1
$\begingroup$

While studying the method of steepest descent, I've come across this exercise

The problem reads: Evaluate the dominant term of the asymptotic expansion of the following integral

$$I_N(q)=\int_{-\infty}^{+\infty} \frac{1}{z-i}\,e^{-N(z^2-2iqz)}\,dz\qquad q\in\mathbb{R}$$ for $N\rightarrow \infty$

Now, the integrand has a saddle point in $$z_0=iq$$ and also has a pole for $$z_1=i$$ For $q<1$, I have no problem finding the dominant term by simply applying the method of steepest descent.

For $q>1$ however, I know that I have to take into account the pole of the integrand but I don't know how to do that.

Also, most books on the subject that I've read through only go over the "simple" case in which the integrand has no singularities, so I would appreciate if you have any suggestions (books, articles, etc) on the subject.

$\endgroup$
1
  • $\begingroup$ I'm rusty on this subject, can you specify the path that you are using when $q<1$? This should help illuminate the issues that arise when $q>1$. $\endgroup$
    – Ian
    Sep 4 '20 at 16:26
0
$\begingroup$

When $q>1$, write \begin{align*} & \int_{ - \infty }^{ + \infty } {\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} \\ & \!= \int_{ - \infty }^{ + \infty } {\!\!\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} - \int_{iq - \infty }^{iq + \infty } {\!\!\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} + \int_{iq - \infty }^{iq + \infty } {\!\!\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} \\ & \!= 2\pi i\mathop {{\mathop{\rm Res}\nolimits} }\limits_{z = i} \left[ {\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} } \right] + \int_{iq - \infty }^{iq + \infty } {\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} \\ & \!= 2\pi ie^{N(1 - 2q)} + \int_{iq - \infty }^{iq + \infty } {\frac{1}{{z - i}}e^{ - N(z^2 - 2iqz)} dz} . \end{align*} Now apply the method of steepest descents to the final integral.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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