3

$$\lim_{x\rightarrow 0} \,\,\left( \sqrt[3]{1+2x+x^3} - \frac{2x}{2x+3} \right) ^ {\frac1{x^3}} $$

I have already tried several options, but the only answer I have gotten so far is $e^{\infty}$, which is incorrect. The correct answer is $e^\frac{43}{81}$, which is easy to get by using Taylor series, but our task was to get the same one by using L'Hospital's rule. Can you help me with it?

lunary
  • 85
  • 9
  • 5
    Maybe this was set to demonstrate the superiority of Taylor series methods over the Hospital? – Angina Seng Nov 24 '19 at 08:33
  • 1
    Welcome to MSE. It is in your best interest that you type your questions (using MathJax) instead of posting links to pictures. – José Carlos Santos Nov 24 '19 at 08:33
  • 1
    @Lord Shark the Unknown could be so, but still I was required to solve it using both approaches :< – lunary Nov 24 '19 at 08:41
  • Write this as the exponential of a logarithm, take the limit inside exponential, use the logarithm to pull the power off, use L'Hopital's rule until you get 43/81 on the inside of the exponential – Brevan Ellefsen Nov 24 '19 at 08:41

3 Answers3

3

The standard approach here is to take logarithm and then evaluate the limit. If $f(x) $ denotes the expression under limit and it tends to a limit $L$ then \begin{align} \log L&=\lim_{x\to 0}\frac{1}{x^3}\log\left(\sqrt[3]{1+2x+x^3}-\frac{2x}{2x+3}\right)\notag\\ &=\lim_{x\to 0}\frac{g(x)}{x^3}\cdot\frac{\log(1+g(x))}{g(x)}\notag\\ &=\lim_{x\to 0} \frac{g(x)} {x^3}\notag\\ &=\lim_{x\to 0}\frac{1}{x^3}\left(\sqrt[3]{1+2x+x^3}-1-\frac{2x}{2x+3}\right)\notag \end{align} The algebraic limit at the end can be evaluated using standard techniques. Here is one approach which requires some algebra. If $$A=\sqrt[3]{1+2x+x^3},B=\frac{4x+3}{2x+3}$$ then $A,B$ both tend to $1$ and $$A^3-B^3=x^3+\frac{(1+2x)(2x+3)^3-(4x+3)^3}{(2x+3)^3}=x^3+\frac{16x^3(1+x)} {(2x+3)^3}$$ so that $(A^3-B^3)/x^3\to 43/27$. And therefore $$\frac{A-B} {x^3}=\frac{A^3-B^3}{x^3}\cdot\frac{1}{A^2+AB+B^2}\to \frac{43}{81}$$ Thus the desired limit is $L=e^{43/81}$.

Advanced tools like L'Hospital's Rule and Taylor series are often unnecessary for simple limit problems and instead a little bit of algebra helps.


Incidentally this exercise helps us to know that $(3+2x)/(3+x)$ is a very good approximation to $\sqrt[3]{1+x}$ for small values of $x$. For example if $x=0.1$ then we have $$\sqrt[3]{1.1}=1.032280\dots,\frac{3.2}{3.1}=1.032258\dots$$

  • Well, I do think that working out the Maclaurin series of $g(x)$ up to the $x^3$ term is rather easier than what you've done here.... – Angina Seng Nov 24 '19 at 11:24
  • @LordSharktheUnknown: Taylor is in general the most efficient approach (I have upvoted your comment to the question), but I prefer to demonstrate the use of theoretically simpler approach here. – Paramanand Singh Nov 24 '19 at 13:56
2

Let $L$ be the limit, then taking the natural log, $$\begin{split}\log L&=\lim_{x\to0}\frac1{x^3}\log\left(\sqrt[3]{1+2x+x^3}-\frac{2x}{2x+3}\right)\\&\stackrel{\mathrm{L'H}}{=}\lim_{x\to0}\frac{1}{3x^2}\frac{1}{\sqrt[3]{1+2x+x^3}-(2x)/(2x+3)}\left(\frac{3x^2+2}{3(1+2x+x^3)^{2/3}}-\frac{6}{(2x+3)^2}\right).\end{split}$$ The latter limit is rather tedious to evaluate, but can be done. (If we should encounter indeterminate forms again, then just apply L'Hospital's rule again.) You should end up finding that $\log L=43/81$, after which exponentiating both sides gives the answer. The key idea here is taking the logarithm, which allows us to convert the given limit into the form required to apply L'Hospital's rule.

YiFan Tey
  • 17,431
  • 4
  • 28
  • 66
  • But actually can I apply L'Hospital's rule again after taking the first derivative if I have a number in numerator and 0 in denominator (not 0/0 or infinity/infinity)? – lunary Nov 24 '19 at 08:56
  • @MariaWhite No, of course you can't do that. – YiFan Tey Nov 24 '19 at 08:57
  • Well, the problem is that if you try to substitute x for 0 in your derivative (by the way, previously I have gotten the same result), it turns out that we have number/0 and I have no idea what to do next – lunary Nov 24 '19 at 09:03
  • @MariaWhite Yes, I have just realised that as well. After checking with a computer I think I missed out a term when taking the derivative. I will correct the answer once I've figured out where it went wrong. – YiFan Tey Nov 24 '19 at 09:04
  • @MariaWhite Fixed now, I forgot a factor of $3$ in the denominator. Now when you substitute $x=0$ the numerator becomes $0$ as well. – YiFan Tey Nov 24 '19 at 09:06
  • Now I get it. Thank you so much! :> – lunary Nov 24 '19 at 09:14
  • Glad to help! ${}$ – YiFan Tey Nov 24 '19 at 09:14
0

Taking the logarithm, what I should do is to consider $$\frac{\log \left(\sqrt[3]{x^3+2 x+1}-\frac{2 x}{2 x+3}\right)}{x^3}$$ and you need to apply L'Hospital rule three times (just because of the denominator). Not funny but doable. And the limit is effectively $\frac{41}{81}$.

Now, take the exponential of it.