right|\le 2\left|\sin\frac{x-y}2\right|$$ it suffices $$2\left|\sin\frac{x-y}2\right|<\epsilon$$ when $$\left|x-y\right|<\delta\implies \left|\frac{x-y}2\right|<\delta$$ SInce $\left|\sin x\right|\le \left|x\right|$, $$2\left|\sin\frac{x-y}2\right|\le 2\left|\frac{x-y}2\right|<2\delta$$

Choosing $\delta=\frac{\epsilon}{2}>0$ will do the trick. Because $\delta$ doesn't depend on $x,y$, the continuity is uniform

$\endgroup$
  • $\begingroup$ Hi, thanks for the fast response! Why $$\left|2\cos\frac{x+y}2\sin\frac{x-y}2\right|\le 2\left|\sin\frac{x-y}2\right|$$ is correct? Also, I know how to prove $sinx<x$, but how do I show $|sinx|<|x|$? Thanks for the reply again! $\endgroup$ – Harold Jan 1 '13 at 19:32
  • 1
    $\begingroup$ because $\cos \alpha \leqslant 1 $, $\forall \alpha \in \mathbb{R}$. $\endgroup$ – Amihai Zivan Jan 1 '13 at 19:34
  • 2
    $\begingroup$ Because $\left|\cos(\bullet)\right|\le 1$. $\endgroup$ – Michael Hardy Jan 1 '13 at 19:34
  • 1
    $\begingroup$ @Harold $\sin x$ is odd $\endgroup$ – Nameless Jan 1 '13 at 19:48
  • 2
    $\begingroup$ @Harold You want $\left|\sin y\right|\le \left|y\right|\iff -\sin y\le -y\iff \sin y\ge y$ for $y<0$ near $0$. But $\sin x\le x$ for $x>0$. Multipliying by $-1$ gives $-\sin x\ge -x\iff \sin (-x)\ge -x\iff \sin y\ge y$ $\endgroup$ – Nameless Jan 1 '13 at 20:10
21
$\begingroup$

By Mean Value Theorem,

$$ |\sin{x}- \sin{y}| \leq |x-y| |\cos{\xi}| \leq |x-y|, \quad x\leq\xi \leq y.$$

Hence, you may choose $\epsilon=\delta$.

$\endgroup$
  • 5
    $\begingroup$ This is a circular reasoning. For MVT you need that sine is differentiable. Since each differentiable function is continuous, you have used what you are proving. $\endgroup$ – wdacda Feb 1 '16 at 19:55
  • 2
    $\begingroup$ @wdacda How is that circular reasoning? $\endgroup$ – 3x89g2 Sep 5 '16 at 5:27
13
$\begingroup$

Since $\sin x$ is a periodic continuous function with a period $2\pi$, it suffices to prove that it is uniformly continuous on $[0, 2\pi]$. Since $[0, 2\pi]$ is compact, this follows from the well-known theorem.

$\endgroup$
3
$\begingroup$

There is an elementary geometric way to do this. Let $x$ and $y$ be real numbers; for now, assume $x, y\in(-\pi, \pi]$. Start off at $(1,0)$ and march off signed distance $x$ to get to point $a$ and $y$ to get to point $b$ on the unit circle. Then $|x - y|$ is the distance from $a$ to $b$ along the unit circle. $|\sin(x) - \sin(y)|$ is the distance between the $y$-coordiates of $a$ and $b$. Hence, in this case $$|\sin(x) - \sin(y) | \le |x - y|.$$

This gives us uniform continuity on $(-\pi, \pi]$, so by periodciity the sine function is uniformly continuous on the entire line.

$\endgroup$
  • $\begingroup$ Another way to prove this identity: math.stackexchange.com/questions/620305/… $\endgroup$ – GinKin Dec 28 '13 at 10:08
  • $\begingroup$ This answer deliberately does not appeal to the MVT; it uses purely geometric properties of the sine and cosine functions. $\endgroup$ – ncmathsadist Dec 29 '13 at 19:56
  • $\begingroup$ Yeah I know. But for a test I'm not sure if your approach would be rigorous enough. No offense. $\endgroup$ – GinKin Dec 29 '13 at 20:03
  • $\begingroup$ What is not rigorous about the fact that, given a line and a point, the shortest distance between them is achieved along the perpendicular dropped from the point to the line? In fact, this solution is unique. $\endgroup$ – ncmathsadist Dec 30 '13 at 0:46

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.