1
$\begingroup$

The formal definition of limit at infinity usually starts with a statement requiring an open interval. An example from OSU is as follows:

Limit At [Negative] Infinity: Let $f$ be a function defined on some open interval $(a, \infty)$ [$(-\infty, a)$]. Then we say the limit of $f(x)$ as $x$ approaches [negative] infinity is $L$, and we write $\lim_{x\to[-]\infty} f(x) = L$ if for every number $\epsilon$, there is a corresponding number $N$ such that $\left|f(x) - L\right| < \epsilon$ whenever $x > N\;[x < N]$.

But, I think it is also valid when the interval is half-open as in the following: Let $f$ be a function defined on some right-open interval $[a, \infty)$ [left-open interval $(-\infty, a]$]. Then we say the limit of $f(x)$ as $x$ approaches [negative] infinity is $L$, and we write $\lim_{x\to[-]\infty} f(x) = L$ if for every number $\epsilon$, there is a corresponding number $N$ such that $\left|f(x) - L\right| < \epsilon$ whenever $x \geq N\;[x \leq N]$.

So, why the formal definition of limit at infinity does not start with a statement requiring a half-open interval, which is more general?

Is it because people want to match it with the formal definition of limit? I understand that the formal definition of one-sided limit requires an open interval because that is necessary to define a limit at a point $a$. But, such requirement does not exist for limit at infinity, and therefore, why the more general version of half-open interval is not used in the formal definition of limit at infinity?

$\endgroup$
1
$\begingroup$

There is not really a difference between both approaches: If $f$ is defined on the open interval $(-\infty,a)$, then we may as well consider the restriction to the closed interval $(-\infty,a-1]$ and similarly vice versa. The reason that open interval may be preferred is that the limit requires $f$ to be defined on a topological neighbourhood of $\infty$. A neighbourhood of $\infty$ is a set that contains an open set containing $\infty$ and the basic open sets are open intervals. So the the following definition might be considered "best", but I'm afraid it is way less intuitive for the learner:

Limit At Infinity: Let $f\colon A\to\mathbb R$ be a function where $A$ is a punctured neighbourhood of $\infty$ in the two-point compactification of $\mathbb R$. Then we say ...

$\endgroup$
1
$\begingroup$

It's great to ask what hypotheses in a theorem or definition are really germane. What is important is that $f$ is defined on some interval $(a,\infty)$ for sufficiently large $a$ so that when we write down the $\epsilon-\delta$ definition, we don't have to worry about pathologies.

$\endgroup$
0
$\begingroup$

It's not more general. If your function is defined on a closed interval $[a,\infty)$ then it's also defined on an open interval $(b,\infty)$ where $b > a$. And the values of the function on the interval $[a,b]$ don't matter at all when it comes to the limit of the function at $+\infty$.

The question of whether the limit of $f(x)$ as $x\to +\infty$ exists is basically the same as asking whether $f$ can be extended from $(a,\infty)$ to $(a,+\infty]$ (see Wikipedia on the extended real line) such that it's continuous at $+\infty$.

$\endgroup$

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.