1.

2 Derivatives of multivariable functions

In first year we started to extend the basic concepts, such as convergence, limits, continuity and deriva-
tives in calculus for single variable functions to calculus of several variables. For example, we can use
Newton’s law of cooling to model the temperature T = T (t) of a cup of coffee over time t. However,
we could model the temperature over time throughout the volume of the cup of coffee, thus requiring
four independent variables; x, y, z, t. Therefore, we would have T = T (x, y, z, t). In this section we
will continue that journey into multivariable calculus.

1.2.1 Partial derivatives

df
Recall that for a function of a single variable, say f = f (x), the derivative of f is the function
dx
 
df f (x + ∆x) − f (x)
= lim .
dx ∆x−→0 ∆x

Recall that occasionally this derivative may be denoted as f ′ (x), which makes it explicit that the
derivative is a function of x itself.
For a function of several variables, say f = f (x, y, z, t), consider the first order partial derivatives of f
with respect to any one of the independent variables:

▶ The first order partial derivative of f with respect to x is the function

 
∂f f (x + ∆x, y, z, t) − f (x, y, z, t)
= lim
∂x ∆x−→0 ∆x

where y, z, t have all been held constant.

▶ The first order partial derivative of f with respect to y is the function
 
∂f f (x, y + ∆y, z, t) − f (x, y, z, t)
= lim
∂y ∆y−→0 ∆y

where x, z, t have all been held constant.

▶ The first order partial derivative of f with respect to z is the function

 
∂f f (x, y, z + ∆z, t) − f (x, y, z, t)
= lim
∂z ∆z−→0 ∆z

where x, y, t have all been held constant.

▶ The first order partial derivative of f with respect to t is the function
 
∂f f (x, y, z, t + ∆t) − f (x, y, z, t)
= lim
∂t ∆z−→0 ∆t

where x, y, z have all been held constant.

ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.1
Consider the four-variable function

f (x, y, z, t) = x2 + yz t

then the first order partial derivatives of f are

∂f ∂f ∂f ∂f
= 2xt, = zt, = yt, = x2 + yz
∂x ∂y ∂z ∂t

1.2.2 The chain rule

Recall that in single-variable calculus, if z = f (y) and y = g(x) are two functions of a single variable,
then the derivative of the composite function f ◦ g(x) = f (g(x)) is

dz df dy
=
dx dy dx
It is important to remember; we are not cancelling out the “dy” differentials, since derivatives are not
fractions. This is more obvious when written in Newton’s notation

z ′ (x) = f ′ (y) g ′ (x)

We can extend the chain rule to a function of n-variables, where each variable depends upon m-variables.

Definition: Chain rule

Let f (x1 , x2 , . . . , xn ) be a differentiable function dependent upon n variables, where each vari-
able xk (u1 , u2 , . . . , um ) is a differentiable function dependent upon m variables. The first order
derivative of f with respect to uk is given by the chain rule
n  
∂f X ∂f ∂xi
=
∂uk i=1
∂xi ∂uk

that is,
∂f ∂f ∂x1 ∂f ∂x2 ∂f ∂xn
= + + ··· +
∂uk ∂x1 ∂uk ∂x2 ∂uk ∂xn ∂uk

▶ again; we are not cancelling out the “∂xi ” symbol, since derivatives are not fractions.

▶ if a specific function depends upon only one variable then the partial derivative is just an ordinary
∂  
derivative, for example x3 (u1 ) only depends upon one variable u1 , thus we write x3 (u1 ) =
∂u1
dx3
.
du1

12
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.2
2
Let z = f (y) = y 2 and y = g(x) = sin(x), then the composite function f ◦g(x) is z = f (g(x)) = (sin(x)) .
By chain rule the derivative of z with respect to x is

dz df dy
=
dx dy dx
d  2 d  
= y sin(x)
dy dx
= (2y) (cos(x))
= 2 sin(x) cos(x) .

Now, consider a function of three independent variables w = f (x, y, z) and let x = x(t), y = y(t),
z = z(t) be functions of a single variable t. The composite function w = f (x(t) , y(t) , z(t)) = F (t) is
a function of the (single) independent variable t. Using the chain rule, the derivative of the function
w = F (t) with respect to t is
dF ∂f dx ∂f dy ∂f dz
= + + .
dt ∂x dt ∂y dt ∂z dt

Example 1.2.3
1
Let w = f (x, y, z) = xy 2 + x2 z, and let x(t) = t2 , y(t) = t, z(t) = then applying the chain rule, the
t
derivative of w with respect to t is
dw ∂f dx ∂f dy ∂f dz
= + +
dt ∂x dt ∂y dt ∂z dt
∂ 2  d  ∂ 2  d  ∂ 2  d 1
2 2 2 2
= xy + x z t + xy + x z t + xy + x z
∂x dt ∂y dt ∂z dt t
 
1
= y 2 + 2xz (2t) + (2xy) (1) + x2 − 2



t
   
2   1 
2 2 1
(2t) + 2 t (t) (1) + t2
2




= (t) + 2 t − 2
t t
= 4t3 + 3t2 .

Alternatively, the composite function is

w = f (x(t) , y(t) , z(t))

 
2

2
1
2 2
= t (t) + t
t
= t4 + t3

and then the derivative of w with respect to t is

dw d 4 
= t + t3
dt dt
= 4t3 + 3t2 .

The chain rule for functions of several variables can get more complicated. Consider, for example, w =
f (x, y, z), which is a function of three independent variables, and assume that x = x(u, v), y = y(u, v)
and z = z(u, v), are functions of two independent variables u and v. Hence the composite function

13
ENG2005 Advanced Engineering Mathematics Monash University

w = f (x(u, v) , y(u, v) , z(u, v)) = F (u, v) is a function of the independent variables u and v. Using the
chain rule, the partial derivative of the function w = F (u, v) with respect to u is
∂F ∂f ∂x ∂f ∂y ∂f ∂z
= + +
∂u ∂x ∂u ∂y ∂u ∂z ∂u
and the partial derivative of the function w = F (u, v) with respect to v is
∂F ∂f ∂x ∂f ∂y ∂f ∂z
= + +
∂v ∂x ∂v ∂y ∂v ∂z ∂v

Example 1.2.4
Let w = f (x, y) = x2 − y 2 , and x(r, θ) = r cos(θ) and y(r, θ) = r sin(θ). Using the chain rule, the partial
derivative of the function w = F (r, θ) with respect to r is
∂F ∂f ∂x ∂f ∂y
= +
∂r ∂x ∂r ∂y ∂r
∂ 2  ∂  ∂ 2  ∂ 
= x − y2 r cos(θ) + x − y2 r sin(θ)
∂x ∂r ∂y ∂r
= (2x) (cos(θ)) + (−2y) (sin(θ))
= (2r cos(θ)) (cos(θ)) + (−2r sin(θ)) (sin(θ))
= 2r cos2 (θ) − sin2 (θ)

= 2r cos(2θ)
and the partial derivative of the function w = F (r, θ) with respect to θ is
∂F ∂f ∂x ∂f ∂y
= +
∂θ ∂x ∂θ ∂y ∂θ
∂ 2  ∂  ∂ 2  ∂ 
= x − y2 r cos(θ) + x − y2 r sin(θ)
∂x ∂θ ∂y ∂θ
= (2x) (−r sin(θ)) + (−2y) (r cos(θ))
= (2r cos(θ)) (−r sin(θ)) + (−2r sin(θ)) (r cos(θ))
= −4r2 cos(θ) sin(θ)
= −2r2 sin(2θ)

Alternatively, the composite function is

w = f (x(r, θ) , y(r, θ))
2 2
= (r cos(θ)) − (r sin(θ))
= r2 cos2 (θ) − sin2 (θ)

= r2 cos(2θ)
Therefore, the derivative of w = F (r, θ) with respect to r is
∂w ∂ 2 
= r cos(2θ)
∂r ∂r
= 2r cos(2θ)
and the derivative of w = F (r, θ) with respect to θ is
∂w ∂ 2 
= r cos(2θ)
∂θ ∂θ
= −2r2 sin(2θ)

The chain rule allows us to determine how a partial derivative in one coordinate system, such as
Cartesian coordinates, is described in another coordinate system, such as polar coordinates.

14
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.5
Find the first order partial derivatives of a function u(x, y) with respect to x and y in terms of the polar
coordinates r and θ.
The polar coordinates are defined by
x(r, θ) = r cos(θ) and y(r, θ) = r sin(θ)
where
y
r2 = x2 + y 2 and tan(θ) =
x
Using the chain rule, the first order partial derivative of u with respect to x is given by
∂u ∂u ∂r ∂u ∂θ
= +
∂x ∂r ∂x ∂θ ∂x
and the first order partial derivative of u with respect to y is given by
∂u ∂u ∂r ∂u ∂θ
= +
∂y ∂r ∂y ∂θ ∂y

Applying implicit differentiation we can find the first order partial derivative of r with respect to x and
the first order partial derivative of r with respect to y.
∂  2 ∂ 2  ∂  2 ∂ 2 
r = x + y2 r = x + y2
∂x ∂x ∂y ∂y
∂r ∂r
2r = 2x 2r = 2y
∂x ∂y
∂r x ∂r y
= =
∂x r ∂y r
∂r ∂r
= cos(θ) = sin(θ)
∂x ∂y
Applying implicit differentiation we can find the first order partial derivative of θ with respect to x and
the first order partial derivative of θ with respect to y.
∂  ∂ y ∂  ∂ y
tan(θ) = tan(θ) =
∂x ∂x x ∂y ∂y x
∂θ y ∂θ 1
sec2 (θ) =− 2 sec2 (θ) =
∂x x ∂y x
∂θ y ∂θ 1
=− 2 2 =
∂x x sec (θ) ∂y x sec2 (θ)
∂θ y ∂θ 1
=− 2 =
x 1 + tan2 (θ)


2


∂x x 1 + tan (θ) ∂y
∂θ y ∂θ 1
=− 
2  = 
2 
∂x x2 1 + y ∂y x 1+ y
x x
∂θ y ∂θ x
=− 2 = 2
∂x r ∂y r
∂θ sin(θ) ∂θ cos(θ)
=− =
∂x r ∂y r

Hence, by using the chain rule, the first order partial derivative of u with respect to x is given by
∂u ∂u sin(θ) ∂u
= cos(θ) −
∂x ∂r r ∂θ
and the first order partial derivative of u with respect to y is given by
∂u ∂u cos(θ) ∂u
= sin(θ) +
∂y ∂r r ∂θ

15
ENG2005 Advanced Engineering Mathematics Monash University

1.2.3 Scalar Fields

Definition: Scalar field

A scalar field (or scalar function) is a function f defined in a region of either two-dimensional
space (R2 ) or three-dimensional space (R3 ) which assigns a unique scalar value (that is, a real
number) to each point P in the domain of f

f = f (P ) = f (r)

where r denotes the position vector from the origin to the point P .

Any point P in R3 with Cartesian coordinates (x, y, z) has a corresponding position vector r = xi +
yj + zk, and then the scalar field can be written as

f = f (P ) = f (r) = f (x, y, z) .

Example 1.2.6
Atmospheric pressure at points on the surface of the Earth is a scalar field.

Example 1.2.7
The temperature at points throughout the volume of the Sun is a scalar field.

Example 1.2.8
The distance function p
d(x, y, z) = x2 + y 2 + z 2
describing the distance from the origin to any three-dimensional point P : (x, y, z) is a scalar field.

Example 1.2.9
The function
f (x, y) = x2 + y 2
is a (two-dimensional) scalar field.

Definition: Level curves

Let f (x, y) be a two-dimensional scalar function. A level curve (or contour) of height c is the
set of all points (x, y) such that f (x, y) = c.

Geometrically, a level curve of height c is the intersection of the z = c plane with the surface z = f (x, y).

16
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.10
Consider the scalar field
f (x, y) = x2 + y 2 .
The surface represented by this function is
The surface f(x,y) = x2+y2

50

40

30
z

20

10

0
5
5
0
0

y −5 −5
x

The level curves of f (x, y) are concentric circles x2 + y 2 = c for c ≥ 0.

Level Curves for x2+y2 = c for values c = 1,4,9,16,25

5
25 25
4 16

3 9
16
25

2
9

25
4

1
16

1
4

0
y

9
25

1
9

−1
16

−2 4
25

16
−3 9

−4 25 16

−5 25
−6 −4 −2 0 2 4 6
x

17
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.11
Consider the scalar function
f (x, y) = x2 − y 2 .
The surface represented by this function is
The surface f(x,y) = x2−y2

30

20

10

0
z

−10

−20

−30
5
5
0
0

y −5 −5
x

The level curves of f (x, y) are rectangular hyperbolae x2 − y 2 = c for any c ∈ R.

Level Curves for x2−y2 = c for values c = −9, −4,−1,1,4,9
5
−4
−9 −4 1
−
4 1 − −9 1
1

−9
9

3
−4 −4 4
4

2 −1
−1
1
1

0
y

−1
4

−1
−2 1 −1 1
−4 −4
−3 −9

−1
9

−4 4 −9 4
−9 −1 1
1 −4 −4
−5
−5 −4 −3 −2 −1 0 1 2 3 4 5
x

Definition: Level surface

Let f (x, y, z) be a three-dimensional scalar function. A level surface of value c is the set of all
points (x, y, z) such that f (x, y, z) = c.

18
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.2.12
Consider the scalar field
f (x, y, z) = x2 + y 2 − z.
The level surfaces of f (x, y, z) are nested paraboloids x2 + y 2 − z = c for any c ∈ R.

Level surfaces for x2+y2−z=c for values c = −2,0,2

2
z

−1

−2
−2 −2

0 0

2 2
x
y

19
 1.3 Double Integrals

1.3.1 Double Integrals

For the moment, we will assume that the region of integration is in the xy-plane, that is, a region of
the two-dimensional plane. To define a region in the plane, we need to be able to define the “edges” of
the region.

Example 1.3.1
The rectangular region in the xy-plane

which can be described as

(x, y) ∈ R2 : −1 ≤ x ≤ 2, 0 ≤ y ≤ 1 .

R=

Example 1.3.2
The region bounded by the curves y = x2 and x = y 2 in the xy-plane

could be described as √
(x, y) ∈ R2 : 0 ≤ x ≤ 1, x2 ≤ y ≤

R= x
or equivalently as
√
(x, y) ∈ R2 : 0 ≤ y ≤ 1, y 2 ≤ x ≤

R= y
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.3.3
The region bounded by the lines x = 0, y = 4 and the curve y = x2 in the xy-plane

could be described as
(x, y) ∈ R2 : 0 ≤ x ≤ 2, x2 ≤ y ≤ 4

R=
or equivalently as
√
(x, y) ∈ R2 : 0 ≤ y ≤ 4, 0 ≤ x ≤

R= y

Example 1.3.4
The region bounded by the lines y = x, y = 2x and x = 2 in the xy-plane

could be described as
(x, y) ∈ R2 : 0 ≤ x ≤ 2, x ≤ y ≤ 2x

R=
or equivalently as R = R1 ∪ R2 where R is split into two regions
n y o n y o
R1 = (x, y) ∈ R2 : 0 ≤ y ≤ 2, ≤ x ≤ y and R2 = (x, y) ∈ R2 : 2 ≤ y ≤ 4, ≤x≤2
2 2

21
ENG2005 Advanced Engineering Mathematics Monash University

Definition: Double Integrals

Let R be a bounded region in the xy-plane, which is completely contained within a rectangular
region D. We divide D into M × N uniform small rectangles each with area ∆A = ∆x∆y and
centre points (xi , yj ). Then define the function
(
f (xi , yj ), if (xi , yj ) is in R,
F (xi , yj ) = .
0 otherwise.

Then the double integral of f (x, y) over the region R is given by

  
ZZ   M X
X N
f (x, y) dA = lim  lim  F (xi , yj ) ∆A ,
R M →∞ N →∞
i=1 j=1

where the limits exist as both ∆x → 0 and ∆y → 0.

1.3.2 The Order of Integration for Double Integrals

Theoretically, if the integrand and the region of integration are “suitably behaved”, then the order of
integration does not matter. However, in practice the order of integration often matters in deriving an
analytic solution.
In Cartesian coordinates, a double integral may be expressed as either the iterated integrals
ZZ  Z x=b Z y=g2(x)  !
 
f (x, y) dA = f (x, y) dy dx
R x=a y=g1(x)

where g1 and g2 are single-variable functions of x, or the iterated integrals

ZZ  Z y=d Z x=h2(y)  !
 
f (x, y) dA = f (x, y) dx dy
R y=c x=h1(y)

where h1 and h2 are single-variable functions of y.

When deciding what order of integration to try, you should always first try sketch the bounded region.

Example 1.3.5
Consider the double integral ZZ  
x + 2y dA
R
over the region bounded by the curves y = x2 and x = y 2 in the xy-plane. We could evaluate this
double integral by first integrating over y
√
R = (x, y) ∈ R2 : 0 ≤ x ≤ 1, x2 ≤ y ≤ x


and then the double integral becomes

√ !
ZZ   Z x=1 Z y= x  
x + 2y dA = x + 2y dy dx
R x=0 y=x2
Z x=1 h iy=√x 
= xy + y 2 dx
x=0 y=x2
Z x=1  
3
= x 2 + x − x3 − x4 dx
x=0
9
= .
20

22
ENG2005 Advanced Engineering Mathematics Monash University

For this double integral, the order of integration is irrelevant, the double integral could be evaluated by
integrating over x first
√
R = (x, y) ∈ R2 : 0 ≤ y ≤ 1, y 2 ≤ x ≤ y


and then the double integral becomes

√ !
ZZ   Z y=1 Z x= y  
x + 2y dA = x + 2y dx dy
R y=0 x=y 2
Z y=1 h
1 ix=√y 
= x2 + 2xy dy
y=0 2 x=y 2
Z y=1  
1 3 1
= y + 2y 2 − y 4 − 2y 3 dx
y=0 2 2
9
= .
20

Example 1.3.6
Consider the double integral ZZ
(4 − x − y) dA
R
over the region bounded by the lines y = x, y = 2x and x = 2 in the xy-plane. We could evaluate this
double integral by first integrating over y

R = (x, y) ∈ R2 : 0 ≤ x ≤ 2, x ≤ y ≤ 2x


and then the double integral becomes

ZZ Z 2 Z 2x   
(4 − x − y) dA = 4 − x − y dy dx
R 0 x
Z 2  2x !
1 2
= 4y − xy − y dx
0 2 x
Z 2 
5 2
= 4x − x dx
0 2
4
= .
3

We could evaluate this double integral by changing the order of integration to first integrate over x
which would require splitting R into two regions
n y o n y o
R1 = (x, y) ∈ R2 : 0 ≤ y ≤ 2, ≤ x ≤ y and R2 = (x, y) ∈ R2 : 2 ≤ y ≤ 4, ≤x≤2
2 2
and then the double integral becomes
ZZ ZZ   ZZ  
(4 − x − y) dA = 4 − x − y dx dy + 4 − x − y dx dy
R R1 R2
Z 2 Z y ! Z 4 Z 2 !
 
= 4 − x − y dx dy + 4 − x − y dx dy.
y y
0 2 2 2

This would involve a lot more work and is left as an exercise.

23
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.3.7
Consider the double integral ZZ  
xy dA
R
over the region R

We could evaluate this double integral by first integrating over y which would require splitting R into
two regions R1 ∪ R2 where
√ √
R1 = (x, y) ∈ R2 : 0 ≤ x ≤ 1, − x ≤ y ≤ x


and √
R2 = (x, y) ∈ R2 : 1 ≤ x ≤ 4, x − 2 ≤ y ≤ x


and then the double integral becomes

ZZ ZZ   ZZ  
(xy) dA = xy dy dx + xy dy dx
R R1 R2
Z 1 Z √x   ! Z 4 Z √x   !
= √
xy dy dx + xy dy dx.
0 − x 1 x−2

Instead we can evaluate the double integral by changing the order of integration to first integrate over
x. For this situation the region can be described as

R = (x, y) ∈ R2 : −1 ≤ y ≤ 2, y 2 ≤ x ≤ y + 2 .


and then the double integral becomes

ZZ   Z 2 Z y+2   
xy dA = xy dx dy
R −1 y2
Z 2   !
y+2
1 2
= x y dy
−1 2 y2

1 2 3
Z 
= y + 4y 2 + 4y − y 5 dy
2 −1
45
= .
8

24
ENG2005 Advanced Engineering Mathematics Monash University

Example 1.3.8
Consider the double integral ZZ  
sin(y)
dA
R y
over the region R

If we attempt to integrate with respect to y first then the region is described as

R = (x, y) ∈ R2 : 0 ≤ x ≤ 1, x ≤ y ≤ 1


and the double integral becomes

ZZ   Z 1 Z 1   
sin(y) sin(y)
dA = dy dx.
R y 0 x y

sin(y)
However we would have to integrate with respect to y, therefore it is best to change the order
y
of integration. We will evaluate this double integral by first integrating with respect to x, where the
region is described as
R = (x, y) ∈ R2 : 0 ≤ y ≤ 1, 0 ≤ x ≤ y


and then the double integral becomes

ZZ   Z 1 Z y   
sin(y) sin(y)
dA = dx dy
R y 0 0 y
Z 1  y 
sin(y)
= x dy
0 y 0
Z 1 
= sin(y) dy
0
= 1 − cos(1) .

25