Chapter 5

Multiple integrals

5.1. Definitions and examples

In this section we want to integrate a function of two variables, f (x, y). With
functions of one variable we integrated over an interval (i.e. a one-dimensional
space) and so it makes some sense then that when integrating a function of two
variables we will integrate over a region of R2 .
We will start out by assuming that the region in R2 is a rectangle which we will
denote as R = [a, b] × [c, d]. This means that the ranges for x and y are a ≤ x ≤ b
and c ≤ y ≤ d. Also, we will initially assume that f (x, y) ≥ 0 although this doesn’t
really have to be the case. Let’s start out with the graph of the surface S given by
graphing f (x, y) over the rectangle R.
what the volume of the region under S and above the xy-plane is. We will
approximate the volume much as we approximated the area in the integral. We
will first divide up a ≤ x ≤ b into n subintervals and divide up c ≤ y ≤ d into m
subintervals. This will divide up R into a series of smaller rectangles and from each
of these we will choose a point (xi , yj ). Here is a sketch of this set up.
Now, over each of these smaller rectangles we will construct a box whose height
is given by f (xi , yj ).
The base of each of the boxes is one of the rectangles from the previous image
and the height of the rectangles corresponds to the function value as noted above
in the text. The tops of all the rectangles forms an approximation to the surface S
above.
Each of the rectangles has a base area of ∆A and a height of f (xi , yj ) so the
volume of each of these boxes is f (xi , yj )∆A. The volume under the surface S is
then approximately,

n X
X m
f (xi , yj )∆A
i=1 j=1

47
48 5. Multiple integrals

To get a better estimation of the volume we will take n and m and

n X
X m
lim f (xi , yj )∆A
n,m→+∞
i=1 j=1

This looks a lot like the definition of the integral of a function of single variable.
In fact, this is also the definition of a double integral, or more exactly an integral
of a function of two variables over a rectangle:
ZZ Xn X m
f (x, y) dA = lim f (xi , yj )∆A
R n,m→+∞
i=1 j=1

Note the similarities and differences in the notation to single integrals. We have
two integrals to denote the fact that we are dealing with a two dimensional region
and we have a differential here as well. Note that the differential is dA instead of
the dx and dy that we’re used to seeing. Note as well that we don’t have limits
on the integrals in this notation. Instead we have the R written below the two
integrals to denote the region that we are integrating over.
As indicated above one interpretation of the double integral of f (x, y) over the
rectangle R is the volume under the function f (x, y) (and above the xy-plane). Or,
ZZ
Volume = f (x, y) dA
R

The following theorem tells us how to compute a double integral over a rectan-
gle.
Theorem 5.1 (Fubini). If f (x, y) is continuous on R = [a, b] × [c, d], then
ZZ Z bZ d Z dZ b
f (x, y) dA = f (x, y) dy dx = f (x, y) dx dy
R a c c a

These integrals are called iterated integrals. Note that there are in fact two
ways of computing a double integral over a rectangle and also notice that the
inner differential matches up with the limits on the inner integral and similarly
for the outer differential and limits. In other words, if the inner differential is dy
then the limits on the inner integral must be y limits of integration and if the outer
differential is dy then the limits on the outer integral must be y limits of integration.
Now, on some level this is just notation and doesn’t really tell us how to com-
pute the double integral. Let’s just take the first possibility above and change the
notation a little.
ZZ Z "Z b
#
d
f (x, y) dA = f (x, y) dy dx
R a c

We will compute the double integral by first computing

Z d
f (x, y) dy
c
and we compute this by holding x constant and integrating with respect to y as if
this were a single integral. This will give a function involving only x’s which we
can in turn integrate.
5.1. Definitions and examples 49

We’ve done a similar process with partial derivatives. To take the derivative of
a function with respect to y we treated the x’s as constants and differentiated with
respect to y as if it was a function of a single variable.
Double integrals work in the same manner. We think of all the x’s as constants
and integrate with respect to y or we think of all y’s as constants and integrate
with respect to x.
Example 5.2. Find
"Z #
ZZ Z b d
2
6xy dA = f (x, y) dy dx
R a c

where R = [2, 4] × [1, 2]. We use the first way:

ZZ Z 4 Z 2 
2 2
6xy dA = 6xy dy dx
R 2 1

and start computing

Z 2 Z 2
2
6xy dy = 6x y 2 dy = 14x
1 1
Then Z 4
14x dx = 7(16 − 4) = 84
2
so ZZ
6xy 2 dA = 84
R
The other method would produce the same value. 4
Example 5.3. Find ZZ
xexy dA
R
where R = [−1, 2] × [0, 1]. We use the first way:
ZZ Z 2 Z 1 
xy xy
xe dA = xe dy dx
R −1 0

and start computing

Z 1 Z 1
xexy dy = x exy dy = ex − 1
0 0
Then Z 2
(ex − 1) dx = e2 − 2 − e−1 − 1 = e2 − e−1 − e − 3
−1
so ZZ
6xy 2 dA = e2 − e−1 − e − 3
R
Integrating with respect to x first would be much more difficult (you need integra-
tion by parts) or even impossible. As with many other integrals, some techniques
are better than others. 4
50 5. Multiple integrals

If the region of integration is not rectangular, we can sometimes enclose it

between the graph of two functions (of x or of y). For example D might be defined
as D = {(x, y) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)} and the double integral for is defined
in terms of iterated integrals as follows:
ZZ Z b Z g2 (x)
f (x, y) dA = f (x, y) dy dx
D a g1 (x)

An analogous definition is made for the other case.

Example 5.4. Find
ZZ
4xy − y 3 dA
D
√
where D is the region bounded by y = x and y = x3 .
In this case we need to determine the two inequalities for√x and y. The two
lines intersect at (0, 0) and (1, 1) and x√3 is everywhere below x in this interval.
Thus we have 0 ≤ x ≤ 1 and x3 ≤ y ≤ x.
Then √
ZZ Z 1 Z x
3
4xy − y dA = 4xy − y 3 dy dx.
D 0 x3
Solving the inner integral we have
√
Z √x x
y4 7x2 x12
4xy − y 3 dy = 2xy 2 − = − 2x7 +
x3 4 x3 4 4
and plugging this into the outer integral we get
Z 1 2 1
7x x12 7x3 x8 x13 7 1 1 55
− 2x7 + dx = − + = − + = 4
0 4 4 12 4 52 0 12 4 52 156
Example 5.5. Find
ZZ
6x2 − 40y dA
D
where D is the triangle with vertices (0, 3), (1, 1), and (5, 3).
There are two ways of computing the double integral, that is, of describing D:
• considering D = D1 ∪ D2 with
1 1
D1 = {(x, y) : 0 ≤ x ≤ 1, −2x+3 ≤ y ≤ 3}, D2 = {(x, y) : 1 ≤ x ≤ 5, x+ ≤ y ≤ 3}
2 2
• writing
1 3
D = {(x, y) : − y + ≤ x ≤ 2y − 1, 1 ≤ y ≤ 3}.
2 2
Since the second has only one set, it may be computationally quicker.
ZZ Z 3 Z 2y−1
935
6x2 − 40y dA = 6x2 − 40y dx dy = −
D 1 − 21 y+ 23 3
4
Exercises on section 5.1 51

The double integral on a generic region D can be interpreted as the volume of

the solid that lies below the surface given by f (x, y) and above the region D in the
xy-plane:
ZZ
Volume = f (x, y) dA.
D

Moreover, if f (x, y) = 1 then the volume equals (numerically) the value of the area
of D so we night write:
ZZ
Area of D = dA.
D

Exercises on section 5.1

2x − 4y 3 dA, where R = [−5, 4] × [0, 3].
RR
Exercise 5.1.1. Evaluate R

ex/y dA, where D = {(x, y) : 1 ≤ y ≤ 2, y ≤ x ≤

RR
Exercise 5.1.2. Evaluate D
y 3 }.
RR
Exercise 5.1.3. Evaluate I = D (x + y) dA where D consists of all points (x, y)
√
such that 0 ≤ y ≤ 9 − x2 and 0 ≤ x ≤ 3.

Exercise 5.1.4. Suppose that two numbers are chosen uniformly from the interval
[0, 1]. We call the larger of the two numbers X and the smaller one Y . Their
joint pdf (probability density function) is fX,Y (x, y) = 2 if 0 ≤ y ≤ x ≤ 1 and 0
otherwise. Find the expected value of X and the marginal distribution of X. Note
that the expected value of a random variable Z with pdf f (x, y) is
ZZ
Zf (x, y) dx dy
R2

and the marginal distribution of X is

Z
fX (x) = f (x, y) dy
D

Exercise 5.1.5. At a certain restaurant, customers must wait an average of 10

minutes for a table. From the time they are seated until they have finished their
meal requires an additional 30 minutes, on average. What is the probability that
a customer will spend less than an hour at the restaurant, assuming that waiting
for a table and completing the meal are independent events? We assume that the
probability densities of the time for waiting and finishing are respectively
( (
0 x<0 0 x<0
pW (x) = 1 − x , pF (x) = 1 − x .
10 e 10 x ≥ 0 30 e 30 x ≥0

The two events are independent so the joint probability is the product of the two
probabilities
(
0 x < 0 or y < 0
p(x, y) = pW (x)pF (y) = 1 − x 1 − y
10 e 10
30 e 30 x, y≥0
52 5. Multiple integrals

5.2. Change of variables

In One-variable Calculus the substitution rule is
Z b Z d
f (g(x))g 0 (x) dx = f (u) du
a c

where u = g(x).
There are many reasons for changing variables. For example, one may want to
get an integral that is easier with the new variables, or to convert the region into a
nicer region to work with
We call the equations that define the change of variables a transformation.
Essentially, if we start with a set of variables x then we want to use another set of
variables u. For example, in two-variables, one may want to go from (x, y) to (u, v)
where x = x(u, v) and y = y(u, v). By changing the variables, the description of
the region changes as well.
2
Example 5.6. If R is the ellipse x2 + y36 = 1 and the transformation is x = u2 ,
2 2
y = 3v then R in the (u, v) system is u4 + v4 = 1, that is a circle with radius 2. 4

Similarly to the one-variable case, the rule for changing variables involves
derivatives. We start considering a change of variable in R2 :
x = g(u, v), y = h(u, v)
that is, a function f : R → R2 defined by
2
 
g(u, v)
f (u, v) = .
h(u, v)
We assume this function is differentiable and its differential, the Jacobian matrix,
is
 ∂g ∂g 
 ∂u ∂v 
Jf (u, v) = 
∂h ∂h

∂u ∂v
The factor we need to transform the integral is the absolute value of the determinant
of the Jacobian.
 ∂g ∂g 
 ∂u ∂v  ∂g ∂h ∂g ∂h
det   = −
∂h ∂h ∂u ∂v ∂v ∂u
∂u ∂v
We have the following theorem.
Theorem 5.7. Suppose that we want to integrate f (x, y) over the region R. Under
the transformation x = g(u, v), y = h(u, v) the region becomes S and the integral
becomes ZZ ZZ
f (x, y) dA = f (g(u, v), h(u, v))|det(Jf )| dA
R S

where dA is the “area” element in the coordinates (u, v).

5.2. Change of variables 53

RR
Example 5.8. Evaluate R x+y dA where R is the trapezoidal region with vertices
given by (0, 0), (5, 0), (5/2, 5/2) and (5/2, −5/2) using the transformation x =
2u + 3v and y = 2u − 3v.
First of all, we need to transform the xy-region R into the uv-region R0 . To do
so, we need to express (u.v) in terms of (x, y). This is easy from the transformation
above, since adding the two equations and subtracting them we get
x+y x−y
u= , v=
4 6
Then we have the following correspondence (xy → uv):
(0, 0) → (0, 0), (5, 0) → (5/4, 5/6), (5/2, 5/2) → (5/4, 0), (5/2, −5/2) → (0, 5/6)
and the region R0 is a rectangle in the uv-plane. From the same equations we
also get x + y = 4u. It remains only the determinant of the Jacobian. Since the
transformation is linear, the Jacobian is a constant matrix:
 
2 3
J=
2 −3
and the determinant is −12. Thus we have
ZZ ZZ Z 5/4 Z 5/6
x + y dA = 4u|−12| du dv = du 48u dv
R R0 0 0
Z 5/4 Z 5/6 Z 5/4
5
= 48u du dv = 48 · u du
0 0 6 0
25 125
= 40 · =
32 4
4
Example 5.9. Evaluate ZZ
2
+y 2
ex dx dy
D
where D is the unit disk centered at the origin.
To write the domain in terms of xy variables is quite difficult and, moreover,
2
the integral of et is impossible to write using elementary functions. The solution
here is to change the variables making both the domain and the function easy to
write.
The “obvious” choice is that of the so called polar coordinates. From basic
trigonometry we have
x = ρ cos θ, y = ρ sin θ
so that the region D is described by 0 ≤ ρ ≤ 1 and 0 ≤ θ ≤ 2π. Since x2 + y 2 = ρ2
2
the function to integrate becomes eρ .
The determinant of the Jacobian is
∂x ∂x
 
cos θ −ρ sin θ
!
 ∂ρ ∂θ 
det 
  = det = ρ cos2 θ + ρ sin2 θ = ρ
∂y ∂y  sin θ ρ cos θ
∂ρ ∂θ
54 5. Multiple integrals

Thus the above integral becomes

ZZ ZZ
2 2 2
ex +y dx dy = eρ ρ dρ dθ
D D0
Z 2π Z 1
2
= eρ ρ dρ dθ
0 0
2 1
2π
eρ
Z
= dθ
0 2
0
Z 2π
1
= (e − 1) dθ
2 0
= π(e − 1) 4

Exercises on section 5.2

RR
Exercise 5.2.1. Evaluate D 2xy dA where D is the portion of the region between
the circles of radius 2 and radius 5 centered at the origin that lies in the first
quadrant.
2 2
Exercise 5.2.2. If R is the region inside x4 + y36 = 1 determine the region R0 we
would get applying the transformation x = 2u, y = 6v to R and compute the area
of R
Exercise 5.2.3. Evaluate R xy 3 dA where R is the region bounded by xy = 1,
RR
v
xy = 3, y = 2 and y = 6 using the transformation x = 6u , y = 2u.
78 6. Answers

5.1.1 We have
ZZ Z 4 Z 3
3
2x − 4y dA = 2x − 4y 3 dy dx
R −5 0
Z 4
= (6x − 81) dx
−5
= 48 − 75 − 729 = −756
5.1.2 We have
ZZ Z 2 Z y3
x/y
e dA = ex/y dx dy
D 1 y
Z 2
2
= (y(ey − e)) dy
1
Z 2
2
= yey − ey dy
1
2
1 2 ey 2
= ey −
2 2 1
e4 − 4e e4 − 4e
= −0=
2 2
5.1.3 D consists of all points in the first quadrant inside the circle x2 + y 2 = 9. We
can then compute I as
Z 3 Z √9−x2 Z 3 p
9 − x2
dx (x + y) dy = x 9 − x2 + dx
0 0 0 2
Z 3 p Z 3
9 − x2
= x 9 − x2 + dx
0 0 2
Z 3 p
27 9
= x 9 − x2 dx + −
0 2 2
Z 3 √
t 27 9
= − dt + −
0 2 2 2
27 9
=9+ − = 18
2 2
5.1.4 Since X is larger than Y the region D where f (x, y) is not zero is the triangle
with vertices (0, 0), (1, 1) and (1, 0). We have
ZZ ZZ Z 1 Z x Z 1
f (x, y) dx dy = f (x, y) dx dy = dx 2 dy = 2x dx = 1
R2 D 0 0 0

which shows f (x, y) is indeed a pdf. Then we have

ZZ Z 1 Z x Z 1
2
E(X) = xf (x, y) dx dy = dx2 2 dy = 2x2 dx =
D 0 0 0 3
and Z x
fX (x, y) = 2 dy = 2x
0
6. Answers 79

5.1.5 The required probability is P (x + y ≤ 60) which defines a triangle T in R2

with vertices (0, 0), (60, 0) and (0, 60). Thus the probability is
ZZ Z 60 Z 60−x
1 x y
p(x, y) dx dy = dx e− 10 − 30 dy
T 300 0 0
Z 60 Z 60−x
1 x
− 10 y
= e dx e− 30 dy
300 0 0
Z 60
1 x
h 60−x
i
= e− 10 dx30 1 − e− 30
300 0
Z 60
1 x 2x+60
= e− 10 − e− 30 dx
10 0
 
1 60
− 10x
− 2x+60
= −10e + 15e 30
10 0
1 
−10e−6 + 15e−6 + 10 − 15e−2

=
10
1  −6
5e + 10 − 15e−2

=
10
= 0.5e−6 + 1 − 1.5e−2 = 0.7982 . . .
5.2.1 We use polar coordinates:
ZZ ZZ
2xy dx dy = 2ρ2 sin θ cos θρ dρ dθ
D D0
Z π/2 Z 5
= ρ3 sin(2θ) dρ dθ
0 2
Z π/2
609
= sin(2θ) dθ
4 0
π/2
609 cos(2θ)
= (− )
4 2 0
609 609
= ·1=
4 4
5.2.2 Under the given transformation, the equation becomes u2 + v 2 = 1. The area
of R is ZZ ZZ
dA = |J| dA = π
R R0
where J is the determinant of the Jacobian. Since the transformation is linear, the
Jacobian is  
2 0
0 6
and the determinant is 12. Thus,
ZZ ZZ ZZ
dA = |J| dA = 12 dA = 12π
R R0 R0

since the area of R0 in the (u, v) coordinates is that of a circle of radius 1.

80 6. Answers

5.2.3 From the equations of the transformation we get u = y and v = 6xu = 6xy.
Moreover, the Jacobian is
1
− 6uv 2
!
6u

0 2
1
and its determinant is 3u .
The transformation changes the boundaries of R into that of R0 (in the (u, v)
coordinates). y = 2 is transformed into u = 1 and y = 6 into u = 3. xy = 1 is
transformed into v = 3 and xy = 3 into v = 9. Then, R0 is a rectangle in the (u, v)
plane delimited by these straight lines.
Finally, we have
ZZ ZZ
3 v 1
xy dA = · (2u)3 · du dv
R R0 6u 3u
ZZ
v 1
= · 8u3 · du dv
0 6u 3u
ZRZ
4
= v · u du dv
9 R0
Z 3Z 9
4
= u du v dv
9 1
  3 
4 9 1 81 9
= − −
9 2 2 2 2
4
= · 4 · 36 = 64
9