1
$\begingroup$

Suppose $\lim \limits_{n \to \infty} a_n = 0$. Find the limit

$$\lim \limits_{n \to \infty} \left(1+a_n \frac{x}{n}\right)^n$$

It's kind intuitive that the answer is 1, but clearly I can't just say that the limits equal $\lim \limits_{n \to \infty} 1^n = 1$.

I feel like I should let $y =(1+a_n \frac{x}{n})^n $. Then take $log$ so that $log(y) = nlog(1+a_n \frac{x}{n})$ But L'hopital doesn't apply here either

$\endgroup$
3
  • $\begingroup$ hint :$$n\log(1+a_n \frac{x}{n}) = \dfrac{\log(1+a_n \frac{x}{n})}{1/n}$$ $\endgroup$ – AgentS Mar 18 '15 at 14:25
  • $\begingroup$ @ganeshie8 so we apply limit, and using L'hopital rule. But know do I take the derivative of a sequence? $\endgroup$ – Ellery Lai Mar 18 '15 at 14:27
  • $\begingroup$ You definitely can't say it is like $\lim_{n\to\infty} 1^n$, because that same reasoning would show that $\lim (1+x/n)^n$ is $1$, when it is $e^x$. You need that $a_n\to 0$. $\endgroup$ – Thomas Andrews Mar 18 '15 at 14:29
3
$\begingroup$

If $a_n \to 0$ then, no matter how big $x$ is, for any $n$ big enough we have $|a_n x|\leq \varepsilon$, hence: $$\left(1-\frac{\varepsilon}{n}\right)e^{-\varepsilon}\leq\left(1+\frac{a_n x}{n}\right)^n \leq e^{\varepsilon}.$$ Since $\varepsilon$ is an arbitrary positive number, the claim follows by squeezing.

$\endgroup$
5
  • 1
    $\begingroup$ Could you explain why is $$ \lim \limits_{n \to \infty} e^{\epsilon} = 0? $$ I know that $\epsilon$ can be arbitrary small positive, but why does it have anything to do with $ n \to \infty ?$ $\endgroup$ – Ellery Lai Mar 18 '15 at 14:50
  • $\begingroup$ @ElleryLai: I have not stated that. The point is the following: given any positive $\varepsilon$, the sequence given by $b_n=\left(1+\frac{a_n x}{n}\right)^n$, for any $n$ big enough, is between the LHS and the RHS of what I wrote, hence: $$\lim_{n\to +\infty} b_n = \color{red}{1}.$$ $\endgroup$ – Jack D'Aurizio Mar 18 '15 at 15:00
  • $\begingroup$ I couldn't figure out why is that $$\left(1+\frac{a_n x}{n}\right)^n \leq e^{\varepsilon}.$$ ?? because of the limit definition of limit? $\endgroup$ – Ellery Lai Mar 18 '15 at 15:30
  • $\begingroup$ @ElleryLai: because for any positive $A$, the sequence $$(1+A/n)^n $$ is an increasing sequence that converges towards $e^A$. Basic fact. $\endgroup$ – Jack D'Aurizio Mar 18 '15 at 15:33
  • $\begingroup$ Since $-\epsilon < a_{n}x < \epsilon$ the inequality should be like $$\left(1 - \frac{\epsilon}{n}\right)^{n} < \left(1 + \frac{a_{n}x}{n}\right)^{n} < \left(1 + \frac{\epsilon}{n}\right)^{n}$$ and this would lead to $$e^{-\epsilon} \leq \left(1 + \frac{a_{n}x}{n}\right)^{n}\leq e^{\epsilon}$$ for sufficiently large values of $n$. $\endgroup$ – Paramanand Singh Apr 4 '15 at 16:43
1
$\begingroup$

Hint: Supose $\lim_{n\to \infty}n\cdot\frac{1}{ x\cdot a_n}=0$. Note that $\lim_{n\to \infty} x\cdot a_n=0$, $$ \left( 1+a_n\frac{x}{n} \right)^{n} = \left( 1+\frac{1}{n\cdot\frac{1}{ x\cdot a_n}} \right)^n = \left[ \left( 1+\frac{1}{n\cdot\frac{1}{ x\cdot a_n}} \right)^{n\cdot\frac{1}{\color{red}{ x\cdot a_n}}} \right]^{\color{red}{ x\cdot a_n}} $$ and $ \lim_{n\to \infty} \left( 1+\frac{1}{\color{blue}{n\cdot\frac{1}{ x\cdot a_n}}} \right)^{\color{blue}{n\cdot\frac{1}{ x\cdot a_n}}}=e $. If $\lim_{n\to \infty}n\cdot\frac{1}{ x\cdot a_n}=0$ then $$ \lim_{n\to \infty} \left[ \left( 1+\frac{1}{n\cdot\frac{1}{ x\cdot a_n}} \right)^{n\cdot\frac{1}{\color{red}{ x\cdot a_n}}} \right]^{\color{red}{ x\cdot a_n}} =\ldots =e^0=1 $$ The case $\lim_{n\to \infty}n\cdot\frac{1}{ x\cdot a_n}=L\in\mathbb{R}$ is trivial.

$\endgroup$
4
  • $\begingroup$ why is that $ \lim \limits_{n \to \infty} \frac{n}{xa_n} = 0 ?$ $\endgroup$ – Ellery Lai Mar 18 '15 at 14:56
  • $\begingroup$ also, for you last part, you seem to use if $$ \lim \limits_{n \to \infty} f(n) = l$$ and $$ \lim \limits_{n \to \infty} g(n) = m$$, then $$\lim \limits_{n \to \infty} f(n)^{g(n)} = l^m $$. I'm not sure if this is one of the limit properties $\endgroup$ – Ellery Lai Mar 18 '15 at 16:24
  • $\begingroup$ @Ellery If $\lim_{n\to \infty}\frac{n}{x\cdot a_n}=L\neq 0$ and $L\in\mathbb{R}$ then the question is trivial. In your secont coment you are right. I hope helped you. $\endgroup$ – Elias Costa Mar 18 '15 at 16:28
  • $\begingroup$ @ElleryLai Do not intend to exhaust all cases to be analyzed. In the case of home work, is not purpose of this site provide full answers. $\endgroup$ – Elias Costa Mar 18 '15 at 16:37

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.