8
$\begingroup$

Prove the following limit without using approximations and derivatives: $$\lim_\limits{x\to -\infty}{\frac{\ln\left(1+3^{x}\right)}{\ln\left(1+2^{x}\right)}}=0$$

I cannot think of any possible factorization or inequality (so that I could use the Squeeze Theorem) that doesn't use derivatives so as to find this limit. Any hint?

$\endgroup$
8
$\begingroup$

HINT:

$$\frac{z}{z+1} \le \log (1+z)\le z \tag 1$$

for $z>-1$. Then, use the squeeze theorem.

SPOILER ALERT: Scroll over the highlighted region to reveal the solution.

Using $(1)$, we have $$\frac{3^x}{2^x(1+3^x)}<\frac{\log (1+3^x)}{\log (1+2^x)}<\frac{3^x(1+2^x)}{2^x}\tag 2$$As $x\to -\infty$ both the left-hand and right-hand sides of $(2)$ approach $0$. Therefore, by the squeeze theorem, the limit of the sequence of interest goes to $0$.


NOTE 1:

We can prove the inequalities in $(1)$ using standard tools, which don't require the use of derivatives. We need only to note that

$$e^x\ge 1+x \tag 2$$

For $x>-1$, we take the log of both sides to reveal

$$\log(1+x)\le x$$

which is the right-hand side inequality of $(1)$. Likewise, if we make the substitution $x=-z/(z+1)$ in $(2)$, we obtain

$$e^{-z/(1+z)}\ge 1-\frac z{z+1}=\frac1{z+1}$$

whereupon rearranging yields for $z>-1$

$$\log(1+z)\ge \frac{z}{z+1}$$

as was to be shown.


NOTE 2:

We can prove the inequality given in $(2)$ using the limit definition of $e^x$

$$e^x=\lim_{n\to \infty}\left(1+\frac xn\right)^n$$

I showed in THIS ANSWER, without the use of derivatives, that $\left(1+\frac xn\right)^n$ is a monotonically increasing function of $n$ (If $x<0$, this is true for $n>|x|$). Then, we have

$$e^x\ge \left(1+\frac xn\right)^n\ge 1+x$$

where this last inequality follows from Bernoulli's inequality. And we are done!

$\endgroup$
  • $\begingroup$ (1) is not known by theory we have until now, so I cannot use it without proving it. But it is proven only with the use of derivatives, so I cannot "officially" prove it... $\endgroup$ – Jason Sep 22 '15 at 19:51
  • $\begingroup$ One does not need differential or integral calculus to prove those inequalities. $\endgroup$ – Mark Viola Sep 22 '15 at 20:16
  • $\begingroup$ I lost you, after "an increasing function of n". How do you conclude the following using monotony? $$e^x\geq \left(1+\frac{x}{n}\right) ^n$$ $\endgroup$ – Jason Sep 23 '15 at 17:44
  • 1
    $\begingroup$ @Jason You're welcome! My pleasure. And thank you for the "best vote!" Much appreciative. $\endgroup$ – Mark Viola Sep 24 '15 at 17:30
8
$\begingroup$

Notice, $$\lim_{x\to -\infty}\frac{\ln(1+3^x)}{\ln(1+2^x)}$$ substituting $x=-t$, we get
$$\lim_{t\to \infty}\frac{\ln(1+3^{-t})}{\ln(1+2^{-t})}$$ $$=\lim_{t\to \infty}\left(\frac{\frac{\ln(1+3^{-t})}{3^{-t}}}{\frac{\ln(1+2^{-t})}{2^{-t}}}\right)\times \frac{2^t}{3^t}$$

$$=\left(\frac{\lim_{t\to \infty}\frac{\ln(1+3^{-t})}{3^{-t}}}{\lim_{t\to \infty}\frac{\ln(1+2^{-t})}{2^{-t}}}\right)\times\lim_{t\to \infty} \left(\frac{2}{3}\right)^t$$ $$=\frac{1}{1}\times (0)=\color{red}{0}$$

$\endgroup$
  • 1
    $\begingroup$ How do you know the $\lim_\limits{x\to +\infty}{\frac{\ln(1+x)}{x}}$ without derivatives? I think it cannot be used... $\endgroup$ – Jason Sep 22 '15 at 19:49
  • 2
    $\begingroup$ @Jason $$\ln(1+x)=\int_0^x\frac{du}{1+u}\implies\frac{x}{1+x}\leqslant\ln(1+x)\leqslant x\implies\lim_{x\to0}\frac{\ln(1+x)}{x}=\cdots$$ $\endgroup$ – Did Sep 22 '15 at 21:30
  • $\begingroup$ @Did You are using integrals. We haven't even be taught "officially" derivatives, so integrals are a bit... $\endgroup$ – Jason Sep 23 '15 at 17:37
  • 1
    $\begingroup$ @Jason Quote: "Prove the following limit without using approximations and derivatives". $\endgroup$ – Did Sep 23 '15 at 18:55
  • $\begingroup$ @Did Those are the bounds that I used in my answer and I obtained them quickly using the identical approach as you "Did" (sorry for the superfluous pun). After the OP requested, I provided a way forward to establish those bounds without appealing to differential or integral calculus. But, I agree with you ... the quote was explicit. $\endgroup$ – Mark Viola Sep 23 '15 at 19:42

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.