0

How to prove that the improper integral $$\int_0^{\infty} \sin(x^3)\, dx$$ converges?

I wanted to use the Taylor series for $\sin(x)$ , and with that I get $\lim \sum \frac{(-1)^n a^{6n+4}}{(2n+1)!(6n+4)} $ as $a$ converges to infinity, but I don't know what to do afterwards. Any ideas?

Henno Brandsma
  • 250,824
mida
  • 65

2 Answers2

1

Consider $$ T(A) := \int_0^A \sin(x^3)\;dx . $$ We want to show this converges as $A \to \infty$.
Change variables $y=x^3$, $$ T(A) = \int_0^{A^3} \frac{\sin(y)}{3y^{2/3}}\;dy . $$ So $T(A) = T_1+T_2(A)$, where $$ T_1 = \int_0^1 \frac{\sin(y)}{3y^{2/3}}\;dy \\ T_2(A) = \int_1^{A^3} \frac{\sin(y)}{3y^{2/3}}\;dy $$ Now for $T_1$, note $$ \left|\frac{\sin(y)}{3y^{2/3}}\right| \le \frac{1}{3y^{2/3}} $$ and $$ \int_0^1 \frac{dy}{3y^{2/3}} = 1 $$ so $T_1$ conveges.
Next for $T_2$, integrate by parts \begin{align} T_2(A) &= \int_1^{A^3} \frac{\sin(y)}{3y^{2/3}}\;dy = \left.-\frac{\cos(y)}{3y^{2/3}}\right|_{y=1}^{y=A^3} - \frac{2}{9}\int_1^{A^3}\frac{\cos(y)}{y^{5/3}}\;dy \\& = \frac{-\cos(A^3)}{3A^2} +\frac{\cos(1)}{3} -\frac{2}{9}\int_1^{A^3}\frac{\cos(y)}{y^{5/3}}\;dy \end{align} Note that $$ \lim_{A \to\infty} \frac{-\cos(A^3)}{3A^2} = 0 . $$ Also, $$ \left|\frac{\cos(y)}{y^{5/3}}\right| \le \frac{1}{y^{5/3}} $$ and $$ \int_1^\infty \frac{dy}{y^{5/3}} = \frac53 $$ Thus $\lim_{A\to\infty}T_2(A)$ exists.

GEdgar
  • 117,296
0

If you have seen the video link I mentioned then, Set $a=0,b=1,n=3,s=\frac{1}{3}$ Then it will follow that, $$\int_{0}^{\infty}\sin(x^{3})dx = \frac{\Gamma(1/3)}{6}$$

RAHUL
  • 1,600
  • I did(and thanm you for that) but my prof informed me that i can't use gamma function. Any other ideas? – mida Jan 27 '22 at 14:18
  • What your professor said is not a surprise. The problem is: show the integral converges. That is easier to do than: evaluate the integral. – GEdgar Jan 27 '22 at 14:29
  • Well, I happened to see this video which derives the Fresnel integral in awesome way. You can do that for your power cubed case too. Btw, I agree with @GEdgar. – RAHUL Jan 27 '22 at 14:37
  • Well that's exactly what i want to do. I dont want to evaluate the integral itself. How do i prove that it converges? I did some stuff, do you happen to have discord so that we could talk about it? If you have time. @GEdgar – mida Jan 27 '22 at 16:34