Evaluate the limit

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

My Attempt:

To simplify notation, let $A = \left(\sqrt[3]{x^3+x^2+1}\right)$ and $B = \left(\sqrt[3]{x^3-x^2+1}\right)$. Now

$$ \begin{align} 2x^2 &= A^3-B^3\\ x &= \sqrt{\frac{A^3-B^3}{2}} \end{align} $$

So the limit becomes

$$\lim_{x\rightarrow \infty}\left(\sqrt{\frac{A^3-B^3}{2}}-A-B\right)$$

How can I complete the solution from this point?

  • $\begingroup$ Factor x out of the roots and write the expression as x*(2-something-something). That may help. $\endgroup$ – Johannes Ernst Jun 30 '14 at 3:34
  • $\begingroup$ $\sqrt[3]{x^3+x^2+1}-\sqrt[3]{x^3-x^2+1}=2x-\cfrac{2x^{-1}}{9}+o(x^{-1})$ $\endgroup$ – Fabien Jun 30 '14 at 3:48

Note that $\sqrt[3]{x^3+x^2+1}=x\sqrt[3]{1+\frac{1}{x}+\frac{1}{x^3}}$, with a similar expression for the other one. Make the substitution $t=1/x$. Our expression becomes $$\frac{2-\sqrt[3]{1+t+t^3}-\sqrt[3]{1-t+t^3}}{t},$$ and we want to find the limit of this as $t$ approaches $0$ from the right. Note that the conditions for using L'Hospital's Rule are met. So our limit is $$\lim_{t\to 0^+} -\frac{1}{3}(1+3t^2)(1+t+t^3)^{-2/3}-\frac{1}{3}(-1+3t^2)(1-t+t^3)^{-2/3}.$$ This one is easy.


It is easier to break into two limits, $x-\sqrt[3]{x^3+x^2+1}$ and $x-\sqrt[3]{x^3-x^2+1}$ then we shall see the first limits to $-\frac{1}{3}$ and the second $\frac{1}{3}$ so the sum limits to zero.

Note the more general result,

$$x-\sqrt[n]{x^n +ax^{n-1}+\cdots}\rightarrow -\frac{a}{n}$$ If we write $A=\sqrt[n]{x^n +ax^{n-1}+\cdots}$ then

$$x-A=\frac{x^n-A^n}{x^{n-1}+x^{n-2}A +\cdots +A^{n-1}}=\frac{-ax^{n-1}+\cdots}{x^{n-1}+x^{n-2}A +\cdots +A^{n-1}}$$

and since $\frac{A}{x}\rightarrow 1$ the rest is easy.



Since $x$ goes to infinity, let us write $$\sqrt[3]{x^3+x^2+1}=x\sqrt[3]{1+\frac{1}{x}+\frac{1}{x^2}}$$ and let us define $y=\frac{1}{x}+\frac{1}{x^2}$. Now, look at the Taylor expansion of $$\sqrt[3]{1+y}=1+\frac{y}{3}-\frac{y^2}{9}+O\left(y^3\right)$$ and replace $y$ by its definition and expand to get finally $$\sqrt[3]{x^3+x^2+1}=x \Big(1+\frac{1}{3 x}-\frac{1}{9 x^2}+...\Big)\simeq x-\frac{1}{9 x}+\frac{1}{3}$$ Similarly, the same method should give $$\sqrt[3]{x^3-x^2+1} \simeq x-\frac{1}{9 x}-\frac{1}{3}$$

I am sure that you can take from here.

More intuitively, you could have noticed that, since $x$ goes to $\infty$,$$\sqrt[3]{x^3 \pm x^2+1}=x\sqrt[3]{1 \pm\frac{1}{x}+\frac{1}{x^2}}$$ behaves just as $x$ and then the limit of the expression is $0$. However, this approach allows you to show "how" the expression goes to $0$ (by positive or negative values).


I think this identity can be used to simplify your expression.
Let $a=\sqrt[3]{x^3+x^2+1}$ and $b=\sqrt[3]{x^3-x^2+1}.$
As $x\rightarrow \infty$ we have $a+b\rightarrow 2x.$ Hence $$\lim_{x\rightarrow \infty}\left\{2x-\left(\sqrt[3]{x^3+x^2+1}+\sqrt[3]{x^3-x^2+1}\right)\right\}\\= \lim_{x\rightarrow \infty} {2x-(a+b)}=0$$

  • 1
    $\begingroup$ This still needs some minor corrections, although basic idea seems to be ok. You cannot say that $a(x)+b(x)\to 2x$ when $x\to\infty$. (Since a and b depend on x.) $\endgroup$ – Martin Sleziak Aug 21 '14 at 8:49

As $x \to\infty$, $A\to x$ and $B\to x$. Therefore the expression would tend to $0$.

  • 2
    $\begingroup$ This kind of reasoning is faulty for example $x-\sqrt[3]{x^3+x^2+1} \rightarrow -\frac{1}{3}$ $\endgroup$ – Rene Schipperus Jun 30 '14 at 4:08

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.