Partial Derivatives

Given a function z = f(x, y), it represents a surface in Re (1K)3 and so a rate of change at a point depends on the direction from the point. We look at these rates of change along lines parallel to the x- and y-axis.

Consider the function l5g1 (1K) at the point (1/2, 1). We calculate the following limits

l5g2 (4K)

l5g3 (3K)
These limits are the slopes of the tangent lines to the curve f(x, 1) and f(1/2,y) respectively and are called the partial derivative of f with respect to x at (1/2, 1) and the partial derivative of f with respect to y these are denoted by fx(1/2,1) and fy. The situation is shown in the graph below. The curves f(x,1) and f(1/2,y) are the traces from the planes y = 1 and x = 1/2 cutting the surface z = f(x,y).

pderv1 (41K)

These partial derivatives are calculated in the same way as derivatives are calculated for functions of a single variable with the exception that we treat the other variable as a constant.

The following example shows this

f(x, y) = x cos(xy)

fx = cos(xy) + x (­sin(xy) y) = cos(xy) ­ xy sin(xy)

fy = x (­sin(xy) x) = ­x2 sin(xy)
Other Notations

Below is the graph of the function f(x, y) = x e(– x2 – y2) and following that are the graphs of fx and fy. Can you tell which is which?

Partial Derivatives

There are higher order derivatives these have the notations

It should come as no surprise that if the partial derivatives are all continuous then fxy = fyx. Below are the graphs of fxx, fyy, and fxy for the function f(x, y) = x e(­ x2 ­ y2)

fxx fyy

fxy

Back