We can solve harder problems involving derivatives of integral
functions. For example, what do we do when our upper limit
of integration is not simply our variable x, as it must be to
use FTC 1, but is rather a
function of x? We will use FTC 2
to solve this FTC 1 problem.
Example: Compute ddx∫x21tan−1(s)ds.
Solution: Let F(x) be the antiderivative of
tan−1(x). Finding a formula for F(x) is hard, but we
don't actually need the antiderivative , since we will not
integrate. Recall that by FTC 2,
∫x21tan−1(s)ds=F(x2)−F(1),
so
ddx∫x21tan−1(s)ds=ddx(F(x2)−F(1)).
By the chain rule, we now get F′(x2)2x−0=2xf(x2)=2xtan−1(x2).
since F′=f by our assumption.
This method generalizes, but please do not try to memorize this; you
do not need to, because it is simply applying FTC
2 and the chain rule, as you see in the box below and
in the following video.
If f is a
continuous function with antiderivative F, and g and h
are differentiable functions, then ddx∫h(x)g(x)f(s)ds=ddx[F(h(x))−F(g(x))]
=F′(h(x))h′(x)−F′(g(x))g′(x)
=f(h(x))h′(x)−f(g(x))g′(x).
There is an an
alternate way to solve these problems, using FTC 1 and the chain
rule. We will illustrate using the previous
example.
Example: Compute
ddx∫x21tan−1(s)ds.
Solution:
We let u=x2 and let g(u)=∫u1tan−1(s)ds, and use the fact
that ddxg(u)=g′(u)dudx to get ddx∫x21tan−1(s)ds=ddx(∫u1tan−1(s)ds)=ddx(g(u))=g′(u)dudx,
By FTC 1, g′(u)=tan−1(u), and
dudx=ddx(x2)=2x, giving us
tan−1(u)dudx=tan−1(x2)⋅2x.