The composite Simpson's rule approximates a definite integral by applying Simpson's rule to successive
pairs of equal subintervals. For and
, it is
|
(1)
|
If
has a continuous fourth derivative, then for some
the error is
|
(2)
|
where the error means the exact integral minus its approximation. Thus the error satisfies
|
(3)
|
Here
is big-O notation indicating that as
, the error magnitude is bounded by a constant multiple
of
for all sufficiently small positive
.