Use of Green's functions in the numerical solution of two-point boundary value problems