We use partial differentiation to differentiate a function of two or more variables. For example,
is a function of two variables.
If we want to find the partial derivative of a two-variable function with respect to x, we treat y as a constant and use the notation ∂x∂f.
If we want to find the partial derivative of a two-variable function with respect to y, we treat x as a constant and use the notation ∂y∂f.
You can think of ∂ as the partial derivative symbol, sometimes called “del.” When you see this symbol, it shows that we’re taking a partial derivative.
This notation should look familiar — it’s just like the derivative of a function in Leibniz’s notation, expressed dxdy. In this case, ∂ replaces the letter “d” with a stylized curly d.
In calculus, derivatives measure the rate of change of a function with respect to a change in its input variable x. Since the input of a multivariable function is more than one variable, we call ∂x∂f and ∂y∂f partial derivatives because they only reveal the rate of change of f when one variable changes, instead of both.
The partial derivative allows us to understand the behavior of a multivariable function when we let just one of its variables change, while the rest stay constant.
How to Do Partial Derivatives
How do partial derivatives work? To find a partial derivative, we find the derivative of a function of two or more variables by treating one of the variables as a constant.
To find ∂x∂f:
Treat y as a constant.
Differentiate the function normally.
To find ∂y∂f:
Treat x as a constant.
Differentiate the function normally.
After we designate one variable as a constant, we can use the derivative rules that are already familiar to us to differentiate the function.
Partial Derivative Rules
Derivative rules help us differentiate more complicated functions by breaking them into pieces. Here are some of the most common derivative rules to know:
Logarithmic and Exponential Rules
You can view more about these rules in an explanation by one of our instructors Dr. Tim Chartier.
For example, let’s take another look at the function f(x,y)=xy+x2y. Suppose we want to find ∂x∂f, the partial derivative with respect to x. The first thing to do is treat y as a constant.
What does it mean to treat y as a constant? A constant is a fixed, unchanging value. Examples of constants are 1, 3.5, 17, and 100,000. To treat y as a constant, we imagine that y is any of these infinite constant values.
We can do this because of the constant rule, which states that the derivative of any constant is 0.
To make it easier to imagine y as a constant, we can replace y with c or k, which are two variables that are commonly used to represent constant values.
Using this trick and replacing y with k, we have
Now, we can find the partial derivative ∂x∂f using the derivative rules. Remember to change k back to y when you have your final answer.
We can do the same to find ∂y∂f, the partial derivative with respect to y. The first thing to do is treat x as a constant. Remember that the square of any constant is simply another constant.
Now, we can find the partial derivative ∂y∂f using the derivative rules. Remember to change k back to x when you have your final answer.
Partial Derivatives Examples
Let’s take a look at some more partial derivative examples.
Find the partial derivatives of f(r,h)=πr2h.
This function represents the volume of a cylinder. When we find the partial derivative ∂r∂(πr2h), we find the rate of change of the cylinder’s volume as only the radius changes.
When we find the partial derivative ∂h∂(πr2h), we find the rate of change of the cylinder’s volume as only the height changes. So,
Find the partial derivatives of f(x,y,z)=xy3−zx+z.
How do partial derivatives work in more than two variables? Just the same! For a function with three variables, we change only one variable and treat the other two as constants. So,
Find the partial derivatives of f(x,y)=x2sin(y)−y2cos(x).
We can use the trigonometry derivative rules for this problem. Remember that sin(y) acts as a constant when we calculate ∂x∂(f(x,y)), and cos(x) acts as a constant when we calculate ∂y∂(f(x,y)). So,