This is easy enough to simulate and find the answer is somewhere around .812. However I am not finding it so easy to solve the integral involved which I believe is...

$$\int_0^1 \sqrt{x^2+(1-x)^2} dx$$

I don't seem to get the right answer if I use this derivation. Bonus points for finding the expected area which is a much easier problem!

  • $\begingroup$ The linked derivation has some problems. The first is that the display following "we now have" should have $c_1$, not $c_1^2$. The second is that in the final equation, the RHS is the integral of $\sqrt{P(x)}$, not of $\sqrt{(x+b_1)^2+c_1}$. If you make these changes you'll get the answer you expected. $\endgroup$ – rogerl Dec 6 '19 at 22:25

You have $\sqrt{x^2+(x-1)^2}=\sqrt{2x^2-2x+1}=\sqrt2\sqrt{(x-1/2)^2+1/4}$. Take $m=x-1/2$ and the integral becomes$$2\sqrt2\int_0^{1/2}\sqrt{m^2+(1/2)^2}dm$$which is a standard form that evaluates to$$2\sqrt2\left[\frac m2\sqrt{m^2+(1/2)^2}+\frac{(1/2)^2}2\log\left|m+\sqrt{m^2+(1/2)^2}\right|\right]_0^{1/2}\approx0.812$$The expected area is similarly given by$$\frac12\int_0^1x(1-x)dx=1/12$$


$$\begin{align}I&=\int_0^1 \sqrt{x^2+(1-x)^2} dx\\ &=\int_0^1\sqrt{2x^2-2x+1}dx\\ &=\int_0^1\sqrt2\sqrt{\left(x-\frac12\right)^2+\frac14}dx\end{align}$$ Let $x-\frac12=\frac12\tan\theta$. Then, $dx=\frac12\sec^2\theta d\theta$. $$\begin{align}I&=\frac{1}{2\sqrt2}\int_{-\frac{\pi}{4}}^{\frac{\pi}{4}}|\sec\theta|\sec^2\theta d\theta\\ &=\frac{1}{2\sqrt2}\int_{-\frac{\pi}{4}}^{\frac{\pi}{4}}\sec^3\theta d\theta\\ \end{align}$$

which can be integrated by parts.

Then, $$\begin{align}I&=\frac{1}{2\sqrt{2}}\left(\frac12\sec\theta\tan\theta+\frac12\ln|\sec\theta +\tan\theta|\right)|_{-\frac{\pi}{4}}^{\frac{\pi}{4}}\\ &=\frac{1}{2\sqrt{2}}\left(\sqrt{2}+\frac12\ln\left|\frac{\sqrt2+1}{\sqrt2-1}\right|\right)\end{align}$$


Use the substitution $u=x-\frac{1}{2}$

$$\int_{-\frac{1}{2}}^{\frac{1}{2}} \sqrt{\left(u+\frac{1}{2}\right)^2+\left(u-\frac{1}{2}\right)^2}du = \int_{-\frac{1}{2}}^{\frac{1}{2}} \sqrt{2u^2 + \frac{1}{2}}du = 2\sqrt{2}\int_0^{\frac{1}{2}} \sqrt{u^2+\frac{1}{4}}du$$

Then there are two ways to approach this problem, either let $u=\frac{1}{2}\tan\theta$ or $u=\frac{1}{2}\sinh t$. Personally I find the second one easier as it doesn't require an integration by parts

$$ \frac{1}{\sqrt{2}} \int_0^{\sinh^{-1}\left(1\right)} \cosh^2 t dt = \frac{1}{2\sqrt{2}} \int_0^{\sinh^{-1}\left(1\right)} 1 + \cosh(2t) dt = \frac{1}{2\sqrt{2}}\Biggr[t + \frac{1}{2}\sinh(2t)\Biggr]_0^{\sinh^{-1}\left(1\right)}$$

$$=\frac{1}{2\sqrt{2}}\Biggr[t + \sinh(t)\cosh(t)\Biggr]_0^{\sinh^{-1}\left(1\right)}=\frac{1}{2\sqrt{2}}\Biggr[t + \sinh(t)\sqrt{1+\sinh^2(t)}\Biggr]_0^{\sinh^{-1}\left(1\right)}$$

$$ = \frac{\sinh^{-1}\left(1\right)}{2\sqrt{2}} + \frac{1}{2} \approx 0.8116$$

by using the hyperbolic identities $\cosh(2t) = \cosh^2(t) + \sinh^2(t)$ and $\cosh^2(t)-\sinh^2(t) = 1$.

  • $\begingroup$ I think you have a typo in the decimal answer. Should it be 0.81161262? This is the most interesting answer so far as it expresses the exact solution in the simplest terms. $\endgroup$ – Goldbug Dec 10 '19 at 18:57
  • $\begingroup$ @Goldbug yes you're absolutely right. I have fixed it. $\endgroup$ – Ninad Munshi Dec 10 '19 at 19:39

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.