22.

Double integrals
Denition 22.1. Let R = [a, b] [c, d] R
2
be a rectangle in the
plane. A partition P of R is a pair of sequences:
a = x
0
< x
1
< < x
n
= b
c = y
0
< y
1
< < y
n
= d.
The mesh of P is
m(P) = max{ x
i
x
i1
, y
i
y
i1
| 1 i k }.
Now suppose we are given a function
f : R R
Pick
c
ij
R
ij
= [x
i1
, x
i
] [y
j1
, y
j
].
Denition 22.2. The sum
S =
n

i=1
n

j=1
f(c
ij
)(x
i
x
i1
)(y
j
y
j1
),
is called a Riemann sum.
We will use the short hand notation
x
i
= x
i
x
i1
and y
j
= y
j
y
j1
.
Denition 22.3. The function f : R R is called integrable, with
integral I, if for every > 0, we may nd a > 0 such that for every
mesh P whose mesh size is less than , we have
|I S| < ,
where S is any Riemann sum associated to P.
We write
__
R
f(x, y) dx dy = I,
to mean that f is integrable with integral I.
We use a sneaky trick to integrate over regions other than rectangles.
Suppose that D is a bounded subset of the plane. Then we can nd a
rectangle R which completely contains D.
Denition 22.4. The indicator function of D R is the function
i
D
: R R,
1
given by
i
D
(x) =
_
1 if x D
0 if x / D.
If i
D
is integrable, then we say that the area of D is the integral
__
R
i
D
dx dy.
If i
D
is not integrable, then D does not have an area.
Example 22.5. Let
D = { (x, y) [0, 1] [0, 1] | x, y Q}.
Then D does not have an area.
Denition 22.6. If f : D R is a function and D is bounded, then
pick D R R
2
a rectangle. Dene

f : R R,
by the rule

f(x) =
_
f(x) if x D
0 otherwise.
We say that f is integrable over D if

f is integrable over R. In this
case
__
D
f(x, y) dx dy =
__
R

f(x, y) dx dy.
Proposition 22.7. Let D R
2
be a bounded subset and let f : D
R and g : D R be two integrable functions. Let be a scalar.
Then
(1) f + g is integrable over D and
__
D
f(x, y) + g(x, y) dx dy =
__
D
f(x, y) dx dy +
__
D
g(x, y) dx dy.
(2) f is integrable over D and
__
D
f(x, y) dx dy =
__
D
f(x, y) dx dy.
(3) If f(x, y) g(x, y) for any (x, y) D, then
__
D
f(x, y) dx dy
__
D
g(x, y) dx dy.
2
(4) |f| is integrable over D and
|
__
D
f(x, y) dx dy|
__
D
|f(x, y)| dx dy.
It is straightforward to integrate continuous functions over regions
of three special types:
Denition 22.8. A bounded subset D R
2
is an elementary region
if it is one of three types:
Type 1:
D = { (x, y) R
2
| a x b, (x) y (x) },
where : [a, b] R and : [a, b] R are continuous functions.
Type 2:
D = { (x, y) R
2
| c y d, (y) x (y) },
where : [c, d] R and : [c, d] R are continuous functions.
Type 3: D is both type 1 and 2.
Theorem 22.9. Let D R
2
be an elementary region and let f : D
R be a continuous function.
Then
(1) If D is of type 1, then
__
D
f(x, y) dx dy =
_
b
a
_
_
(x)
(x)
f(x, y) dy
_
dx.
(2) If D if of type 2, then
__
D
f(x, y) dx dy =
_
d
c
_
_
(y)
(y)
f(x, y) dx
_
dy.
Example 22.10. Let D be the region bounded by the lines x = 0,
y = 4 and the parabola y = x
2
. Let f : D R be the function given
by f(x, y) = x
2
+ y
2
.
3
If we view D as a region of type 1, then we get
__
D
f(x, y) dx dy =
_
2
0
__
4
x
2
x
2
+ y
2
dy
_
dx
=
_
2
0
_
x
2
y +
y
3
3
_
4
x
2
dx
=
_
2
0
4x
2
+
2
6
3
x
4

x
6
3
dx
=
_
4x
3
3
+
2
6
x
3

x
5
5

x
7
3 7
_
2
0
=
2
5
3
+
2
7
3

2
5
5

2
7
3 7
=
2
6
3 5
+
2
8
7
= 2
6
_
1
3 5
+
2
2
7
_
.
On the other hand, if we view D as a region of type 2, then we get
__
D
f(x, y) dx dy =
_
4
0
__

y
0
x
2
+ y
2
dx
_
dy
=
_
4
0
_
x
3
3
+ xy
2
_

y
0
dy
=
_
4
0
y
3/2
3
+ y
5/2
dy
=
_
2y
5/2
3 5
+
2y
7/2
7
_
4
0
=
2
6
3 5
+
2
8
7
= 2
6
_
1
3 5
+
2
2
7
_
.
4