A fixed small-angle approximation alone is insufficient because its error is multiplied by . Introduce the shrinking cutoff , which satisfies but .
On , the Taylor expansion gives uniformly. Thus
uniformly there, and the two central integrals are asymptotically equal. On , choose a fixed with for . The sine-integral tail is at most ; the Gaussian tail has the same exponential bound with . Both are , while the central Gaussian integral is asymptotic to . Hence
This justifies the localized Laplace method rather than interchanging a nonuniform approximation with integration.

Articles by others on the same topic (0)

There are currently no matching articles.