Partial Derivatives
Partial Derivatives
Partial Derivatives
P. Sam Johnson
In this lecture, we see how partial derivatives are defined and interpreted
geometrically, and how to calculate them by applying the rules for
differentiating functions of a single variable.
The slope of the curve z = f (x, y0 ) at the point P(x0 , y0 , f (x0 , y0 )) in the
plane y = y0 is the value of the partial derivative of f with respect to x at
(x0 , y0 ).
The tangent line to the curve at P is the line in the plane y = y0 that
passes through P with this slope. The partial derivative ∂f /∂x at (x0 , y0 )
gives the rate of change of f with respect to x when y is held fixed at the
value y0 .
The slope of the curve z = f (x0 , y ) at the point P(x0 , y0 , f (x0 , y0 )) in the
vertical plane x = x0 is the partial derivative of f with respect to y at
(x0 , y0 ). The tangent line to the curve at P is the line in the plane x = x0
that passes through P with this slope. The partial derivative gives the rate
of change of f with respect to y at (x0 , y0 ) when x is held fixed at the
value x0 .
The partial derivative with respect to y is denoted the same way as the
partial derivative with respect to x:
∂f ∂f
(x0 , y0 ), fy (x0 , y0 ), , fy .
∂y ∂y
Notice that we now have two tangent lines associated with the surface
z = f (x, y ) at the point P(x0 , y0 , f (x0 , y0 )).
Is the plane they determine tangent to the surface at P? We will see that
it is for the differentiable functions, we will learn how to find the tangent
plane. First we have to learn more about partial derivatives themselves.
Example 3.
Find the values of ∂f /∂x and ∂f /∂y at the point (4, −5) if
f (x, y ) = x 2 + 3xy + y − 1.
Solution : To find ∂f /∂x, we treat y as a constant and differentiate with
respect to x:
∂f ∂ 2
= (x + 3xy + y − 1) = 2x + 3 · 1 · y + 0 − 0 = 2x + 3y .
∂x ∂x
The value of ∂f /∂x at (4, −5) is 2(4) + 3(−5) = −7.
P. Sam Johnson Partial Derivatives August 30, 2019 14/117
Solution (contd...)
Implicit differentiation works for partial derivatives the way it works for
ordinary derivatives, as the next example illustrates.
Example 6.
Find ∂z/∂x if the equation yz − ln z = x + y defines z as a function of the
two independent variables x and y and the partial derivative exists.
The slope is the value of the partial derivative ∂z/∂y at (1, 2):
∂z ∂ 2 2
= (x + y ) = 2y |(1,2) = 2(2) = 4.
∂y (1,2)
∂y (1,2)
then
∂f ∂ ∂
= [x sin(y + 3z)] = x sin(y + 3z)
∂z ∂z ∂z
∂
= x cos(y + 3z) (y + 3z) = 3x cos(y + 3z).
∂z
We shall see that if the partial derivatives of f (x, y ) exist and are
continuous throughout a disk centered at (x0 , y0 ), however, then f is
continuous at (x0 , y0 ).
(
0, xy 6= 0
The graph of f (x, y ) = consists of
1, xy = 0
the linesL1 and L2 and the four open quadrants of
the xy -plane. The function has partial derivatives
at the origin but is not continuous there.
(a) Since f (x, y ) is constantly zero along the line y = x (except at the
origin), we have
lim f (x, y ) = lim 0 = 0.
(x,y )→(0,0) y =x (x,y )→(0,0)
(b) Since f (0, 0) = 1, the limit in part (a) proves that f is not continuous
at (0, 0).
(c) To find ∂f /∂x at (0, 0), we hold y fixed at y = 0. Then f (x, y ) = 1
for all x, and the graph of f is the line L1 in the figure. The slope of
this line at any x is ∂f /∂x = 0. In particular, ∂f /∂x = 0 at (0, 0).
Similarly, ∂f /∂y is the slope of line L2 at any y , so ∂f /∂y = 0 at
(0, 0).
Notice the order in which the mixed partial derivatives are taken:
∂2f
Differentiate first with respect to y , then with respect to x.
∂x∂y
fyx = (fy )x Means the same thing.
∂2f ∂2f
∂ ∂f ∂ ∂f
= = − sin y + e x = = − sin y + e x
∂y ∂x ∂y ∂x ∂x∂y ∂x ∂y
∂2f ∂2f
∂ ∂f ∂ ∂f
= = ye x . = = −x cos y .
∂x 2 ∂x ∂x ∂y 2 ∂y ∂y
We have noticed in the above example that the mixed second-order partial
derivatives
∂2f ∂2f
and
∂y ∂x ∂x∂y
are equal. This is not a coincidence. They must be equal whenever
f , fx , fy , fxy , and fyx are continuous, as stated in the following theorem.
ey
w = xy + .
y2 + 1
∂w ∂2w
=y and = 1.
∂x ∂y ∂x
∂3f
= fyyx ,
∂x∂y 2
∂4f
= fyyxx ,
∂x 2 ∂y 2
and so on. As with second-order derivatives, the order of differentiation is
immaterial as long as all the derivatives through the order in question are
continuous.
fy = −4xyz + x 2
fyx = −4yz + 2x
fyxy = −4z
fyxyz = −4.
The starting point for differentiability is not the difference quotient we saw
in studying single-variable functions, but rather the idea of increment.
Recall that if y = f (x) is differentiable at x = x0 , then the change in the
value of f that results from changing x from x0 to x0 + ∆x is given by an
equation of the form
in which ε → 0 as ∆x → 0.
in the value of f that results from moving from (x0 , y0 ) to another point
(x0 + ∆x, y0 + ∆y ) in R satisfies an equation of the form
Existence alone of the partial derivatives at that point is not enough, but
continuity of the partial derivatives guarantees differentiability.
The Chain Rule for functions of a single variable says that when
then
For functions of two or more variables, the Chain Rule has several forms.
The form depends on how many variables are involved, but once this is
taken into account, it works like the Chain Rule for functions of a single
variable.
We can now probably predict the Chain Rule for functions of three
variables, as it only involves adding the expected third term to the
two-variable formula.
Theorem 20 (Chain Rule for Functions of Three Independent
Variables).
If w = f (x, y , z) is differentiable and x, y , and z are differentiable
functions of t, then w is a differentiable function of t and
dw ∂w dx ∂w dy ∂w dz
= + + .
dt ∂x dt ∂y dt ∂z dt
Under the conditions stated below, w has partial derivatives with respect
to both r and s that can be calculated in the following way.
dy Fx
=− .
dx Fy
We have seen several different forms of the Chain Rule in this section, but
each one is just a special case of one general formula. When solving
particular problems, it may help to draw the appropriate branch diagram
by placing the dependent variable on top, the intermediate variables in the
middle, and the selected independent variable at the bottom.
1. fx = y 2 , fy = 2xy , fz = −4z
2. fx = −x(x 2 + y 2 + z 2 )−3/2 , fy = −y (x 2 + y 2 + z 2 )−3/2 ,
fz = −z(x 2 + y 2 + z 2 )−3/2
3. fx = √ yz2 2 2 , fy = √ xz2 2 2 , fz = √ xy2 2 2
1−x y z 1−x y z 1−x y z
1 2 3
4. fx = x+2y +3z , fy = x+2y +3z , fz = x+2y +3z
5. fx = sech2 (x + 2y + 3z), fy = 2 sech2 (x + 2y + 3z),
fz = 3 sech2 (x + 2y + 3z)
6. fx = y cosh(xy − z 2 ), fy = x cosh(xy − z 2 ), fz = −2z cosh(xy − z 2 )
V δv 2
W (P, V , δ, v , g ) = PV +
2g
4. Wilson lot size formula
km hq
A(c, h, k, m, q) = + cm +
q 2
xy + z 3 x − 2yz = 0
xz + y ln x − x 2 + 4 = 0
∂2w 2
2∂ w
= c ,
∂t 2 ∂x 2
where w is the wave height, x is the distance variable, t is the time
variable, and c is the velocity with which the waves are propagated.
1. w = cos(2x + 2ct)
2. w = ln(2x + 2ct)
3. w = f (u), where f is a differentiable function of u, and u = a(x + ct), where a is a
constant.
4. Does a function f (x, y ) with continuous first partial derivatives throughout an open region
R have to be continuous on R? Give reasons for your answer.
5. If a function f (x, y ) has continuous second partial derivatives throughout an open region
R, must the first-order partial derivatives of f be continuous on R? Give reasons for your
answer.
∂f ∂2f
= .
∂t ∂x 2
Show that u(x, t) = sin(αx) · e −βt satisfies the heat equation for
constants α and β.
E (a + h, b + k)
lim √ = 0.
(h,k)→(0,0) h2 + k 2
The function f is differentiable if it is differentiable at each point of its
domain. The function L(x, y ) is called the local linearization of f (x, y )
near (a, b).
Theorem 40.
Show that if f is a differentiable function with local linearization
E (a + h, b) E (a + h, b)
0 = lim √ = lim
h→0 h2 + k 2 h→0 h
f (a + h, b) − L(a + h, b)
= lim
h→0 h
f (a + h, b) − f (a, b) − mh
= lim
h→0 h
f (a + h, b) − f (a, b)
= lim −m
h→0 h
= fx (a, b) − m.
Therefore, if any of the partial derivatives fail to exist, then the function
cannot be differentiable.
Zooming in around (0, 0) does not make the graph look like a plane.
p
If we zoom in on the graph of the function f (x, y ) = x 2 + y 2 at the
origin, as shown in the figure, the sharp point remains; the graph never
flattens out to look like a plane. Near its vertex, the graph does not look
like it is well approximated (in any reasonable sense) by any plane.
Any plane passing through the point (0, 0, 0) has the form
L(x, y ) = mx + ny
Taking k = 0 gives
|h| − mh h
lim = 1 − m lim .
h→0 |h| h→0 |h|
This limit exists only if m = 0 for the same reason as before. But then the
value of the limit is 1 and not 0 as required. Thus, we again conclude f is
not differentiable.
In the above example the partial derivatives fx and fy did not exist at the
origin and this was sufficient to establish nondifferentiability there. We
might expect that if both partial derivatives do exist, then f is
differentiable. But the next example shows that this not necessarily true:
the existence of both partial derivatives at a point is not sufficient to
guarantee differentiability.
and similarly
fy (0, 0) = 0.
So, if there did exist a linear approximation near the origin, it would have
to be L(x, y ) = 0. But we can show that this choice of L(x, y ) does not
result in the small relative error that is required for differentiability.
If this limit exists, we get the same value no matter how h and k approach
0. Suppose we take k = h > 0. Then the limit becomes
Thus, this function is not differentiable at the origin, even though the
partial derivatives fx (0, 0) and fy (0, 0) exist. The figure confirms that near
the origin the graph of z = f (x, y ) is not well approximated by any plane.
One can show that f (x, y ) is not continuous at the origin. Show that the
partial derivatives fx (0, 0) and fy (0, 0) exist. Could f be differentiable at
(0, 0)?
f (h, 0) − f (0, 0) 1 0 0
fx (0, 0) = lim = lim . 2 2
= lim = 0,
h→0 h h→0 h h + 0 h→0 h
As we have seen from the last two examples, it is not enough that the
partial derivatives exist.
The class of functions with continuous partial derivatives is given the name
C 1.
2x 2y
fx = and fy = .
x2 + y2 x2 + y2
Since each of fx and fy is the quotient of continuous functions, the partial
derivatives are continuous everywhere except the origin (where the
denominators are zero). Thus, f is differentiable everywhere in its domain.
the ∂ 2 f /∂z 2 term from the previous equation, describes potentials and
steady-state temperature distributions in a plane (see the accompanying
figure). The plane (a) may be treated as a thin slice of the solid (b)
perpendicular to the z-axis.
∂2f ∂2f
(a). ∂x 2 + ∂y 2 = 0
∂2f ∂2f ∂2f
(b). ∂x 2
+ ∂y 2
+ ∂z 2
=0
∂2w 2
2∂ w
= c ,
∂t 2 ∂x 2
where w is the wave height, x is the distance variable, t is the time
variable, and c is the velocity with which the waves are propagated.
In our example, x is the distance across the ocean’s surface, but in other
applications, x might be the distance along a vibrating string, distance
through air (sound waves), or distance through space (light waves). The
number c varies with the medium and type of wave.
P. Sam Johnson Partial Derivatives August 30, 2019 115/117
Summary