Adaptive Simpson and Romberg integration methods

[robve]

(03-24-2021 10:19 PM)Albert Chan Wrote:  FYI, you can avoid infinity by doing atan(1/x) = sign(x)*pi/2 - atan(x)

Quote right. Point taken. But not everyone will see or understand your smart change.

I picked this example a bit arbitrarily to show that there are cases when the integral may be not be evaluable at an endpoint when in fact it is if no floating point exception is raised. If an exception is raised, then there are two options: either rewrite the integrand (if possible) or use quadratures that are applicable to open intervals.

- Rob