Chapter 2 Partial Derivatives

2.1 First Order Partial Derivatives

2.2 Higher Order

2.3 Chain Rule

2.4 Partial Derivatives of Implicit Functions

2.5 Application of Partial Derivatives

2.5.1 Increments and Differential

2.5.2 Local Extrema

1
2.1 First order partial derivatives

Definition:
The partial derivative f x ( x, y ) of f ( x, y ) with respect
to x is obtained by holding y constant and
differentiating with respect to x. The partial
derivative f y ( x, y ) with respect to y is obtained by
holding x constant and differentiating with respect to
y.

f ( x  x , y )  f ( x , y )
f x ( x, y )  lim
x  0 x
f ( x , y  y )  f ( x , y )
f y ( x, y )  lim
y  0 y

f f
Other notations: f x  , fy 
x y

Partial derivatives at a point  xo , yo :

f f f
f x  xo , yo   x o , yo 
x  xo , y o  x x  xo , y  y o x

2
Example: Find f x (0,1) and f y (0,1) if
(i) f ( x, y )  2 x 2  3 x 2 y  y 2  x  y  1

sin(x  2 y )
(ii) f ( x, y ) 
(1  x)

For function of three variables 𝑥, 𝑦 and 𝑧, that is 𝑓(𝑥, 𝑦, 𝑧),

there are three partial derivatives:
𝑓𝑥 , 𝑓𝑦 , 𝑓𝑧
𝑓𝑥 is calculated by differentiating with respect to 𝑥, and
hold the other variables which are 𝑦 and 𝑧 as constants.

𝑓𝑦 is calculated by differentiating with respect to 𝑦, and

hold the other variables which are 𝑥 and 𝑧 as constants.

𝑓𝑧 is calculated by differentiating with respect to 𝑧, and

hold the other variables which are 𝑥 and 𝑦 as constants.

𝜕𝑤 1
Example: Given 𝑤 = 𝑥 sin(𝑥𝑦𝑧). Find at (1, , 𝜋)
𝜕𝑥 2

3
Partial Derivative as a Slope

Z Curve C Tangent line

𝑦0
Y

𝑥0

The tangent line to the curve C in the plane 𝑦 = 𝑦0 has

slope 𝑓𝑥 (𝑥, 𝑦0 ).

4
 Z Curve D Tangent line

𝑦0
Y

𝑥0

The tangent line of curve D in the plane 𝑥 = 𝑥0 has slope

𝑓𝑦 (𝑥0 , 𝑦).

5
Example:
Compute the slope of the tangent line to the surface
𝑓(𝑥, 𝑦) = 𝑥𝑦 3 + 𝑥 3 𝑦 at point 𝑃(1, −1, −2) in the direction
parallel to the 𝑥𝑧-plane.

Partial Derivative as a Rate of Change

2
A partial derivative is the rate of change of a multivariable
function when we allow only one of the variables to
change.

Example:
2
Point P moves along the intersection of plane x  with
3
the upper hemisphere 𝑧 = √1 − 𝑥 2 − 𝑦 2 . Find the rate of
 2 1 2
change of z with respect to y at point P , ,  .
 3 3 3

6
2.2 Higher order Partial Derivatives

If f ( x, y ) is a function with two variables, then f x ( x, y ) and

f y ( x, y ) are also functions with two variables. Partial
derivatives of f x ( x, y ) and f y ( x, y ) are the second-order
partial derivatives of f ( x, y ) and they are defined as

   f   2 f
f x  f xx     2
x x  x  x

   f   2 f
f x  f xy    
y y  x  yx

   f   2 f
f y  f yx    
x x  y  xy

   f   2 f
f y  f yy     2
y y  y  y

If z  f ( x, y ) , then
2 2 z
f ( x, y )  f xx ( x, y )   z xx
x 2
x 2

7
Example:
(i) Find f xx , f xy , f yx , f yy of f ( x, y )  x 2 ye y
(ii) If 𝑧 = sin(3𝑥 + 2𝑦), show that
3𝑧𝑦𝑦 − 2𝑧𝑥𝑥 = 6𝑧.

Theorem 2.1 (Euler Theorem)

If f , f x , f y , f xy and f yx are continuous and defined in an
open space that contains a point (a, b) , then
f xy (a, b)  f yx (a, b) .

8
2.3 Chain rules

Two intermediate variables and two independent

variables

Theorem 2.3 If z  f (u, v) , u  g ( x, y ) and v  h( x, y )

and the partial derivatives of f, g, and h exist, then

z z u z v
 
x u x v x

z z u z v
 
y u y v y

z z
example: By using the chain rules find and if
x y
z  u 2  v 2 with u  x 2 y and v  x 2 cos y .

9
Three intermediate variables and three independent
variables

Chain rule can be extended to function with any multiple

variables.

If w  f (u , v, r ) , u  g ( x, y, z ), v  h( x, y, z ) and
r  s ( x, y, z ) then
w w u w v w r
  
x u x v x r x

w w u w v w r
  
y u y v y r y

w w u w v w r
  
z u z v z r z

Example:

w w w
(i) Find , and if 𝑤 = 𝑟 2 + 𝑠 + 2𝑡,
x y z
r  x 2  y 2  z 2 , s  xyz and t  yz 2 .

10
Three intermediate variables and one independent
variable

Example:
dw
Find by using the chain rule if w  xyz with
dt
x  t 3 , y  t 2 and z  t 4

11
2.4 Implicit Function
Let F ( x, y )  0 with y  f (x) is defined implicitly.
We define F as:

w  F (u , y ) with u  x, y  f ( x)

u x
w
y x

Using the fact that u and y are functions with one variable,
we obtain
dw w du w dy
  .
dx u dx y dx
dw
w  F ( x, y )  F ( x, f ( x))  0 for all x, then  0.
dx
du dy
u  x and y  f (x) , then  1 and  f ' ( x)
dx dx

dw w du w dy
 
dx u dx y dx
w w
0  (1)  f ' ( x)
u y
dy w / u w / x F ( x, y )
 f ' ( x)     x
dx w / y w / y Fy ( x, y )

12
Theorem 2.4 If F ( x, y )  0 defines an implicit
function with 𝑓(𝑥), then

dy F ( x, y )
 x
dx Fy ( x, y )

dy
example: Find , if y  f (x) is defined as
dx
𝑦 4 + 3𝑦 − 4𝑥 3 = 5𝑥 + 1

Theorem 2.5 If F ( x, y, z )  0 implicitly defines a

differentiable function z as a function of two
variables x and y , then
z F ( x, y , z ) z Fy ( x, y, z )
 x , 
x Fz ( x, y, z ) y Fz ( x, y, z )

z z
example: Find and y if z as a function of two
x
variable x and y is implicitly defined as
𝑥𝑦 2 + 𝑧𝑦 = −𝑥 sin 𝑧

13
2.5 Applications of Partial Derivatives

2.5.1 Increment and Differential

Let z  f ( x, y ) . If x increases by x and y increases by y ,

then z will increase by z , where

z  f ( x  x , y  y )  f ( x , y )

f ( x  x, y )  f ( x, y ) f ( x, y  y )  f ( x, y )
fx  , fy 
x y
which implies
𝑓(𝑥 + 𝛿𝑥, 𝑦) − 𝑓(𝑥, 𝑦) ≈ 𝑓𝑥 (𝑥, 𝑦)𝛿𝑥

𝑓(𝑥, 𝑦 + 𝛿𝑦) − 𝑓(𝑥, 𝑦) ≈ 𝑓𝑦 (𝑥, 𝑦)𝛿𝑦

If we take
z   f ( x  x, y)  f ( x, y)   f ( x, y  y)  f ( x, y),
then
𝛿𝑧 ≈ 𝑓𝑥 (𝑥, 𝑦)𝛿𝑥 + 𝑓𝑦 (𝑥, 𝑦)𝛿𝑦

Definition:
Suppose 𝑧 = 𝑓(𝑥, 𝑦), and 𝑓𝑥 and 𝑓𝑦 exist. If 𝛿𝑥 and 𝛿𝑦 are
increments of x and y respectively, then an approximate
increment 𝛿𝑧 of z is given by

𝛿𝑧 = 𝑓𝑥 (𝑥, 𝑦)𝛿𝑥 + 𝑓𝑦 (𝑥, 𝑦)𝛿𝑦

14
If 𝛿𝑥 ⟶ 0 and 𝛿𝑦 ⟶ 0 (very small), then the increment in
𝛿𝑥 is denoted as 𝑑𝑥, increment in 𝛿𝑦 is denoted as 𝑑𝑦, and
the increment in 𝛿𝑧 is denoted as 𝑑𝑧. The 𝑑𝑧 is known as
the differential of 𝑧. Therefore
𝛿𝑧 ≈ 𝑑𝑧

And the total differential df of f is given by

𝑑𝑧 = 𝑑𝑓 = 𝑓𝑥 (𝑥, 𝑦)𝑑𝑥 + 𝑓𝑦 (𝑥, 𝑦)𝑑𝑦

Definition 2.3 Let w  f ( x, y, z ) and f x ( x, y, z ) , f y ( x, y, z )

and f z ( x, y, z ) exist. If x increases by x , y increases by y
, and z increases by z , then the small increment in w is

𝛿𝑤 = 𝑓𝑥 (𝑥, 𝑦, 𝑧)𝛿𝑥 + 𝑓𝑦 (𝑥, 𝑦, 𝑧)𝛿𝑦+𝑓𝑧 (𝑥, 𝑦, 𝑧)𝛿𝑧

and the total differential of f is given

𝑑𝑤 = 𝑑𝑓 = 𝑓𝑥 (𝑥, 𝑦, 𝑧)𝑑𝑥 + 𝑓𝑦 (𝑥, 𝑦, 𝑧)𝑑𝑦 + 𝑓𝑧 (𝑥, 𝑦, 𝑧)𝑑𝑧

15
 .
Example:
(i) If z  f ( x, y )  3 x 2  xy , use a total differential
to approximate the change in f ( x, y ) as ( x, y )
varies from the point (1,2) to the point 1.01,1.98

(ii) Let the measurements of a cuboid vary from 9

cm, 6 cm and 4 cm to 9.02, 5.97 and 4.01 cm.
Use a total differential to approximate the change
in volume.

We can use differential to estimate change and estimating

error

Let 𝑧 = 𝑓(𝑥, 𝑦), then 𝛿𝑧 ≈ 𝑓𝑥 (𝑥, 𝑦)𝛿𝑥 + 𝑓𝑦 (𝑥, 𝑦)𝛿𝑦

 𝛿𝑧 is the error in calculating z

 |𝛿𝑧| is the maximum error

𝛿𝑧
 is the relative error
𝑧

𝛿𝑧
 × 100% is the percentage error
𝑧

𝛿𝑧
 | 𝑧 | × 100% maximum percentage error

16
2.5.2 Local Extrema

Extrema refers to either maximum or minimum.

Definition 2.5
Suppose 𝑧 = 𝑓(𝑥, 𝑦) is a function of two variables defined
in a domain D, and that (𝑎, 𝑏) is in D.

(i) f (a, b) is a local maximum of the function f, if

there exist region 𝑅 ⊂ 𝐷 , such that (𝑎, 𝑏) in R
and 𝑓(𝑎, 𝑏) ≥ 𝑓(𝑥, 𝑦) for every (𝑥, 𝑦) in R.

(ii) f (a, b) is a local minimum of the function f, if

there exist region 𝑅 ⊂ 𝐷 , such that (𝑎, 𝑏) in R
and 𝑓(𝑎, 𝑏) ≤ 𝑓(𝑥, 𝑦) for every (𝑥, 𝑦) in R.

Theorem 2.6 Suppose f (a, b) is a local maximum

or local minimum of 𝑓(𝑥, 𝑦). If f x and f y exist at
(a, b) , then
f x (a, b)  0 and f y (a, b)  0

note: The reverse statement of the theorem is not

necessarily true.

17
Theorem 2.7 Second Derivatives Test

A pair (𝑎, 𝑏) such that 𝑓𝑥 (𝑎, 𝑏) = 0 and 𝑓𝑦 (𝑎, 𝑏) = 0 is

called a critical point or stationary point.

Suppose that 𝑓(𝑥, 𝑦) is a function of two variables

that has continuous second partial derivatives on a
region R and (𝑎, 𝑏) is a critical point in R.

Let
2
𝐺 (𝑎, 𝑏) = 𝑓𝑥𝑥 (𝑎, 𝑏)𝑓𝑦𝑦 (𝑎, 𝑏) − [𝑓𝑥𝑦 (𝑎, 𝑏)]

𝑓𝑥𝑥 (𝑎, 𝑏) 𝑓𝑥𝑦 (𝑎, 𝑏)

𝐺(𝑎, 𝑏) = | |
𝑓𝑦𝑥 (𝑎, 𝑏) 𝑓𝑦𝑦 (𝑎, 𝑏)

Then

(a) 𝑓(𝑎, 𝑏) is a local maximum if 𝐺 (𝑎, 𝑏) > 0 and

𝑓𝑥𝑥 (𝑎, 𝑏) < 0.

(b) 𝑓(𝑎, 𝑏) is a local minimum if 𝐺 (𝑎, 𝑏) > 0 and

𝑓𝑥𝑥 (𝑎, 𝑏) > 0.

(c) (𝑎, 𝑏, 𝑓(𝑎, 𝑏)) is a saddle point if 𝐺 (𝑎, 𝑏) < 0.

(d) No conclusion can be made if 𝐺 (𝑎, 𝑏) = 0.

18
Examples: Find the critical point(s) of the following
functions and determine the nature of those points.
.
(i) 𝑓(𝑥, 𝑦) = 2𝑦 3 − 6𝑥𝑦 − 𝑥 2
(ii) 𝑓(𝑥, 𝑦) = 𝑥 2 − 𝑦 2 + 2𝑥 + 6𝑦 − 4

19
2.6 Absolute Extrema

Definition 2.6 The absolute extremum for a function is the

extremum values of the function on any given region.

Definition:
1. A region in ℝ2 is called closed if it includes its
boundary. A region is called open if it doesn’t include
any of its boundary points.
2. A region in ℝ2 is called bounded if it can be
completely contained in a disk/rectangle. In other
words, a region will be bounded if it is finite.

Theorem 2.8 (Extreme Value Theorem)

If a function is continuous on a region that is closed and
bounded, then it has both absolute maximum and absolute
minimum in the region. Further, the absolute extrema must
occur either at a critical point in the region, or on the
boundary of the region.

20
Steps to find the absolute extrema of a function on a closed
and bounded region.
1. Find the critical points of the function in the region,
and compute the values of 𝑓 at the critical points.

2. Find the extremum values of 𝑓 on the boundary.

3. Compare the values of 𝑓at the critical points with the

maximum and the minimum values of 𝑓 on the
boundary of the region. The highest value of 𝑓 is the
absolute maximum and the lowest value of 𝑓 is the
absolute minimum.
Note: in this case, no need to use the second derivatives test

Open and unbounded Region

Theorem 2.9 Suppose f(x,y)is continuous in a region R.
(i) If f(x,y) assumes only a local maximum and there
is no local minimum and no saddle point in the
region, then the local maximum is the absolute
maximum in the region.
(ii) If f(x,y) assumes only a local minimum and there
is no local maximum and no saddle point in the
region, then the local minimum is the absolute
minimum in the region.

21