2

Would this be the proper way to prove a limit of a function using the $\epsilon-\delta$ definition?

Use the definition of limit to show that $\lim_\limits{x\to2}x^2+4x=12$

Let $\epsilon>0$, then there exists $\delta>0$, such that $x\in \mathbb{R}$, $x\neq 2$ if

$$|x-2|<\delta\implies |x^2+4x-12|<\epsilon$$

$$|x^2+4x-12|=|x+6|\cdot|x-2|$$

Let $\delta = \frac {\epsilon}{|x+6|}$, then

$$|f(x)-12|<|x+6|\cdot \delta$$

$$|f(x)-12|<|x+6|\cdot \frac {\epsilon}{|x+6|}$$

$$|f(x)-12|<\epsilon$$

$$\therefore \lim f(x)=12$$

Michelle Drolet
  • 629
  • 5
  • 14
  • 1
    You have a beautfiul handwriting, but you should post your effort not as a picture. – Cornman Nov 27 '17 at 16:11
  • @Cornman thank you for the compliment. I just have no knowledge of programming so it would be very messy if I posted it without a picture. – Michelle Drolet Nov 27 '17 at 16:15
  • Michelle - Latex editing is fairly easy, almost next to writing it on paper. As a starter you can refer this link: Quick Latex Reference – jonsno Nov 27 '17 at 16:18
  • If you write each equation between a pair of $ symbols (or $$ before and after to get the equation in the middle of a separate line), and pay attention to what the question looks like in the review window, you can do a lot without knowing much. Another place to start learning the format is https://math.stackexchange.com/help/notation – David K Nov 27 '17 at 16:19
  • I edited it with proper formatting. Be aware, most people won't bother answering your question, unless you show that you put some effort into your post. (I.e. proper latex formatting). – Mark Pineau Nov 27 '17 at 16:19
  • @MarkPineau thank you for the edit, I really appreciate it! I guess I have to learn proper latex formatting lol... – Michelle Drolet Nov 27 '17 at 16:20
  • Now that it is edited, is this the proper way to solve these types of questions? – Michelle Drolet Nov 27 '17 at 16:21
  • Thank you! May I ask why we restrict delta less than 1? – Michelle Drolet Nov 27 '17 at 16:24
  • Note that $\delta$ should not depend on $x$. – Jason Nov 27 '17 at 16:28

2 Answers2

3

No. Its not correct. $\delta$ cannot depend on $x$. This is one of those $\delta = min \{1, \epsilon /M\}$ problems. Use the fact that if $x \in (1,3)$ then $|x+6| <9 $. In particular, if $|x-2|< 1 $ then $|x+6| < 9$. Sometimes, when you can't factor or uniformly bound a term you have to settle for bounding it near the limit point. This is perfectly fine because all that matters when taking limits is the behavior of the function near the limit point anyways. If you set $\delta = min\{1, \epsilon /9\}$ , then $$|x-2| < \delta \implies |x-2||x+6|< \epsilon$$

EDIT:

Btw, the reason that delta cannot depend on x is because you have to find ONE delta that works for EVERY x. Having delta depend on x will give you a different delta for each value of x.

EDIT 2:

In response to your comment. You're welcome. I was crucified on my first analysis test because I was absent the day he went over this. So now I really harp on people when I see them do it. I can virtually guarantee you that your text has some worked examples of this type of problem (never have come across one that doesn't), and also that a question like this will be on the exam. Another way to see this is to pay attention to order of quantifiers.

$$(\forall \epsilon > 0)(\exists \delta > 0)(\forall x \in I)\big(0 < |x-a| < \delta \implies |f(x) - L| < \epsilon\big)$$

$ \\ $

Your approach would require the order to be $(\forall \epsilon > 0)(\forall x \in I)(\exists \delta > 0)$

Getting familiar with quantifiers will help you a lot in this class in both understanding subtle distinctions in definitions (like uniform continuity vs continuity) and in proofs by contradiction.

For example, how would one prove a limit doesn't exist? It's easy to see by using quantifier negation:

$$ \begin{align} \sim (\forall \epsilon > 0)(\exists \delta > 0)(\forall x \in I)\big(0 < |x-a| < \delta \implies |f(x) - L| < \epsilon\big) \\ (\exists \epsilon> 0)\sim(\exists \delta > 0)(\forall x \in I)\big(0 < |x-a| < \delta \implies |f(x) - L| < \epsilon\big) \\ (\exists \epsilon > 0)(\forall \delta > 0)\sim(\forall x \in I)\big(0 < |x-a| < \delta \implies |f(x) - L| < \epsilon\big) \\ (\exists \epsilon > 0)(\forall \delta > 0)(\exists x \in I)\sim\big(0 < |x-a| < \delta \implies |f(x) - L| < \epsilon\big) \\ (\exists \epsilon > 0)(\forall \delta > 0)(\exists x \in I)\big(0 < |x-a| < \delta \, \land \, |f(x) - L| \geq \epsilon\big) \\ \end{align} $$

Thus to show a limit doesn't exist you have to find an $\epsilon$ such that for every $\delta$, one can find an $x$ with the following property

$$|x-a| < \delta \ \text{ but } \ |f(x) - L| \geq \epsilon$$

David Reed
  • 3,265
  • @MichelleDrolet yw. I added some more to the post as well. – David Reed Nov 27 '17 at 17:30
  • My text definitely does have some examples like these and I am positive there will be at least one question similar to this on my final. That's why I want to make sure I'm doing it correctly. So thank you for your help and attention to detail. It really is very helpful! – Michelle Drolet Nov 27 '17 at 17:59
0

After this step $|f(x)−12|<|x+6|⋅δ$, you want to try to set a bound on $|x+6|$.

We know $|x-2|<\delta$.

$\implies$ $-\delta<x-2<\delta$.

$\implies$ $8-\delta<x+6<8+\delta$.

We also know for obvious reasons, $-8-\delta<8-\delta$.

$\implies -8-\delta<x+6<8+\delta$

$\implies |x+6|<\delta+8$.

This lets us say, $|f(x)−12|<|x+6|⋅δ<(\delta+8)⋅\delta$

We can choose $\delta$ to be whatever we want, to avoid dealing with $\delta^2$, fix $\delta\leq1$.

Now $(\delta+8)⋅\delta\leq9\delta$

So pick $\delta=min\{1,\frac{\epsilon}{9}\}$