How can we calculate $$ \int_{0}^{\infty}{\sin\left(x\right)\sin\left(2x\right)\sin\left(3x\right)\ldots \sin\left(nx\right)\sin\left(n^{2}x\right) \over x^{n + 1}}\,\mathrm{d}x ? $$

I believe that we can use the Dirichlet integral

$$ \int_{0}^{\infty}{\sin\left(x\right) \over x}\,\mathrm{d}x = {\pi \over 2} $$

But how do we split the integrand?

  • $\begingroup$ Just to be sure before trying something: Is it on purpose that the last factor is $n^2$ and not $n+1$ (that would fit the pattern)? $\endgroup$ – mickep Jun 29 '16 at 7:20
  • $\begingroup$ If it would have been n+1 instead of n^2 .. then I know the answer for sure $\endgroup$ – Aman Rajput Jun 29 '16 at 7:47
  • $\begingroup$ I used mathematica to calculate some results,we set result is I, and $n=1,I=\frac{\pi}{2}$,$n=2,I=\pi$,$n=3,I=3 \pi$,$n=4,I=12 \pi$,$n=5,I=60 \pi$ $\endgroup$ – Young1997 Jun 29 '16 at 8:48
  • $\begingroup$ so the answer is $\frac{n!\pi}{2}$ $\endgroup$ – Young1997 Jun 29 '16 at 8:50
  • $\begingroup$ @Young1997 In this case, that method works. But sometimes, it can fail for large numbers. See this $\endgroup$ – Sophie Agnesi Jun 29 '16 at 13:29

We have (theorem $2$, part $(ii)$, page 6) that:

If $a_{0},\dots,a_{n} $ are real and $a_{0}\geq\sum_{k=1}^{n}\left|a_{k}\right|$, then $$\int_{0}^{\infty}\prod_{k=0}^{n}\frac{\sin\left(a_{k}x\right)}{x}dx=\frac{\pi}{2}\prod_{k=1}^{n}a_{k}.$$

So it is sufficient to note that if we take $a_{0}=n^{2},\, a_{k}=k,\, k=1,\dots,n $ we have $$a_{0}=n^{2}\geq\frac{n\left(n+1\right)}{2}=\sum_{k=1}^{n}a_{k} $$ hence

$$\int_{0}^{\infty}\frac{\sin\left(n^{2}x\right)}{x}\prod_{k=1}^{n}\frac{\sin\left(kx\right)}{x}dx=\frac{\pi n!}{2}.$$


We can use contour integration to show that $$\int_{0}^{\infty} \frac{\sin(x) \sin(2x) \cdots \sin(nx) \sin(n^{2}x)}{x^{n+1}} \, dx = \frac{\pi n!}{2} . $$

Consider the complex function $$f(z) = \frac{\sin(z) \sin(2z) \cdots \sin(nz) e^{in^{2}z}}{z^{n+1}}. $$

If we can argue that the magnitude of the numerator is bounded in the upper half-plane, then it follows from the estimation lemma that $\int f(z ) \, dz$ vanishes along the upper half of the circle $|z|=R$ as $R \to \infty$.

Notice that numerator of $f(z)$ can be expressed as $$\frac{e^{in^{2}z}}{(2i)^{n}}\prod_{k=1}^{n} \left(e^{ikz}-e^{-ikz} \right) . $$

Since $n^2 \ge \frac{n(n+1)}{2} = \sum_{k=1}^{n} k$ , this is a linear combination of exponential functions of the form $e^{i a z}$, where $a \ge 0$. And in the upper half-plane, the magnitude of such exponential functions never exceed $1$.

Therefore, by integrating $f(z)$ around a indented contour that consists of the real axis and the large semicircle above it, we get $$ \text{PV} \int_{-\infty}^{\infty} f(x) \, dx - \pi i \, \text{Res} [f(z), 0] = 0 \, , $$ where

$$ \begin{align} \text{Res} [f(z), 0] &= \lim_{z \to 0} z \, \frac{\sin(z) \sin(2z) \cdots \sin(nz) e^{in^{2}z}}{z^{n+1}} \\ &= \lim_{z \to 0} \frac{\sin(z)}{z} \frac{\sin(2z)}{z} \cdots \frac{\sin(nz)}{z} \, e^{inz^{2}} \\ &=1 \cdot 2 \cdots n \cdot 1 \\ &=n!. \end{align}$$

Equating the imaginary parts on both sides of the equation, we get $$\int_{-\infty}^{\infty} \frac{\sin(x) \sin(2x) \cdots \sin(nx) \sin(n^{2}x)}{x^{n+1}} \, dx = \pi n! . $$

And since the integrand is even, it follows that $$\int_{0}^{\infty} \frac{\sin(x) \sin(2x) \cdots \sin(nx) \sin(n^{2}x)}{x^{n+1}} \, dx = \frac{\pi n!}{2} .$$

  • 1
    $\begingroup$ awesome answer. a very big (+1)! $\endgroup$ – tired Jul 19 '16 at 10:09

If you know the answer can you verify with my answer

$$\frac12[\pi(n^2 n!)-\frac{\pi}{2^n(n+1)!}((1+2+..+n+n^2)-2)^2]$$

If you think it is not correct, then verify it by reducing the terms

  • $\begingroup$ Are you sure this is correct? In the case of $n = 1$ the integral evaluates to $\pi/2$, but your solution gives $\pi$? $\endgroup$ – okrzysik Jun 29 '16 at 8:14
  • $\begingroup$ Sorry this is thee answer for integration between -infty to infty .. I forgot to divide by 2 since limits is changed in question $\endgroup$ – Aman Rajput Jun 29 '16 at 9:29
  • $\begingroup$ I have edited it now... you can verify for other too.. if too wrong then I will check again $\endgroup$ – Aman Rajput Jun 29 '16 at 9:30
  • $\begingroup$ I don't think this gives the correct answer either, my calculations seem to agree with the OPs' in the comment section of $\pi n!/2$. $\endgroup$ – okrzysik Jun 29 '16 at 9:41
  • $\begingroup$ Why don't you include the steps of the solution? $\endgroup$ – mickep Jun 29 '16 at 9:52

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.