5

I was asked to evaluate the integral

$$\int_{-1}^{1} \frac{\sin{x}}{1+x^2}dx$$

if it exists.

This is a problem from Calculus and the student has been taught how to use trigonometric substitution. My intuition was to do trig sub with $$x=\tan{\theta}$$ and eliminating $$\frac{dx}{1+x^2}$$ but when I do that the numerator becomes $$\sin{(\tan{\theta})}$$ which I am not comfortable integrating.

Another queue was that the problem asks "if it exists" so I tried to see if there are any points within $(-1,1)$ that may cause any problem, but I don't think I see any asymptotes or undefined numbers so I'm not sure if I am dealing with an improper integral.

Can someone help me out on this? Thank you.

Jeff Faraci
  • 10,393
hyg17
  • 5,271
  • 4
  • 42
  • 90

3 Answers3

12

Zero by symmetry.

It is an odd integrand integrated over a symmetric bound

or more properly "integrated over an interval that is symmetric about the origin".

Jeff Faraci
  • 10,393
5

The integral is \begin{align} I &= \int_{-1}^{1} \frac{\sin(x)}{1+x^{2}} dx \\ &= \int_{-1}^{0} \frac{\sin(x)}{1+x^{2}} dx + \int_{0}^{1} \frac{\sin(x)}{1+x^{2}} dx \\ &= - \int_{0}^{1} \frac{\sin(x)}{1+x^{2}} dx + \int_{0}^{1} \frac{\sin(x)}{1+x^{2}} dx \\ &= 0 \end{align} where the change of variable $x \rightarrow -x$ was made.

Leucippus
  • 27,174
2

We will begin by evaluating $$I(x) = \int\!\dfrac{\sin(x)}{1+x^2}\mathrm{d}x$$

$$I = \dfrac{1}{2i} \int\!\dfrac{\sin(x)}{x-i}\mathrm{d}x - \dfrac{1}{2i}\int\!\dfrac{\sin(x)}{x+i}\mathrm{d}x$$

$$I = \dfrac{1}{2i} \int\!\dfrac{\sin(x)}{x-i}\mathrm{d}x - \dfrac{1}{2i}\int\!\dfrac{\sin(x)}{x+i}\mathrm{d}x$$

We will now consider $$J_\pm = \int\!\dfrac{\sin(x)}{x\pm i}\mathrm{d}x$$

Substitute $u = x \pm i$.

$$J_\pm = \int\!\dfrac{\sin(u \mp i)}{u}\mathrm{d}x = \int\!\dfrac{\sin(u)\cos(i)\mp\sin(i)\cos(u)}{u}\mathrm{d}x$$

Splitting the integral and using hyperbolic functions

$$J_\pm = \int\!\dfrac{\sin(u)\cosh(1)}{u}\mathrm{d}x \mp\int\!i\dfrac{\sinh(1)\cos(u)}{u}\mathrm{d}x$$

Writing in terms of $\operatorname{Si}(x)$ and $\operatorname{Ci}(x)$

$$J_\pm = \cosh(1)\operatorname{Si}(u) \mp i\sinh(1)\operatorname{Ci}(u)$$

$$J_\pm = \cosh(1)\operatorname{Si}(x\pm i) \mp i\sinh(1)\operatorname{Ci}(x\pm i)$$

$$I = \dfrac{1}{2i}\left(J_--J_+\right)$$

$$I = \dfrac{\left(\cosh(1)\operatorname{Si}(x- i) + i\sinh(1)\operatorname{Ci}(x- i)\right)-\left(\cosh(1)\operatorname{Si}(x+ i) - i\sinh(1)\operatorname{Ci}(x+ i)\right)}{2i}$$

Now evaluate $I$ at $1$ and $-1$.

$$I(1) = I(-1) = -0.324967578038053553$$

$$I(1) - I(-1) = \boxed{0}$$

If you are sneaky you can use symmetry to show that $I(1) = I(-1)$ without actually calculating the value.

Brad
  • 5,252
  • 2
  • 21
  • 51
  • This is not of use to a beginning calculus student. – Ted Shifrin May 14 '14 at 01:48
  • @TedShifrin but it is useful to everyone else. The problem is posted for the benefit of the community rather than simply the person who asks the question. This is why "defacing" your own questions is not allowed. If anyone wants to see how they would evaluate this indefinite integral without W|A they can refer to my post instead of posting a new thread. An "acceptable" answer has already been posted. – Brad May 14 '14 at 01:50
  • This was actually very helpful for me. Thanks. – hyg17 May 14 '14 at 03:10