# Does this integral have a closed-form solution?

I'm very rusty with calculus, and I was hoping someone would be willing to help me with the following definite integral:

$$\int_b^{\infty} \frac{\cos(ax)}{1+x^2} dx$$

$$b>0$$

-
This does it in Mathematica, but it ain't pretty: Integrate[Cos[a*x]/(1 + x^2), {x, b, Infinity}, Assumptions -> {b > 0, a > 0}] // FullSimplify –  Nick Thompson Apr 25 '12 at 0:33
I'm thinking curiosity's going to win and I'm going to end up looking in a table of integrals to see if this is in there. –  Nicholas Stull Apr 25 '12 at 0:43
Thanks for trying everyone. –  Stuck_pls_help Apr 25 '12 at 0:45

For $b = 0$ or $b = -\infty$ you can use contour integration. For other $b$ (assuming $a \neq 0$), you will have to do it numerically.
(If there were a formula in terms of $b$, you would have an elementary formula for an antiderivative of the integrand (just differentiate your formula with respect to $b$ and you'll recover the negative of the integrand!), but there is no elementary formula for an antiderivative of this integrand, I believe).
I had a counterexample in mind when I made that comment. If we evaluate $\int_0^b \sin(\frac{1}{x})\, dx$, we can differentiate the closed form solution (which is expressible in terms of the cosine integral) with respect to $b$, but that integrand does not necessarily admit an elementary antiderivative. Or am I missing something? –  Nicholas Stull Apr 25 '12 at 2:51