Solution

ID: past-exam-of-the-mathematics-course-of-the-university-of-cambridge/2008/ii/paper-1/30a/solution

In put . Expanding the cubic exponential near the endpoint gives
The substitution evaluates the coefficient using the Gamma function. Thus, numbering the first term by ,
For any fixed truncation the exponential's Taylor remainder is bounded by the next power times , and replacing the finite upper limit by infinity adds an exponentially small tail. This justifies the asymptotic expansion for .
For large , the ratio of magnitudes of consecutive terms, by Stirling's formula, is
The least term consequently has index . An indexing convention starting at one changes this by one, without changing the estimate.
For the second integral, both endpoints minimize . The endpoint gives the same expansion with positive coefficients. Its first two contributions are
To check their ordering, near write . Then and , producing a contribution . Because , this is smaller than the second displayed term. It is the third term, ahead of the next contribution of order . Splitting the integral into small endpoint neighborhoods and an interior interval, where the phase is strictly positive, makes the endpoint estimates rigorous.

New to topics? Read the docs here!