0
$\begingroup$

Through some calculations I ended up with the integral $\int_0^\infty\exp(iz(\frac{1}{3}t^3+t))dt$. I would like to obtain the result that the behaviour of the integral is $\sim\frac{i}{z}$ as $z\rightarrow\infty$

I am not very good in obtaining a nice asymptotic approximation but my guess here is to use the method of steepest descents. Calculating the derivative of $\frac{1}{3}t^3+t$ and setting it to zero gives me $t_1=i,t_2=-i$ but how can we gain the directions of steepest descent now?

$\endgroup$

1 Answer 1

2
$\begingroup$

This is not perhaps explicitely about steepest descent: Let $h$ be the inverse of the function $t\mapsto \frac 13 t^3+t$. If we take $s=\frac 13 t^3+t$ as a new variable of integration, the integral becomes $\int_ 0^\infty \textrm{exp}(i zs) h'(s)\,ds $. An integration by parts gives (assuming $Im(z)\geq 0$) $ \int_ 0^\infty \textrm{exp}(i zs) h'(s)\,ds = \Big /_0^\infty \frac 1{iz}\textrm{exp}(i zs) h'(s) - \int_0^\infty \frac 1{iz}\textrm{exp}(i zs) h''(s)\, ds\\= \frac iz + \frac i{z}\int_0^\infty \textrm{exp}(i zs) h''(s)\, ds$,
since $h'(0)=1$. Because $h''$ is integrable, the integral $\int_0^\infty \textrm{exp}(i zs) h''(s)\, ds$ approaches zero as $z\to \infty$ (assuming $Im(z)\geq 0$) and this should give the desired result.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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