2
$\begingroup$

I'm trying to prove: $$\int_0^\infty \frac{\sin x}{x}dx = \frac{\pi}{2}$$

I have defined $\; I(a)= \int_0^\infty\frac{\sin x}{x}e^{-ax}dx\;$ for $a>0$, and I'm trying to show that: $$\frac{dI}{da}(a)=-\int_0^\infty \sin(x) e^{-ax}dx$$

I started like this: $$\frac{dI}{da}(a)= \lim_{h \to 0} \frac{I(a+h)-I(a)}{h}=\lim_{h \to 0} \int_0^\infty \frac{\sin x}{x}\biggl(\frac{e^{-(a+h)x}-e^{-ax}}{h}\biggr)dx$$

Next I would like to put the limit inside the integral, and then I get what I want, but how can I justify this? Thanks

$\endgroup$
  • $\begingroup$ This might help. $\endgroup$ – TheSimpliFire Jan 6 '18 at 14:57
  • $\begingroup$ have you tried the Leibniz integral rule, to derivate under the integral sign ? $\endgroup$ – hamza boulahia Jan 6 '18 at 14:57
  • $\begingroup$ I haven't learned the Leibniz integral rule. The only case I know I am allowed to put limits inside an integral is when I have uniform convergence, which in this case I'm not sure how to prove $\endgroup$ – user401516 Jan 6 '18 at 14:58
  • $\begingroup$ One thing that you can do is just to copy the proof of the Leibniz rule for this specific case. The proof I know uses the DCT. $\endgroup$ – Shashi Jan 6 '18 at 15:49
  • $\begingroup$ Your caution deserves praise, especially since even after showing $\displaystyle I(a)=\arctan\frac1a$, you'll still have to justify $\displaystyle\lim_{a\to0+}\int^\infty_0\frac{\sin x}x\,e^{-ax}\,dx=\int^\infty_0\frac{\sin x}x\,\lim_{a\to0+}e^{-ax}\,dx$. $\endgroup$ – Professor Vector Jan 6 '18 at 16:01
1
$\begingroup$

You can use Fubini to calculate $I(a)$: \begin{align} \int^\infty_0 e^{-ax}\frac{\sin x}x\,dx&=\int^\infty_0 e^{-ax}\int^1_0\cos tx\,dt\,dx \\&=\int^1_0\int^\infty_0 e^{-ax}\cos tx\,dx\,dt \\&=\int^1_0\frac{a}{a^2+t^2}\,dt=\arctan\frac1a \end{align} Since $|\cos tx|\le1$, the double integral is absolutely convergent for any $a>0$.
As I said, this is not enough, we have to justify $\displaystyle I(0)=\lim_{a\to0+}I(a)$. We have $$I(0)-I(a)=\int^\infty_0\sin x\,\frac{1-e^{-ax}}x\,dx=\int^\infty_0\cos x\,\frac{1-(1+ax)\,e^{-ax}}{x^2}\,dx$$ by partial integration. Now, we can use the dominated convergence theorem, because $|\cos x|\le1$, and the fraction is $\ge0$ and $\le$ the minimum of $a^2$ and $x^{-2}$, i.e. dominated by an integrable function on $[0,\infty)$.

$\endgroup$
  • $\begingroup$ How can I use it to push to limit inside the integral? $\endgroup$ – user401516 Jan 6 '18 at 18:01
  • $\begingroup$ @user401516 I've just added the necessary details. $\endgroup$ – Professor Vector Jan 6 '18 at 18:15
0
$\begingroup$

What you have done is possible only if the function is itegrable. Then because of linearity of derivation, integral and limit you can substitute them. This is possible since ${sinx\over x}{e^-ax}$ is integrable and bounded for $x\in [0,\infty)$ when $a\ge0$

$\endgroup$
  • $\begingroup$ When $a=0$, this function is not absolutely integrable over $[0,\infty)$ $\endgroup$ – TheOscillator Jan 6 '18 at 16:31
  • $\begingroup$ It doesn't require for absolute integrability. It suffices only to be integrable.... $\endgroup$ – Mostafa Ayaz Jan 6 '18 at 16:58
  • $\begingroup$ In the sense of an Improper Riemann integrable? Do you have a reference for your claim? Now I am curious. $\endgroup$ – TheOscillator Jan 6 '18 at 17:39
  • $\begingroup$ Sure! Please refer two "Real Mathematical Analysis" written by Charles C. Pugh $\endgroup$ – Mostafa Ayaz Jan 6 '18 at 17:45

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.