0
$\begingroup$

I once saw the notation $O(x) \to 0$ or $o(x) \to 0$ for $x\to 0$ but I don't really know why that makes sense since $O(x)$ is more like a class of functions. So I would interpret this as "if $f(x) = O(x)$, then $f(x) \to 0$ for $x\to 0$". Does this always hold? Can I conclude from $$\lim_{x\to 0}\frac{f(x)}{x} = 0 \implies \lim_{x \to 0}f(x) = 0$$ ?

EDIT: I was referring to the first answer of this post.

2 Answers 2

1

$\displaystyle \lim_{x\to 0}\frac{f(x)}{x} = 0 \implies \lim_{x \to 0}f(x) = 0$ is certainly true since

if $\left|\frac{f(x)}{x}\right| \lt \epsilon$ for all $x$ with $0 \lt |x| \lt \delta$

then $\left|{f(x)}\right| \lt |x| \epsilon \lt \epsilon$ for all $x$ with $0 \lt |x| \lt \min(1,\delta)$

so you will be arbitrarily close to $0$ as $x \to 0$

  • 0
    Thanks. So whenever I have a term like $f(x) + g(x) + o(x)$ and $f(x) \to 0$ and $g(x) \to 0$ for $x\to 0$ it will mean the whole term tends to $0$ as $x\to 0$?2017-01-09
  • 0
    My problem if the I usually see $o(x)$ as describing the behaviour as $x \to \infty$ as infinite asymptotics. But if you are clear in the context that it has meaning as infinitesimal asymptotics as $x \to 0$ then yes.2017-01-09
  • 0
    Okay, thanks! Of course you have to add the context.2017-01-09
1

EDIT: I misunderstood the question and assumed the $O$ referred to big-O notation. The correct interpretation is not this one; therefore my answer is rendered irrelevant.

As you noted, $O(x)$ is a class of functions therefore

$$f(x) = O(x)$$ makes no sense assuming you meant that $f(x)$ is a function.

One can write, however, $f(x) \in O(x)$.

Remember that one writes $f(x) \in O(g(x))$ if, when $x \to \infty$, the function $f$ is bounded above by a scalar multiple of $g$. Note that the definition says nothing about the limit of $x \to 0$. In particular, one has that

$$f(x) = x + 1 \in O(x)$$

and $f(x)$ tends to $1$ as $x \to 0$.

  • 0
    I edited my post with what I was referring to. The answer said $O(x) \to 0$ as $x\to 0$, so would you say that this is wrong?2017-01-09
  • 0
    @lappen68 sorry, I happened to take your $O(x)$ as big-O notation when it is not. Therefore my answer is **not** right.2017-01-09
  • 0
    Also, your result is correct as long as we are talking about $O(x)$. $x + 1 \notin o(x)$ though.2017-01-09
  • 0
    @lappen68 I know; sorry for the confusion.2017-01-09
  • 1
    Nah, It was my mistake, should've made it more clear.2017-01-09