.

Chapter 14 Multiple Integrals

.
1. Double Integrals, Iterated Integrals, Cross-sections

2. Double Integrals over more general regions, Definition, Evaluation of

Double Integrals, Properties of Double Integrals

3 . Area and Volume by Double Integration, Volume by Iterated Integrals,
Volume between Two surfaces,
.4 Double Integrals in Polar Coordinates, More general Regions
5. Applications of Double Integrals, Volume and First Theorem of Pappus,
Surface Area and Second Theorem of Pappus, Moments of Inertia
6. Triple Integrals, Iterated Triple Integrals,

7. Integration in Cylindrical and Spherical Coordinates

8. Surface Area, Surface Area of Parametric Surfaces, Surfaces Area in

Cylindrical Coordinates
. Change of Variables in Multiple Integrals, Jacobian
9
.

. . . . . .

Matb 200 in 2013-2014

 .
Chapter 14 Multiple Integrals
.
1. Taylors Theorem for function of 2 variables

14.1 Riemann integral for function of 1 variable

14.1 Double integral by Riemann sums
14.1 Iterated Integral, Fubini Theorem on rectangular region
14.2 Double Integral over Non-Rectangle Region
14.2 cross-sections, Fubini Theorem on general region
14.2 technique of determining the upper and lower limits
14.3 Area and Volume of Double integration
.

. . . . . .

Matb 200 in 2013-2014

.
Definition. The definite integral of a function f (x) of one variable defined on an
interval [a, b], is given by
b n n
f (x) dx = lim
max xk 0
f (xk )xk = n+
lim f (xk )xk .

.a k =1 k =1

This concept arose from the problem of finding areas under curves. Now we
have similar problem if we replace the function of two variables.
.
Volume Problem. Given a function f (x, y) of two variables that is continuous
and nonnegative on a region R in the xy-plane, find the volume of the solid
enclosed
. between the surface z = f (x, y) and the region R.

. . . . . .

Matb 200 in 2013-2014

.
Volume and Double Integrals
.
Let f (x, y) be a function of two variables defined over a rectangle
R = [a, b] [c, d]. We would like to define the double integral of f (x, y) over R
. the algebraic volume of the solid under the graph of z = f (x, y) over R.
as
b
The idea is similar to the case of integral a f (x)dx in one variable case, in
which we subdivide the interval into smaller subintervals with uniform width
x = b a
n , and then choose arbitrary points xi in the subinterval Ii . Then we
n
have the approximate Riemann sum f (xi )x.
i=1
Of course, this sum depends on the choices of xi and the subinterval.
In fact, this idea can be implemented in 2-dimensional cases as well.

. . . . . .

Matb 200 in 2013-2014

Let R = [a, b] [c, d]. and f (x, y) be a function defined on R. We first subdivide
the rectangle R into mn small rectangles Rij , each having area A, where
i = 1, , m and j = 1, , n. For each pair (i, j), pick an arbitrary point
(xij , yij ) inside Rij . Use the value f (xij , yij ) as the height of a rectangular solid
erected over Rij . Thus its volume is f (xij , yij )A.

The sum of the volume of all these small rectangular solids approximates the
volume of the solid under the graph of z = f (x, y) over R. This sum
m n
f (xij , yij )A is called Riemann sum of f .
i=1 j=1
.
Definition The double integral of f over R is
m n

R
f (x, y) dA = lim
m,n
f (xij , yij )A, if this limit exists.
. i=1 j=1
. . . . . .

Matb 200 in 2013-2014

Remarks. In general, it is very difficult to prove that the limit of Riemann sum
converges, because of the choices of the height f (xij , yij ) involved. The usual
method is replace the height either the maximum and the minimum values of f
within each smaller rectangles, and hence we obtain the upper and lower
Riemann sums respectively.
.
Theorem. If f (x, y) is
continuous on a domain containing the rectangle R, then
the double integral f (x, y) dA always exists.
. R

. . . . . .

Matb 200 in 2013-2014

.
Example. Approximate the value of the integral (4x3 + 6xy2 ) dA over the
R
rectangle R = [1, 3] [2, 1], by means of the Riemann sums, with xi = 1,
and
. yj = 1.

Solution. Partition the rectangle R into six 1 1 squares

Ri with area Ai = 1 (i = 1, , 6). Choose the center
points (xi , yi ) for each square as shown on the right.

6
The desired Riemann sum is f (xi , yi )Ai
i=1
= f ( 32 , 23 ) 1 + f ( 52 , 23 ) 1 + f ( 32 , 21 ) 1
+f ( 52 , 21 ) 1 + f ( 32 , 12 ) 1 + f ( 52 , 12 ) 1
135 385 63 265 63 265
= + + + + + = 294,
4 4 4 4 4 4
which is called the midpoint approximation of the integral (4x3 + 6xy2 ) dA.
R

. . . . . .

Matb 200 in 2013-2014

.
Iterated Integrals
. d
Let f (x, y) be a function defined on R = [a, b] [c, d]. We write f (x, y) dy to
c
mean that x is regarded as a constant and f (x, y) is integrated with respect to y
from
. y = c to y = d.

.
d
Therefore, the value of the integral f (x, y) dy is a function of x, and we can
c
integrate it with respect to x from x = a to x = b. The resulting integral
b ( d )
f (x, y) dy dx is called an iterated integral.
a c

. . . . . .

Matb 200 in 2013-2014

.
Iterated Integrals
. d
Let f (x, y) be a function defined on R = [a, b] [c, d]. We write f (x, y) dy to
c
mean that x is regarded as a constant and f (x, y) is integrated with respect to y
from
. y = c to y = d.

.
d
Therefore, the value of the integral f (x, y) dy is a function of x, and we can
c
integrate it with respect to x from x = a to x = b. The resulting integral
b ( d )
f (x, y) dy dx is called an iterated integral. Similarly one can define
d ( b )
a c

the iterated integral f (x, y) dx dy.

c a
Remark. We call the blue and red segments inside the region R the
cross-sections
. of R cut by the line y = y0 and x = x0 respectively.
. . . . . .

Matb 200 in 2013-2014

.
Example. Evaluate the iterated integrals
3 2 2 3
(a) x2 y dy dx, (b) x2 y dx dy.
. 0 1 1 0

Solution.
3 2 3 2
3 2
(a) x2 y dydx = x2 y dydx = x2 y dydx ==
0 1 0 1 0 1
3 [ 2 ]y=2 3 2 [ 3 ] x=3
y 3x x 27
x2 dx = dx = = .
0 2 y=1 0 2 2 x=0 2
2 3 2 [ 3 ] x=3 2 [ 2 ]y=2
x y 9y 27
(b) x2 y dx dy = dy = 9ydy = = .
1 0 1 3 x=0 1 2 y=1 2

. . . . . .

Matb 200 in 2013-2014

.
Fubinis Theorem for Rectangle case
.
Thoerem. If f (x, y) is continuous on R = [a, b] [c, d], then
b d d b
f (x, y) dy dx = f (x, y)dA = f (x, y) dx dy.
.a c R c a

.
Example. f (x, y) is a positive function defined on a rectangle R = [a, b] [c, d].
The volume V of the solid under the graph of z = f (x, y) over R, is given by
either one of the iterated integrals:
b d d b
f (x, y) dy dx, or f (x, y) dx dy.
. a c c a

. . . . . .

Matb 200 in 2013-2014

.
Proof of Fubini Theorem for rectangle case
.
Thoerem. If f (x, y) is continuous on R = [a, b] [c, d], then
b d d b
f (x, y) dy dx = f (x, y)dA = f (x, y) dx dy.
.a c R c a

Proof. First partition [a, b], and [c, d] each into n equal subintervals, so that we
have n2 smaller rectangles of area A = xy. We construct a Riemann sum
of f on R which is close to the iterated integral. For i = 1, , n, one can
choose any point xi in subinterval [xi1 , xi ]. It follows from the mean value
theorem of integral that there exists yij in subinterval [yj1 , yi ] with
y
j
f (xi , y)dy = f (xi , yij )y. This produces a point (xi , yij ) in each rectangle
yj 1
[xi1 , xi ] [yj1 , yi ]. Then the Riemann sum
( ) ( )
n n n n n yi
f (xi , yij ) A = f (xi , yij ) y x = f (xi , y)dy x =
i,j=1 i=1 j=1 i=1 j = 1 yi 1
n ( d ) n d
f (xi , y) dy x = A(xi )x, where A(x) = c f (x, y) dy. The result
i=1 c i=1
follows as n tends to +.
. . . . . .

Matb 200 in 2013-2014

.
Fubinis Theorem for Rectangle case
.
Thoerem. If f (x, y) is continuous on R = [a, b] [c, d], then
b d d b
f (x, y) dy dx = f (x, y)dA = f (x, y) dx dy.
.a c R c a

.
More generally, this is true if f is bounded on R, f is discontinuous only at a
finite
. number of smooth curves, and the iterated integrals exist.

Furthermore, the theorem is valid for a general closed and bounded region as
discussed in the subsequent sections.

. . . . . .

Matb 200 in 2013-2014

.
Example. Find the volume of the solid S that is bounded by the elliptic
paraboloid x2 + 2y2 + z = 16, the planes x = 2, y = 2, and the 3 coordinate
.planes.
Solution. Let R = { (x, y) 0 x 2, 0 y 2 }, and one can rewrite the
defining equation of the elliptic paraboloid as z = 16 x2 2y2 where
(x, y) R. Then the volume of S is given by the double integral
2 2
f (x, y) dA = (16 x2 2y2 ) dx dy
R 0 0
2[ ]x=2
x3
= 16x 2xy2
dy
0 3 x=0
2
88 88 2 4 22 160
= ( 4y2 ) dy = = .
0 3 3 3 3
Remark. In this problem, one has to decide which is the height function, just
like z = f (x, y) in the previous formulation. In some other cases, one can use
x = h(y, z) or y = g(x, z) as the height function.

. . . . . .

Matb 200 in 2013-2014

.
Proposition. (a) In general, if f (x, y) = g(x)h(y), then
( b ) ( d )
f (x, y)dA = g(x) dx h(y) dy ,
R a c

where R = [a, b] [c, d] is a rectangle.

. The equation above does not hold if the region R is not a rectangle.
(b)
b d b d
Solution. (a) f (x, y)dA = g(x)h(y)dy dx = g(x) h(y)dy dx =
( b ) (R d ) a c a c

g(x) dx h(y) dy .
a c
.
2 2 2 2
Example. exp(x2 + y2 ) = ex +y = ex ey .
Counterexample. One can show that sin(x + y) can not be expressed as
g(x) h(y). However, one can relax the condition into sums of products of
functions:
. sin(x + y) = sin x cos y + cos x sin y.

. . . . . .

Matb 200 in 2013-2014

. Double Integral over Non-Rectangular Region
.
Let f (x, y) be a continuous function defined on
a closed and bounded
region D in R2 . The
double integral f (x, y) dA can be defined
D
similarly as the limit of a Riemann sum
f( xk , yk )Ak , where small rectangle Rk with
k
dimension xk yk lies completely inside
region
. R.

However, due to the irregular shapes of the re-

gion, we subdivide the region by rectangular
grid, and then evaluate the volume of the rect-
angular solid with the base of smaller rectan-
gles the completely lies inside the region R.

. . . . . .

Matb 200 in 2013-2014

.
Example. If R = { (x, y) | 1 x 1, 2 y 2 }, evaluate the double

integral 1 x2 dA.
. R

Solution. We can compute

the integral by interpreting it as a volume of a solid
body D. If we let z = 1 x2 , then z2 = 1 x2 , i.e. x2 + z2 = 1, hence part of
the boundary of D lies in the cylinder. , then and , so the given double integral
represents the volume of the solid S that lies below the circular cylinder and
above the rectangle R. The volume of S is the area of a semicircle with radius
1 times the length of the cylinder. Thus

1
1 x2 dA = Volume of S = 12 (2 (2)) = 2.
R 2

. . . . . .

Matb 200 in 2013-2014

.
Fubini theorem for non-rectangular region
.
Let R be a region in the xy-plane, and suppose there exist two continuous
function ymin (x), ymax (x) defined on the interval [a, b] such that
R = { (x, y) | a x b, ymin (x) y ymax (x) }, then
b ymax (x)
f (x, y) dA = f (x, y) dy dx.
. R a ymin (x)

.
Using Vertical Cross-sections
.
In evaluating R f (x, y)dA, one can ideally use the iterated integral
b ( )
f (x, y)dA = f (x, y) dy dx,
R a ?

the difficulties lies in determining upper and lower limits , ? in the iterated
.integral.

. . . . . .

Matb 200 in 2013-2014

.
Vertical Cross-sections
.
b ( )
f (x, y)dA = f (x, y) dy dx.
. R a ?

We propose the following steps:

1. Sketch and label the bounding curves, and determine the region R of

integration in the double integral.

2. Project the region R onto one the coordinate axes, so that its shadow is an
interval [a, b] or union of intervals on the coordinate axis.
3. Choose any arbitrary point P(x, 0) or P(0, y) in the shadow, draw a line

through P perpendicular to the axis with shadow.

4. Ideally the line meets the boundary R at only two points (x, y
max ) and
(x, ymin ). These two ys depends on x, and hence are functions of x, i.e.
the ones determined by the boundary curves of R. In this case one can
describe the region
R = { (x, y) R2 | a x b, ymin (x) y ymax (x) }.

. . . . . .

Matb 200 in 2013-2014

.
Example. Sketch the region D bounded by the
lines x = 0, y = 0 and
2x + 3y = 1. and evaluate
the double integral x dA.
. D

Solution. The region D is a triangle bounded by

x = 0 and y = 0 and : 2x + 3y = 1. First we
determine the intersection points of these
3 lines. The two coordinates axes meet at (0, 0); the line intersects the x- and
y-axis at (1/2, 0) and (0, 1/3) respectively. Next we determine the order of
b
integration in the iterated integral, for example ( dy)dx. Then the
a
region D has a shadow { x | 0 x 1/2 } on x-axis. Then any vertical red
line through a point (x, 0) on the x-axis will intersect the region at two boundary
points (x, ymax ) and (x, ymin ), where ymin = 0 given by the x-axis, and ymax = y
which satisfies 2x + 3y = 1, i.e. ymax = 132x . So we have
1 12x 1 [ 2 ]1
2 3 2 x(1 2x) 3 2
xdA = x dy dx = dx = x6 2x9 = 721
.
D 0 0 0 3 0

. . . . . .

Matb 200 in 2013-2014

 .
Example. Let R be the region in first quadrant bounded by the two curves
x. 2 + y2 = 1 and x + y = 1.

1. Sketch and label the bounding curves, and determine the region R of
integration in the double integral.
2. Project the region R onto one the coordinate axes, so that its shadow is an

interval [a, b] or union of intervals on the coordinate axis.

. . . . . .

Matb 200 in 2013-2014

 .
Example. Let R be the region in first quadrant bounded by the two curves
x. 2 + y2 = 1 and x + y = 1.

Usually the intersection point of curves C1 : x + y = 1 and C2 : x2 + y2 = 1

gives some important information. 1 = x2 + y2 = x2 + (1 x)2 = 2x2 2x + 1,
i.e. 0 = x(x 1), and hence we know (x, y) = (1, 0) and (0, 1) are the common
intersection. Want to see (by
mathematical means) the relative position of the
curves y = 1 x and y = 1 x2 when x varies in the interval [0, 1].
For 0 x 1, we have 0 1 x 1 + x, so
(1 x)2 (1 + x)(1 x) = 1 x2 ,

it follows that 1 x 1 x2 for 0 x 1, i.e. the line segment C2 is below
the circle C2 .

. . . . . .

Matb 200 in 2013-2014

.
Example. Let R be the region in first quadrant bounded by
the two curves
x2 + y2 = 1 and x + y = 1. Rewrite the double integral f (x, y)dA in iterated
R
integrals.
.
1. Choose any arbitrary point P(x, 0) or P(0, y) in the shadow, draw a line
through P perpendicular to the axis with shadow.
2. Ideally the line meets the boundary R at only two points (x, y
max ) and
(x, ymin ). These two ys depends on x, and hence are functions of x, i.e.
the ones determined by the boundary curves of R. Then the region
R = { (x, y) R2 | a x b, ymin (x) y ymax (x) }.

Answer: f (x, y)dA
1 1 x2
R

= f (x, y) dy dx
0 1x

. . . . . .

Matb 200 in 2013-2014

.
Proposition. The area of a region R in xy-plane is given by dA.
. R

. . . . . .

Matb 200 in 2013-2014

.
Example. Find the area of the region R bounded by the curves C1 : y = x, and
C
. 2 : y = x2 in the first quadrant. Answer in dydx.
Solution. It is observed from the figure above that R =
{ (x, y) | 0 x 1, x2 y x } . We first find the
intersection point (x, y) of C1 and C2 . In this case, both
equations y = x and y = x2 hold, i.e. x = x2 , and so x = 0
or x = 0. It follows that (x, y) = (0, 0) or (1, 1). As x runs
along in the interval [0, 1], i.e. 0 x 1, one needs to
decide which of these two curves Ci (i = 1, 2) lies on top, while the other is at
the bottom. For this, one can compare the y-coordinates of the points (x, x) on
C1 , and (x, x2 ) on C2 , so it follows from 0 x 1 that x x2 = x(1 x) 0.
We know that C1 lies above C2 when x [0, 1]. So
R = { (x, y) | 0 x 1, x2 y x }.
The area of the region R is given by
1 x 1 [ ]1
x2 x3 1
dydx = (x x2 ) dx = = .
0 x2 0 2 3 0 6
. . . . . .

Matb 200 in 2013-2014

. 2 2x
Example. Sketch the region of integration for the integral (4x + 2) dy dx
0 x2
.and write an equivalent integral with the order of integration reversed.

Solution. It follows from the given iterated

integral that the domain of integral R =
{ (x, y) | 0 x 2, x2 y 2x }. In this
case, the top and bottom curves intersect
at (0, 0) and (2, 4).So the region R has
a shadow T = [0, 4] onto y-axis. For any
point P(y, 0) (0 y 4) in the shadow
T,

the red line parallel to x-axis will meet the parabola y = x2 at ( y, y) and the
straight line y = x + 2 at (y 2, y). It follows from 0 y 4, that y2 4y, i.e.

ymin = y/2 y = ymax . Geometrically, meets the boundary curves at two
y
points (xmin , y) and (xmax , y). One has R = { (x, y) | 0 y 4, 2 x y }.
2 2x 4 y
It follows that the (4x + 2) dy dx = (4x + 2) dx dy.
0 x2 0 y/2
. . . . . .

Matb 200 in 2013-2014

.
Example. Find the area of the region R bounded by the curves
C1 : y = x + 2,
and C2 : y = x2 in the 1st quadrant. Answer in integral form dydx.
.

Solution. The region R can be described by means of vertical section as

{ (x, y) | 1 x 2, x2 y x + 2 }. It follows from the description of R that
the area of the region R is given by
2 x+2 2 [ ]x+2 2
dy dx = y dx = (x + 2 x2 ) dx
1 x2 1 x2 1
[ 2 ]2
x x3 9
= + 2x = .
2 3 1 2

. . . . . .

Matb 200 in 2013-2014

.
Example. Find the area of the region R bounded by the curves
C1 : y = x + 2,
and C2 : y = x2 in the 1st quadrant. Answer in integral form dxdy.
.

Solution. If one uses the cross sections parallel to x-axis, then we have the
diagram above, in which the lower limits may be on C1 if y 1, and on C2 if
0 y 1. So one may divide the region R into two subregions R1 and R2 as

above. In fact, R1 = { (x, y) | 0 y 1, y x y }, and

R2 = { (x, y) | 1 y 4, y 2 x y }. So it follows that area of R is
1 y 4 y
1 dA = dA + dA = dxdy + dxdy.
R R1 R2 0 y 1 y2

. . . . . .

Matb 200 in 2013-2014

.
Example. Evaluate (x + 2y)dA, where D is the region bounded by the
D
.parabolas y = 2x2 and y = 1 + x2 .
Solution. We first determine the intersection of the
curves y = 2x2 and y = 1 + x2 as follows. From
2x2 = y = 1 + x2 , we have x2 = 1, and hence x =
1. When x varies within the interval [1, 1], one
has x2 1, i.e. 2x2 1 + x2 . In particular, the
curve y = 2x2 is below the curve y = 1 + x2 . So
D = { (x, y) | 0 x 1, 2x2 y 1 + x2 }. And
we have the figure on the right.
Rewrite the double integral as iterated integral,
1 1 + x2 1 [ ]1+x2
(x + 2y)dA = (x + 2y) dy dx = xy + y2
1 2x2 1 2x2
1D
32
= (3x4 x3 + 2x2 + x + 1) dx = .
1 15
.
Exercise. Evaluate D xydA, where D is the region bounded by the line
y. = x 1 and the parabola y2 = 2x + 6. Answer: 36. . . . . . .

Matb 200 in 2013-2014

. 1 1
Example. Evaluate the iterated integral I = sin(y2 ) dydx by
0 x
.interchanging the order of integration.
Solution. Let D be the region The region of integration, then it follows from the
upper and lower limits of the iterated integral, we have
D = { (x, y) | 0 x 1, x y 1 }, which is given by vertical cross-section.
Hence D is the triangular region bounded by the lines y = x, x = 0 and y = 1.
Rewrite D by means of horizontal cross-section, then 0 y 1, and the
bounding curve will be x = 0 on the left, and x = y on the right of the region D.
Hence, we have another description of D = { (x, y) | 0 y 1, 0 x y }.
1 1 1 y
By Fubinis theorem we have I = sin(y2 ) dydx = sin(y2 ) dxdy =
1 [ ]1 0 x 0 0
1 1 cos 1
y sin(y )dy = cos(y ) =
2 2
0 2 0 2

. . . . . .

Matb 200 in 2013-2014

.
Example. Find the volume of the solid D above the xy-plane and is bounded
2 2
. the cylinder x + y = 1 and the plane z = 0 and z = y.
by
Solution. Since the plane z = 0, z = y is the bottom
and the top faces of the solid D, and the plane z = y
is the top face of the solid, so the height function
f (x, y) = ztop (x, y) zbottom (x, y) = y 0 = y.
Next to determine the region R for double integral,
the planes intersect at points (x, 0, 0), which are on
the x-axis. Therefore, the volume of the solid D can
be computed by integrating this function f over the
bottom face of the solid which is the semi-circular
disk R.
R = { (x, y) | x2 + y2 1, y 0 }
= { (x, y) | 0 y 1, 1 y2 x 1 y2 }.
1 1 y2
The volume of the solid S is y dA = y dx dy (y is a constant)
R 0 1y2
1 [ ] 1
2 2
= 2y 1 y2 dy = (1 y2 )3/2 = .
0 3 0 3
. . . . . .

Matb 200 in 2013-2014

 .
Example. The region common to the interiors of the cylinders x2 + y2 = 1 and
x. 2 + z2 = 1, one-eighth of which is shown in the accompanying figure.

Solution. We just calculate the volume of portion D of the region in 1st octant.
2 2
One immediately recognizes the solid D has a top face given by x + z = 1,
i.e. zmax (x) = 1 x , and the bottom xy-plane given by zmin (x, y) = 0.
2

Moreover, the shadow R of the solid D is a circle disk in the 1st quadrant, so
R = { (x, y) | 0 x 1, 0 y 1 x2 }. The volume of D is given by
1 1 x2
(zmax (x, y) zmin (x, y), ) dA = 1 x dA =
2 1 x2 dy dx
R R 0 0
1 1 [ ]1
x3 2
= 1 x2 1 x2 dx = (1 x2 )dx = x = .
0 0 3 0 3
. . . . . .

Matb 200 in 2013-2014