My proof: The theorem I am trying to apply is, "continous function on a compact set is uniformly continous". Since $\mathbb{R}$ is not compact, we cannot directly apply the above theorem. But we notice that since sin$^2x$ is periodic all I need is the compact domain $[n\pi,(n+2)\pi]$ because by varying $n$, I can cover $\mathbb{R}$. So I do the following, if given $\mid x - y \mid < \delta$, then I fit in $x,y \in [n\pi, n+2)\pi]$ for some $n$. Since $f$ is uniformly continuos in this interval, I will have my $\epsilon$.

Example of my explanation: Let $x = 1.81\pi$ and $= 1.9\pi$, then $\mid x - y\mid < 0.1\pi$ and $x,y \in [0, 2\pi]$, so we can find an $\epsilon$ because $f$ is uniformly continuos in $[0, 2\pi]$. The only way this process could break is when $x = 1.95\pi$ and $y = 2.04\pi$, then $\mid x - y \mid < 0.1\pi$ but $x, y \not\in [0, 2\pi]$ so we cannot find the appropriate $\epsilon$. But here I make a clever change of domain to $[\pi, 3\pi]$ from $[0,2\pi]$. In $[\pi, 3\pi]$ I can use the previous reasoning to find the $\epsilon$.

Is my proof correct?

  • $\begingroup$ In the second part you are confusing the roles of $\delta$ and $\varepsilon$. $\endgroup$ – mucciolo Jan 18 '18 at 21:49
  • $\begingroup$ You still have to show the $\delta$ is independent of $n$. $\endgroup$ – Daniel Schepler Jan 18 '18 at 21:49
  • $\begingroup$ @DanielSchepler: That would follow because sin$^2x$ is periodic, won't it $\endgroup$ – ManishKumar Singh Jan 18 '18 at 21:52
  • $\begingroup$ Yes, but it's still something that would need to be mentioned in the proof. $\endgroup$ – Daniel Schepler Jan 18 '18 at 22:13

The independence of $\delta$ on location ($n$ in your paradigm) is a consequence of periodicity, but requires some additional argument. You can fix your proof as follows.

As you observed, we have uniform continuity on the compact interval $[0,4\pi].$ Thus, for any $\epsilon > 0$ there exists $\delta >0$ such that $|\sin^2 x - \sin^2 y| < \epsilon$ for all $x,y \in [0,4\pi]$ when $|x-y| < \delta$.

For all $x,y \in \mathbb{R}$, if $|x-y| < \delta' = \min(\delta,2\pi)$ then $x,y \in [z, z+2 \pi]$ for some $z \in \mathbb{R}$. But $z = 2m\pi + \alpha$ for some integer $m$ and $0 \leqslant \alpha < 2 \pi$. Hence,

$$|\sin^2 x - \sin^2 y| = |\sin^2(2m\pi+ \alpha + x' )- \sin^2(2m\pi + \alpha + y')| \\= |\sin^2( \alpha +x' )- \sin^2( \alpha +y')| $$

where $x' = x-z$, $y' = y-z,$ and $x',y' \in [0,2 \pi]$

Since $|\alpha + x' - (\alpha+y')| = |x-y| < \delta$ and $\alpha +x', \alpha + y' \in [0,4\pi]$, we have

$$|\sin^2 x - \sin^2 y| = |\sin^2( \alpha +x' )- \sin^2( \alpha +y')| < \epsilon$$

for all $x,y$ such that $|x- y| < \delta' =\min(\delta,2\pi)$.

Since $\delta'$ does not depend upon $x$ and $y$, we have uniform continuity.


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.