Integration by substitution, or u-substitution,
is the most common technique of finding an antiderivative.
It allows us to find the antiderivative of a function by reversing
the chain rule. To see how it works, consider the following
example.
Let f(x)=(x2−2)8. Then f′(x)=8(x2−2)7(2x) by
the chain rule. Remember that we have the "inside part"
(x2−2), and its derivative (2x). We can use this
knowledge to antidifferentiate, if we have an integrand that has
an "inside part" and is multiplied by the derivative of that
"inside part".
To see this, consider ∫8(x2−2)7(2x)dx. We make a
substitution to make it clear what to do: Let u be the inside part, so u=x2−2. We differentiate u,
and use the notation du=2xdx
(instead of dudx=2x; you will see why when we subsitute
back into the integrand). Then ∫8(x2−2)7(2x)dx=∫8u7du by directly replacing every
element in the first integrand, including the dx,
by exactly what it becomes via the substution (in
blue above). This
integral is now easy to evaluate, giving u8+c=(x2−2)8+c.
Notice our integrand above was a composite function, f(g(x))
where g(x)=x2−2 was the inside part. This u-substitution process can be
stated formally as shown below. Notice that the first
integrand is a composite function multiplied by the derivative of
the inside part.
∫f(g(x))g′(x)dx=∫f(u)du.
Another example:
∫cos(x3)⋅3x2dxu=x3,du=3x2dx=∫cos(u)du=sin(u)+c=sin(x3)+c. Do:
Differentiate to check this answer!
By choosing a suitable function u, we can often convert hard
integrals into much easier integrals that we know how to evaluate.
Unfortunately, there is no magic formula for deciding what u
should be — this is a toolkit, not a recipe. Some examples
are in the video that follows.