[Proof-verification] Determining whether the function $f(x)=\sqrt{x^2-x-6} \text{ in } x_0=3$; $x_0\in \mathbb{R}$ is continous $\color{red}{\text{ in }x_0}$ or not with the $\varepsilon$-$\delta$-definition of limit/criterion:

$$f(x)=\sqrt{x^2-x-6} \text{ in } x_0=3$$

Let $\varepsilon>0$ and ${\mid x-x_0\mid}<\delta \iff {\mid x-3\mid}<\delta$

\begin{align} {\mid f(x)-f(x_0)\mid}&= {\mid\sqrt{x^2-x-6}-\sqrt{3^2-3-6}\mid}\\ & ={\mid\sqrt{x^2-x-6}\mid}\\ & ={\mid\sqrt{x+2} \cdot \sqrt{x-3}\mid}\\ & <\sqrt{x+2}\cdot \sqrt{x+2}\\ & = x+2\\ & < x-3+5 \\ & <\delta+5 =: \varepsilon \\ & \iff \delta = \varepsilon -5 \end{align}

$\implies$ the function is continous in $x_0=3 \qquad \qquad \qquad \qquad \qquad_\blacksquare $

Is this proof correct?

  • $\begingroup$ So, if $\varepsilon = 1$, what's the $\delta > 0$? $\endgroup$ – Ennar Nov 6 '18 at 15:54
  • $\begingroup$ @Ennar It wouldn't be positive $\endgroup$ – Doesbaddel Nov 6 '18 at 18:07
  • $\begingroup$ And that's not really good now, is it? $\endgroup$ – Ennar Nov 6 '18 at 18:23
  • $\begingroup$ No, because $\delta$ needs to be positve, right? $\endgroup$ – Doesbaddel Nov 6 '18 at 18:28
  • $\begingroup$ Exactly.$\hphantom{}$ $\endgroup$ – Ennar Nov 6 '18 at 18:35

First of all, the function $f$ is not defined on $\mathbb{R}$ but for $$ x\in(-\infty,-2]\cup[3,+\infty). $$

So one can only talk about its continuity of $f$ at $x=3$ from the right.

For $x>3$, you are right to get $$ |f(x)-f(3)|=\sqrt{x-3}\sqrt{x+2}. $$ Note that you don't need the absolute value for the square root terms.

But then you made a mistake: the $\delta$ you get must be positive.

The term $\sqrt{x-3}$ should not be dropped and it would give you the desired $\delta$.

Consider instead for $0<x-3<1$ the inequality $$ \sqrt{x-3}\sqrt{x+2}\leq 6\sqrt{x-3}\le\varepsilon. $$

  • $\begingroup$ Can you explain more in detail how you came to this conclusion and the estimation upwards? $\Big($Consider instead for $0<x-3<1$ the inequality $\sqrt{x-3}\sqrt{x+2}\leq 6\sqrt{x-3}\le\varepsilon\Big)$. $\endgroup$ – Doesbaddel Nov 6 '18 at 18:25
  • $\begingroup$ Why is $\sqrt{x-3}\sqrt{x+2}\leq 6\sqrt{x-3}$? With your hint I would come to the following conclusion: $\cdots \leq 6 \cdot \sqrt{x-3} < 6 \sqrt{ \delta }=: \varepsilon \implies \delta = \dfrac{\varepsilon^2}{36}$ Is that correct? $\endgroup$ – Doesbaddel Nov 6 '18 at 19:13
  • $\begingroup$ @Doesbaddel: If $0<x-3<1$, then $3<x<4$ and thus $\sqrt{x+2}<\sqrt{6}<6$. I just did a rough estimate. $\endgroup$ – user587192 Nov 6 '18 at 21:38
  • $\begingroup$ @Doesbaddel: you would need $\delta=\min\{1,\frac{\varepsilon^2}{36}\}$. $\endgroup$ – user587192 Nov 6 '18 at 21:39
  • $\begingroup$ Why does $x-3$ needs to be between $0$ and $1$? Where does this step come from? Why do you need $\delta=\min\{1,\frac{\varepsilon^2}{36}\}$ and not $\delta= \frac{\varepsilon^2}{36}$? (Otherwise, your estimation makes sense now and I understood it.) $\endgroup$ – Doesbaddel Nov 6 '18 at 21:53

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.