4

Is this limit solved correctly?

$$\lim_{x\to0^+} \frac{(1+\sin x)^{\frac1x} - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\ln(1+\sin x)\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\ln(1+x+o(x^2)\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\left(\left[x+o(x^2)\right]-\frac12\left[x+o(x^2)\right]^2 + o(x^2)\right)\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\left[x-\frac{x^2}{2} + o(x^2)\right]\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(1-\frac{x}{2} + o(x)\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp(1)\cdot\exp\left(-\frac{x}{2}\right)\cdot\exp(o(x)) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(-\frac{x}{2}\right)\cdot\left(e\cdot\exp(o(x)) - 1\right)}{x^\alpha} = \frac{1\cdot(e\cdot1 -1)}{\lim_{x\to0^+}x^\alpha} = \begin{cases}\frac{e-1}{0^+} = +\infty && \ \text{if } \alpha >0\\\frac{e-1}{1} = e-1 &&\ \text{if } \alpha =0\\\frac{e-1}{+\infty} = 0 &&\ \text{if } \alpha <0\end{cases}$$

EDIT:

A first-order approximation is enough:

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\ln(1+x+o(x^2)\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(\frac1x\left[x+o(x^2)+o(x)\right]\right) - \exp(-\frac x2)}{(\tan x )^\alpha}$$

$$\lim_{x\to0^+} \frac{\exp\left(1+o(1)\right) - \exp(-\frac x2)}{(\tan x )^\alpha} = \frac{e-1}{\lim_{x\to0^+}x^\alpha} = \begin{cases}\frac{e-1}{0^+} = +\infty && \ \text{if } \alpha >0\\\frac{e-1}{1} = e-1 &&\ \text{if } \alpha =0\\\frac{e-1}{+\infty} = 0 &&\ \text{if } \alpha <0\end{cases}$$

aleio1
  • 995
  • 3
    we could rewrite $(1+\sin x)^{\frac 1{\sin x}\times \frac{\sin x}{x}}\to e$ so effectively numerator goes to $e-1$. – zwim Dec 11 '22 at 12:17
  • Thanks to all. @zwim Would it be correct also one of the following approximations?

    $(1+\sin x)^{\frac1x} \sim (1+x)^{\frac1x}\to e $ or $(1+\sin x)^{\frac1x} \sim (1+\sin x)^{\frac1{\sin x}}\to e $ ?

    – aleio1 Dec 12 '22 at 09:19
  • @AdamRubinson thanks I have just edited – aleio1 Dec 12 '22 at 10:53
  • Well it is correct because of the reason I gave, but it is really borderline to push your luck with these manipulations (i.e. $1+\sin x\sim 1+x$ because you have a sum in your equivalent) or to do equivalents with indeterminate forms (i.e. $1^\infty$). Rather it is preferable to make $(1+\frac 1u)^u$ appear which has a known limit $e$ at infinity. Then we manipulate a resulting exponent $\frac{\sin x}{x}$ which is no more infinite so this is ok. The other option is to use Landau $o(\ )$ notation as you did, which is equally correct. – zwim Dec 12 '22 at 20:25
  • I cannot see what could be wrong with a such approximation. I mean, is it not always true that if $g(x)\sim x$ for $x\to0$ then $\lim_{x\to0}f(g(x)) = \lim_{x\to0}f(x)$ ? – aleio1 Dec 13 '22 at 10:00
  • I just say you have to be cautious of not falling in traps like this one for instance https://math.stackexchange.com/a/3053385/399263 – zwim Dec 13 '22 at 10:26

0 Answers0