Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top


$$L = \lim_{n\rightarrow\infty} n \int_{0}^{1} \frac{{x}^{n-2}}{{x}^{2n}+x^n+1} \mbox {d}x$$

share|cite|improve this question
A general strategy here is to find some simpler expression which gives a lower bound and some simpler expression which gives an upper bound and hope that in the limit they become equal. – Qiaochu Yuan Jun 7 '12 at 20:39
I think you should add "super-duper mega awsome" to the title. – Michael Greinecker Jun 7 '12 at 21:04
up vote 5 down vote accepted

Marvis showed that $$ I_n = \int_0^1 \dfrac{nx^{n-2}}{x^{2n} + x^n + 1} dx = \int_0^1 \dfrac{dt}{t^{1/n}(t^2 + t + 1)}. $$ At this point one can use monotone convergence as Davide commented, but since Chris is interested in a more direct approach, here is one.

We have $$ I_n \geq \int_0^1 \dfrac{dt}{t^2 + t + 1}=\frac{\pi}{3\sqrt3}, $$ and for $n>1$ and for any small $\delta>0$ $$ I_n\leq \int_0^\delta \dfrac{dt}{t^{1/n}} + \int_\delta^1 \dfrac{dt}{\delta^{1/n}(t^2 + t + 1)}=\frac{\delta^{1-1/n}}{1-1/n}+\frac1{\delta^{1/n}}\Big(\frac{2\pi}{3\sqrt3}-\frac{2\arctan(\frac{2\delta+1}{\sqrt3})}{\sqrt3}\Big). $$ An inspection of the limit as $n\to\infty$ of the right hand side shows that $$ I_n\leq \delta+\frac{2\pi}{3\sqrt3}-\frac{2\arctan(\frac{2\delta+1}{\sqrt3})}{\sqrt3}+E(\delta,n), $$ where $E(\delta,n)\to0$ as $n\to\infty$ for any fixed $\delta>0$. Since $\arctan(z)$ is continuous at $z=\frac1{\sqrt3}$ with $\arctan(\frac1{\sqrt3})=\frac\pi6$, for any given $\varepsilon>0$, we can choose $\delta>0$ so small that $$ \delta+\frac{2\pi}{3\sqrt3}-\frac{2\arctan(\frac{2\delta+1}{\sqrt3})}{\sqrt3}<\frac{\pi}{3\sqrt3}+\frac\varepsilon2. $$ For sufficiently large $n$, we can also ensure $E(\delta,n)<\frac\varepsilon2$, implying that there is a threshold value $N$ such that $$ \frac{\pi}{3\sqrt3}\leq I_n<\frac{\pi}{3\sqrt3}+\varepsilon, $$ for all $n>N$.

share|cite|improve this answer
nice proof. Thanks! – user 1618033 Jun 7 '12 at 21:59

Let $x^n = t$. Then $n x^{n-1} dx = dt \implies nx^{n-2} dx = \dfrac{dt}x = \dfrac{dt}{t^{1/n}}$.

$$I_n = \int_0^1 \dfrac{nx^{n-2}}{x^{2n} + x^n + 1} dx = \int_0^1 \dfrac{dt}{t^{1/n}(t^2 + t + 1)}$$ Hence, $$\lim_{n \rightarrow \infty} I_n = \int_0^1 \dfrac{dt}{t^2 + t + 1} = \dfrac{\pi}{3 \sqrt{3}}$$

But can I change the limit and integral?...

share|cite|improve this answer
@Chris But I still need to argue why I can change the limit and integral and at the moment I don't have an argument. – user17762 Jun 7 '12 at 20:42
@ZevChonoles I might be missing something trivial, but why is the denominator ($t^{1/n}(t^2 + t + 1)$) in the integral $\geq 1$? For $t$ sufficiently close to $0$, the value is less than $1$. – user17762 Jun 7 '12 at 20:47
@Marvis: I was originally writing my comment about the integral $$\int_0^1\frac{nx^{n-2}}{x^{2n}+x^n+1}dx$$ until I realized that's not what we're taking the limit of. The DCT would work for the above, because the numerator is always $\leq n$ and the denominator is always $\geq 1$. Unfortunately I don't think there's any way of taking the limit before the change of variables (?) – Zev Chonoles Jun 7 '12 at 20:49
Isn't there monotone convergence? – Davide Giraudo Jun 7 '12 at 20:51
I don't see how this is very simple, – Gigili Jun 7 '12 at 21:42

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.