Approximation of error function-type integral

[RedSonja]

Hi! How do I approximate the integral
\begin{equation} \int_0^{\infty} dt \:e^{-iA(t-B)^2} \end{equation}
with [itex]A, B[/itex] real, [itex]A > 0[/itex], and [itex]B=b \cos\theta[/itex] where [itex]0 \leq \theta < 2\pi[/itex]?
I guess for [itex] B\ll 0[/itex] the lower limit may be extended to [itex] - \infty[/itex] to yield a full complex gaussian integral, but what about [itex]B \geq 0[/itex]? And what happens for [itex]A \gg 1[/itex] and [itex]A \ll 1[/itex] respectively?
Thanks for your help!

 

[HallsofIvy]

If you let [itex]x= \sqrt{A}(t- B)[/itex] then the integral becomes
[tex]\frac{1}{\sqrt{A}}\int_{-\sqrt{A}B}^\infty e^{-x^2}dx[/tex]

 

[RedSonja]

Ok, so for

\begin{equation}
\frac{1}{\sqrt{A}} \int_{-\sqrt{A}B}^{\infty} dx \: e^{-ix^2}
\end{equation}

with [itex]\sqrt{A}B[/itex] large and positive we may extend the limit to [itex]-\infty[/itex] and obtain [itex]\sqrt{\frac{\pi}{A}} e^{-i\frac{\pi}{4}}[/itex], and for [itex]\sqrt{A}B\approx 0[/itex] we get half of that, but what happens for [itex]\sqrt{A}B[/itex] negative? I suppose we get 0 for large, negative [itex]\sqrt{A}B[/itex] (?), but I don't know how to handle the approximation for "intermediate" negative [itex]\sqrt{A}B[/itex].

Can anyone help? Thanks!