Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I want show that this improper integral convergence: $$\int_{0}^{+\infty}\frac{\sin x}{x+\sin x}dx$$ please help me.

share|cite|improve this question
Different ways how to do it. Please observe that as a limit for x→0 the function is defined. So as a definite integral from, say (0,1] the integral is convergent. From (1,→), perform integration by parts. You then get behind the integral sign a (x+sinx)² in the denominator. Using Comparison theorem you can find that these integrals are convergent because x+sinx is greater than 0.5x – imranfat May 1 '13 at 18:23
up vote 8 down vote accepted

For $0\lt x\le\pi$, $0\le\dfrac{\sin(x)}{x+\sin(x)}=\dfrac{\sin(x)/x}{1+\sin(x)/x}\le\dfrac12$, so $$ 0\le\int_0^{\pi}\frac{\sin(x)}{x+\sin(x)}\,\mathrm{d}x\le\frac\pi2 $$

Integrate by parts to get $$ \begin{align} \int_\pi^\infty\frac{\sin(x)}{x+\sin(x)}\,\mathrm{d}x &=\left.-\frac{\cos(x)}{x+\sin(x)}\right]_\pi^\infty -\int_\pi^\infty\frac{\cos(x)(1+\cos(x))}{(x+\sin(x))^2}\,\mathrm{d}x\\[6pt] &=-\frac1\pi-\int_\pi^\infty\frac{\cos(x)(1+\cos(x))}{(x+\sin(x))^2}\,\mathrm{d}x \end{align} $$ and the last integral converges since the absolute value of the numerator of the integrand is bounded by $2$ and the denominator is greater than $(x-1)^2$.

share|cite|improve this answer

The integral converges near $x=0$ because the limit of the integrand is $1$ there. And at infinity, the integrand approaches $\sin{x}/x$, which is integrable there. To see this, Taylor expand the denominator:

$$\frac{\sin{x}}{x+\sin{x}} \approx \frac{\sin{x}}{x} - \frac{\sin^2{x}}{x^2}$$

both of which are integrable at infinity. Further terms are still smaller.

Is there an $x \ne 0$ such that the denominator vanishes? Let's see: $x > \sin{x}$ so that there will be no cancellation outside of $x=0$. Thus, we have convergence.

share|cite|improve this answer
You are being careless at infinity, methinks. – Harald Hanche-Olsen May 1 '13 at 18:24
Better now, yes. I am not surprised that you can do it, but consider your audience. Also, there is the example $\sum_n (-1)^n/(\sqrt{n}+(-1)^n)$, which diverges. This shows that you can't be too careful. – Harald Hanche-Olsen May 1 '13 at 18:29
@GitGud: thanks. My wife likes the new look, but she is really tiring of the bats. – Ron Gordon May 1 '13 at 18:35
$$\frac{1}{x+\sin{x}} = \frac{1}{x} \frac{1}{1+(\sin{x}/x)} \approx \frac{1}{x} \left( 1 - \frac{\sin{x}}{x}\right )$$ – Ron Gordon May 1 '13 at 19:23
@RonGordon: ah, I think what may have been causing some confusion is that the Taylor series you are using is for $$\frac{u}{1+u}=u-u^2+u^3-u^4+u^5-u^6+\dots$$which converges for $|u|\lt1$ and using $u=\sin(x)/x$ in $$\frac{\sin(x)}{x+\sin(x)}=\frac{\sin(x)/x}{1+\sin(x)/x}$$ I was not certain of that at first since you mentioned Taylor expanding the denominator. – robjohn May 1 '13 at 20:20

We have $$\lim_{x\to 0}\frac{\sin x}{x+\sin x}=\lim_{x\to 0}\frac{ x}{x+ x}=\frac{1}{2}$$ hence the integral $$\int_0^1\frac{\sin x}{x+\sin x}dx$$ is convergent

Moreover, by integration by part we have $$\int_1^A\frac{\sin x}{x+\sin x}dx=- \left[\frac{\cos x}{x+\sin x}\right]_1^A-\int_1^A\frac{\cos x(\cos x+1)}{(x+\sin x)^2}dx$$ and with $$|\frac{\cos x}{x+\sin x}|\leq |\frac{1}{x-1}|\to 0,\quad x\to\infty$$ and $$|\frac{\cos x(\cos x+1)}{(x+\sin x)^2}|\leq \frac{2}{(x-1)^2}$$ and the integral $$\int_2^\infty \frac{1}{(x-1)^2}dx$$ is convergent so the integral $$\int_1^\infty\frac{\sin x}{x+\sin x}dx$$ is also convergent and the desired result follows

share|cite|improve this answer

The problem is at $+\infty$, as the function tends to $1/2$ at $0$ and it is continuous on $(0,+\infty)$. Now $$ \frac{\sin x}{x+\sin x}-\frac{\sin x}{x}=\frac{x\sin x-\sin x(x+\sin x)}{x(x+\sin x)}=-\frac{\sin^2 x}{x(x+\sin x)}\sim-\frac{\sin^2x}{x^2} $$ at $+\infty$, so the former is integrable. And for $\frac{\sin x}{x}$, it is easy to prove convergence of the improper integral with an integration by parts on $[1,x]$.

Note: or as pointed out by imranfat, you can directly do integration by parts of the original integrand on $[1,x]$. It amounts essentially to the same.

share|cite|improve this answer
For $(\sin x)/x$, the question you refer to is overkill if all you want to show is that the integral converges. It is easier to put $a_n=\int_{\pi n}^{\pi(n+1)}(\sin x)/x\,dx$ and show that $\sum a_n$ converges by the alternating series test. – Harald Hanche-Olsen May 1 '13 at 18:28
@HaraldHanche-Olsen Or even better, an integration by parts. That's just the only reference I could find... I agree it does not make a great answer... I'll edit. – 1015 May 1 '13 at 18:33

The integral can be compared to the convergent integral $\int_0^\infty {\sin(x)\over x}dx={\pi\over2}$. The difference is given by $\int_0^\infty{\sin^2(x)\over {x(x+\sin(x))}}dx$, which is absolutely convergent, so the integral converges.

The convergence of $\int_0^\infty{\sin(x)\over x} dx$ is well-known, and can be established by splitting the integral up into integrals of the form $\int_{2k\pi}^{(2k+2)\pi} {\sin(x)\over x} dx$, with $k$ a positive integer. Each of these integrals has a positive and negative component, and it is clear that $$|{\int_{(2k)\pi}^{(2k+2)\pi} {\sin(x)\over x} dx}| < {2\over{(2k)\pi}}-{2\over{(2k+1)\pi}}={1\over{k(2k+1)\pi}}\,.$$ If the upper bound is not a multiple of $2\pi$, then there will be an additional contribution of $\int_{2n\pi}^{R+2n\pi}{\sin(x)\over x}dx$, with $n$ the last positive integer for which the upper limit is smaller than $2\pi n$ and $R$ smaller than $2\pi$, but this integral is clearly less than $R\over{2n\pi}$. The integral $\int_0^{R+2n\pi}{\sin(x)\over x}dx$ is therefore smaller than $2\pi+\sum_{k=1}^{n}{1\over{k(2k+1)\pi}}+{R\over{2n\pi}}$, with the $2\pi$ coming from the region from $x=0$ to $x=2\pi$. Since the sum is convergent as $n$ tends to $\infty$ (limit comparison to the $p$-series with $p=2$) and the additional piece tends to zero in this limit, we can say for sure that this integral converges. Showing that it is equal to $\pi\over2$ is a different matter, but it can be done in multiple ways (contour integration and relation to the Gamma function are the two that come to mind, but I am sure there are others).

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.