If we were to prove $\lim_{x \to 4} \sqrt{x} = 2$, then by using $\epsilon$–$\delta$ definition, we have:
$0 \leq |\sqrt{x} – 2| < \epsilon$
$-\epsilon< \sqrt{x} – 2 < \epsilon$
$2 -\epsilon< \sqrt{x} < 2 + \epsilon$
$(2 -\epsilon)^{2}< x < (2 + \epsilon)^{2}$
$4 -4\epsilon +\epsilon^{2}< x < 4 +4\epsilon -\epsilon^{2}$
Now, we set $\delta \leq \min(4\epsilon +\epsilon^{2}, 4\epsilon -\epsilon^{2})$ and choose whichever is smaller.
Since $\epsilon > 0$, the minimum is $\delta \leq 4\epsilon -\epsilon^{2}$
Source: Hartman, G. (n.d.). 1.2: Epsilon-Delta Definition of a Limit. LibreTexts. Retrieved April 29, 2021, from https://math.libretexts.org/Bookshelves/Calculus/Book%3A_Calculus_(Apex)/01%3A_Limits/1.02%3A_Epsilon-Delta_Definition_of_a_Limit
Remark: I paraphrased what the author said, including only the key points, but the computations remain the same.
From the above, I have some questions, which I'll list as follows:
- The minimum, according to the source, is $\delta \leq 4\epsilon -\epsilon^{2}$. Is this because that's the only value $\delta$ that guarantees, for every $x$ within $\delta$ units from $4$, $f(x)$ is to be within $\epsilon$ units from 2?
- Does the need for minimum come from the fact that the distance between $4$ and $4 – \delta$ and the distance between $4$ and $4 + \delta$ is not the same?
- Is the distance between $4$ and $4 – \delta$ and the distance between $4$ and $4 + \delta$ not the same because the function we're evaluating is non-linear? Does this mean that the distances from $4$ are to be the same when the function is linear?
- Why can $\delta$ be made smaller than the minimum?
Please let me know if I've made myself understood as I'm not a native english speaker.
Best Answer
Because for any $\epsilon$, with that choose of $\delta$, $$ 4-\delta < x < 4+\delta \implies 2 - \epsilon < f(x) < 2 + \epsilon $$ And this satifies(actually, this is) the $\epsilon-\delta$ definition. So your statement is basically correct, the only problem is "the only value $\delta$ that satifies". Fix $\epsilon$, then with smaller choose of $\delta$, let's say some $\delta'$ < $\delta$, $$ 4-\delta' < x < 4+\delta' \implies 4-\delta < x < 4+\delta \implies 2 - \epsilon < f(x) < 2 + \epsilon $$ So in fact, any smaller choose of $\delta$ satifies the inequality.
I think you mean the distance between f(4) and f(4−δ) copmared to f(4) with f(4+δ). Yes. Because you need different choose of $\delta$ to satify both sides of the inequality, and you have to make the both sides correct to prove it satifies the definition.
I think you mean the distance between $f(4)$ and $f(4-δ)$ copmared to $f(4)$ with $f(4+δ)$. Yes, because we're evaluating a non-linear function. When the function is linear, it's clear that $$ f(4) - f(4-\delta) = f(4+\delta) - f(4)$$
Look 1.