2
$\begingroup$

Solve $$ \int_0^\pi \frac {x} {1+\sin x} dx $$ using $\sin x = \frac {2\tan(x/2)} {1+\tan^2(x/2)}$ and the substitution $t = \tan(x/2)$. I tried doing this but I got to a point where my integral limits were $0$ to $\infty $. This happened when I substituted for $\tan(x/2)$ Is there a way of doing this using this substitution only? And also why does this happen?

UPDATE: What I did - $$ I = \int_0^\pi \frac {\pi-x} {1+\sin x} dx = \int_0^\pi \frac {\pi} {1+ \sin x} dx - I$$ Using $\sin x = \frac {2\tan(x/2)} {1+\tan^2(x/2)}$ $$ $$ $$ 2I = \int_0^\pi \frac {\pi} {1+\sin x} dx = \pi\int_0^\pi \frac {\sec^2(x/2)} {1+ \tan^2(x/2)+2 \tan(x/2)} dx$$ Now if there were no limits, this could've been solved easily by $t = \tan x$. But I can't do that because if I did, the limits would become $0$ to $\infty$. A way to solve this would be multiplying and dividing by $1+ \sin x$ but I don't want to do that. I want to use $t=\tan x$

$\endgroup$
  • $\begingroup$ Maybe you should try integration by parts first to get rid of that free $x$ on the top there. $\endgroup$ – Bob Jones Mar 17 '17 at 8:00
  • $\begingroup$ Set $x\rightarrow x-\pi$ and see what happpens..nearly the same question was asked a few days ago by the user @bui so have a look in his profile $\endgroup$ – tired Mar 17 '17 at 8:08
1
$\begingroup$

After the main symmetry trick, another symmetry trick and a rationalization: $$ \int_{0}^{\pi}\frac{du}{1+\sin u}=2\int_{0}^{\pi/2}\frac{1-\sin u}{\cos^2 u}\,du =2\left[\tan u-\frac{1}{\cos u}\right]_{0}^{\pi/2}=\color{red}{\large2}.$$

$\endgroup$
  • $\begingroup$ I don't want to use rationalization, I want to do it using the substitution u = tanx. Tricks are fine but the substitution has to be that. $\endgroup$ – Cowgirl Mar 17 '17 at 9:38
  • $\begingroup$ @AnswerQuicklyPlease: you still have to split the integration range in halves since $\sin(x)$ is not injective on $(0,\pi)$. And after that, rationalization leads to the solution way faster than the given substitution. $\endgroup$ – Jack D'Aurizio Mar 17 '17 at 9:44
1
$\begingroup$

If we must solve the problem with the suggested substitution, the integral becomes $\int_0^\infty\frac{4\arctan t dt}{(1+t)^2}$. The identity $\int_0^\infty f(t)dt=\int_0^1\left(f(t)+\tfrac{1}{t^2}f\left(\tfrac{1}{t}\right)\right)dt$ changes this to $\int_0^1\frac{2\pi dt}{(1+t)^2}=\pi$.

$\endgroup$
0
$\begingroup$

Hint. Alternatively, by the change of variable $u=\pi-x$, one gets $$ I=\int_0^\pi \frac {x} {1+\sin x} dx=\int_0^\pi \frac {\pi-u} {1+\sin (\pi-u)} du=\pi\int_0^\pi \frac {1} {1+\sin u} du-I $$ the latter integral being easier to evaluate.

$\endgroup$
  • 3
    $\begingroup$ Always the same old good trick ! $\endgroup$ – Claude Leibovici Mar 17 '17 at 8:10

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.