I want to evaluate $$ \int_0^1 ( 1 - x^2)^{10} dx $$ One way I can do this is by expanding out $(1 - x^2)^{10}$ term by term, but is there a better way to do this?

  • $\begingroup$ Yes, by coming up with the reduction formula! $\endgroup$ – user209663 Jun 12 at 21:31

When you encounter this type of problem, the first thing to do is to come up with a reduction formula. Let us find the reduction formula for $\int(k^2 - x^2)^n dx $.

Using integration by parts: $u = (k-x^2)^n $ and $dv = dx$ then $du = n(k^2 - x^2)^{n-1}(-2x)dx $ and $v =x$


$$\int(k^2 - x^2)^n dx = x(k^2 - x^2)^{n-1} + 2n \int x^2(k^2 - x^2)^{n-1} \\ = x(k^2 - x^2)^{n-1} + 2n \int \bigg[ k^2(k^2 - x^2)^{n-1} - (k^2 - x^2)^{n} \bigg] dx \\ = x(k^2 - x^2)^{n-1} + 2n \int k^2(k^2 - x^2)^{n-1} dx - 2n \int (k^2 - x^2)^{n} dx $$ From here you can see that: $$2n\int(k^2 - x^2)^n dx + \int(k^2 - x^2)^n dx = x(k^2 - x^2)^{n-1} + 2n \int k^2(k^2 - x^2)^{n-1} dx - 2n \int (k^2 - x^2)^{n} dx $$

Thus, $$\int(k^2 - x^2)^n dx = \frac{x(k^2 - x^2)^{n-1}}{2n+1} + \frac{2nk^2}{2n+1} \int (k^2 - x^2)^{n-1} dx$$

For your problem, $k=1$ and $n=10$. Can you take it from here?

  • $\begingroup$ wow! I knew there must be a shorter way. Thank you! $\endgroup$ – MathStudent Jun 12 at 21:45
  • 2
    $\begingroup$ Is this really a shorter way? Evaluating all of the terms of the recursion seems like just as much effort as simply using the binomial expansion. $\endgroup$ – Mortified Through Math Jun 12 at 21:55
  • $\begingroup$ @MathStudent No problem. $\endgroup$ – user209663 Jun 12 at 21:58
  • $\begingroup$ @SZN Note that this is a definite integral... look at the recursion.... what happen to the first part? This is similar to $\int_0^{\pi} sin^n(x) dx$. You also get a recursion... but the trick is the bound of the integral. It makes it easier to evaluate $\endgroup$ – user209663 Jun 12 at 21:58
  • 5
    $\begingroup$ @user209663 what you say is perfectly fine. But by using the binomial expansion I can evaluate the integral in just a few lines without any recursion or argument about bounds. $\endgroup$ – Mortified Through Math Jun 12 at 22:06

This answer is exactly what the asker is NOT asking for. However, I believe it will be useful to show the asker that expanding out "term-by-term" is not actually that hard.

We have

\begin{align} \int_0^1 (1-x^2)^{10}dx &= \int_0^1 \sum_{i=0}^{10} C(10,i)(-x^2)^idx \\ % & = \sum_{i=0}^{10} (-1)^iC(10,i)\int_0^1 x^{2i}dx\\ % &= \sum_{i=0}^{10} \frac{(-1)^iC(10,i)}{2i+1}, \end{align}

where $$C(n,m) = \frac{n!}{m!(n-m)!}.$$


An alternate approach is through the beta function. We start by u-substitution and try to make our way towards an integral form of the beta function:

$$\begin{align*} \int_0^1(1-x^2)^{10}\,\mathrm dx &= \frac{1}{2} \int_0^1 (1-u)^{10}u^{-1/2}\,\mathrm du,\qquad x=\sqrt{u} \\ &= \frac{1}{2} \int_0^1 (1-u)^{11-1}u^{1/2-1}\,\mathrm du\\ &= \frac{1}{2}\cdot\frac{\Gamma(11)\Gamma(1/2)}{\Gamma(11+1/2)}\\ &= \frac{1}{2}\cdot\frac{10!\sqrt{\pi}}{\frac{22!}{4^{11}\cdot11!}\sqrt{\pi}}\\ &= \frac{262144}{969969}. \end{align*}$$

  • 6
    $\begingroup$ Upvoted. There's a special place in my heart for using special functions to solve problems that look like they only admit basic methods :-) $\endgroup$ – Mortified Through Math Jun 12 at 22:10
  • $\begingroup$ what formula did you use to calculate $\Gamma(11.5)$? $\endgroup$ – Henry Lee Jun 14 at 23:34
  • 1
    $\begingroup$ @Henry Lee It's a rearrangement of the Legendre duplication formula. $\endgroup$ – dxdydz Jun 14 at 23:36
  • 1
    $\begingroup$ @Henry Lee If you like that, it gets even better. The duplication formula is a special case of that one. $\endgroup$ – dxdydz Jun 14 at 23:46
  • 1
    $\begingroup$ So it has a general formula! These are the kinds of equations I love finding but never bump into doing my usual work. @dxdydz thank you so much $\endgroup$ – Henry Lee Jun 14 at 23:49

Similar to dxdydz's answer, let $x=\sin(t)$ $$\int_0^1 ( 1 - x^2)^{10}\, dx=\int_0^{\frac \pi 2}\cos^{21}(t)\,dt$$ and remember that $$\int_0^{\frac \pi 2}\cos^{n}(t)\,dt=\frac{\sqrt{\pi }}2 \frac{ \Gamma \left(\frac{n+1}{2}\right)}{ \Gamma \left(\frac{n+2}{2}\right)}$$


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.