Suppose you have a function like \(f(x)=(x^3+1)^5\) and you want to find \(f'(x)\). We would use the chain rule, which tells us to first derive the “outside function” \((—-)^5\), leave the inside alone, and then multiply by the derivative of the inside function. This gives you \(f'(x)= 5\cdot (x^3+1)^4\cdot 3x^2 = 15x^2\cdot (x^3+1)^4\).
Now, what if instead we were given the function \(f(x)=15x^2\cdot (x^3+1)^4\) and we are asked to antiderive it? Noting that antideriving is just “backwards deriving,” I can put on my Sherlock Holmes hat and observe that the chain rule was probably used to get the derivative \(f(x)=15x^2\cdot (x^3+1)^4\) because the derivative of the inside is \(3x^2\) which is close to what we have on the outside. So, in order to antiderive this, if I could somehow work backwards by removing the derivative of the inside function from that \(15x^2\) on the outside and then antiderive the outside \((—-)^5\), that would give me an antiderivative of \(f(x)\). This idea is entirely encapsulated in the idea of U-Substitution given explicitly below.
First we identify the inside function and call it \(u\):
Then we derive our \(u=x^2+1\) to get \(\frac{du}{dx}=2x\). Now, every place we see \(x^2+1\) we will replace it with a \(u\) and every place we see \(2x\) we will replace it with \(\frac{du}{dx}\), and then “cancel” the \(dx\)’s.
Now that we have no \(x\)’s in the integrand and only have \(u\)’s, we will antiderive the current integrand in terms of \(u\) as usual:
Lastly, we replace all \(u\)’s with \(x^2+1\) to get the final answer (because the original problem was in terms of \(x\), not these \(u\)’s that we substituted to help with the problem).
Suppose we want to find \(\int (x^5+3)^6\cdot x^4\ dx\). Now if we let \(u=x^5+3\), then \(\frac{du}{dx}=5x^4\) which is NOT EXACTLY what we have being multiplied on the outside. As such, we cannot carry out the substitution just yet because of this mismatch. We need to make a few alterations to the \(\frac{du}{dx}\) equation before we can substitute. This is illustrated in the quick example below.
Let \(u=x^5+3\). Then note that \(\frac{du}{dx}=5x^4\). Since we have \(x^4\) in the integrand, we will get \(x^4\) by itself on one side of the equation \(\frac{du}{dx}=5x^4\) so we have something to replace the \(x^4\) with. Thus, we have:
Now we can replace \(x^5+3\) with \(u\) and \(x^4\) with \(\frac{1}{5}\cdot \frac{du}{dx}\) respectively. This gives us
Now we can antiderive as usual in terms of \(u\), bringing the \(\frac{1}{5}\) along for the ride.
Lastly, we replace the \(u\) with the original \(x^5+3\) and call it a day:
Let \(u=x^3+1\)
\(\frac{(x^3+1)^{21}}{21}+C\)
Let \(u=2x+1\)
\(\frac{(2x+1)^{301}}{602}+C\)
Let \(u=3x+5\)
\(\frac{(3x+5)^{3.5}}{10.5}+C\)
Let \(u=x^2+x\). Also, convert radical to fractional power.
\(\frac{4(x^2+x)^{\frac{3}{2}}}{3}+C\)
Let \(u=x^2+4\)
\(\frac{-(x^2+4)^{-3}}{6}+C\)
Let \(u=x-20\)
\(10(x-20)^{\frac{1}{2}}+C\)
Let \(u=3x+1\)
\(\frac{2(3x+1)^{\frac{1}{2}}}{3}+C\)
Let \(u=x^3+\frac{3}{2}x^2\)
\(\frac{(x^3+\frac{3}{2}x^2)^\frac{2}{3}}{2}+C\)
Let \(u=\sqrt{x}+1\). It would also be helpful to convert radicals to fractional powers
\(\frac{2(x^{\frac{1}{2}}+1)^3}{3}+C\)