Our current list of derivative rules take into account how the derivative operation interacts with various function arithmetic operations: e.g., addition, scalar multiplication, product, quotient. We round out this discussion by considering another very important function operation: namely, function composition.
Assume that the function \(f\) is defined as the composition of the functions \(g\) and \(u\text{,}\) so that \(f(x)=g(u(x))\text{.}\) If \(u\) is differentiable at \(x\text{,}\) and \(g\) is differentiable at \(u(x)\text{,}\) then \(f\) is differentiable at \(x\text{,}\) and we have
The main difficulty in using the chain rule is being able to recognize how a given function \(f\) can be described as a composition of functions \(f(x)=g(u(x))\text{.}\) The procedure below is designed to help you identify the inner function \(u\) in this representation. You do not need to use it when computing derivatives with the chain rule, but it may help you get the hang of this technique. See the first computation in ExampleΒ 1.22.3 for an illustration of ProcedureΒ 1.22.2.
To compute the derivative of a function \(f\) that is built from other functions using composition, proceed as follows.
Identify inner function.
Identify a function \(u=u(x)\) so that \(f(x)=g(u(x))\) for some function \(g\text{.}\) The function \(u(x)\) might be easily identified as an βinner functionβ, or may be a common expression that appears in the definition of \(f\text{.}\)
Remark1.22.4.βBlah formulationβ of the chain rule.
The main purpose of ProcedureΒ 1.22.2 is to give you a streamlined, procedural way of identifying a function as a composition of an inner function \(u\) and outer function \(g\text{,}\) and then applying the chain rule appropriately. As you get more accustomed to these computations, you may be able proceed more fluidly. A slightly more casual approach might be described as the βblah formulationβ of the chain rule. Here is how it works:
We describe a function of the form \(f(x)=g(u(x))\) as β\(g\)-of-blahβ, where βblahβ stands for \(u(x)\text{.}\)
Consider the application of the chain rule formula to \(f(x)=\sqrt{x^2+1}\text{.}\) We consider this function as the square root of blah, where blah is \(x^2+1\text{.}\) The chain rule then says the derivative of \(f\) should be the derivative of the square root function, evaluated at blah, times the derivative of blah. This yields
Using the chain rule in combination of our existing derivative formulas, leads to simple generalizations of these formulas where we replace the variable \(x\) with an arbitrary differentiable function \(u(x)\text{.}\) We call these βchain rule enhanced formulasβ.
We end this section with an application of the chain to a physical setting, where we investigate the rate of change of volume of an inflating balloon. Not surprisingly this depends both on the current radius of the balloon, and how fast that radius itself is expanding. The chain rule makes this observation much more precise! This is a first example of what is called a related rates problem. We will see more of these in SectionΒ 1.24.
The volume \(V\) (in cm\(^3\)) of a spherical inflatable balloon is computed as \(V=\frac{4}{3}\pi r^3\text{,}\) where \(r\) is the radius of the balloon (in cm).
Compute the rate of change of the volume \(V\) with respect to the radius \(r\text{.}\)
Suppose now that while inflating the balloon, the radius \(r\) is given by the function \(r=h(t)\text{.}\) Compute the rate of change of \(V\) with respect to \(t\text{.}\) Leave your answer in terms of \(h\) and \(h'\text{.}\)
We thus see that at any given time \(t_0\text{,}\) the rate of change with respect to \(t\) is the product of the rate of change of volume with respect to \(r\) for the radius value at \(t_0\) (i.e., for \(r_0=h(t_0)\)) and the rate of change of \(r\) with respect to time at that time \(t_0\text{.}\) (Of course this is just as the chain rule predicts: $\frac{dV}{dt}=\frac{dV}{dr}\cdot \frac{dr}{dt}$.) This makes general sense, physically. The rate of change of the radius is acting as a multiplier for the rate of change of the volume with respect to \(t\text{:}\) in particular the greater the rate at which the radius increases, the greater the rate at which the volume increases.