Lecture 36

1 Multiple Integrals
1.1 Double integrals
Let f (x, y) be a real valued function defined over a domain Ω ⊂ IR2 . To start with, let us
assume that Ω be the rectangle R = (a, b) × (c, d). We partition the rectangle with node
points (xk , yk ), where

a = x1 < x2 < ...., xn = b, and c = y1 < y2 < ...., yn = d.

Let Rnm be the small sub-rectangle with above vertices. Now we can define Upper and lower
Riemann sum as
X
U (Pn , f ) = sup f (x, y)|Rnm |
n,m Rnm

and
X
L(Pn , f ) = inf f (x, y)|Rnm |
Rnm
n,m

where |Rnm | is the area of the rectangle Rnm . Here we may define the norm of partition Pn
p
as kPn k = max |xi − xi−1 |2 + |yi − yi−1 |2 . Then by our understanding of definite integral
i
we can define the upper, lower integrals and f (x, y) is integrable if and only if

inf{U (P, f ) : P } = sup{L(P, f ) : P }.

The definite integral is defined as

ZZ
f (x, y)dA = inf{U (P, f ) : P } = sup{L(P, f ) : P }.
Ω
ZZ X
It can be shown that f (x, y)dA = lim S(Pn , f ), where S(Pn , f ) = f (xk , yk )|Rnm |.
Ω kPn k→0
n,m

We have the following Fubini’s theorem for rectangle:

Suppose f (x, y) is integrable over R = (a, b) × (c, d), then
ZZ Z b Z d  Z d Z b 
f (x, y)dA = f (x, y)dy dx = f (x, y)dx dy
R a c c a

In case of f (x, y) ≥ 0 we may interpret this as the volume of the solid formed by the
surface z = f (x, y) over the rectangle R. This is precisely the ”sum” of areas of the cross
Rd
section A(x) = c f (x, y)dy between x = a and x = b. Since x varies over all of (a, b), this
Rb
sum is nothing but the integral a A(x)dx.

1
For any general bounded domain Ω, we can divide the domain into small sub domains Ωk and
consider the upper, lower sum exactly as above by replacing Rnm by Ωk . Then a function
f (x, y) : Ω → IR is integrable if the supremum of lower sums and infimum of upper sums exist
and are equal. We may define
ZZ X
f (x, y)dA = lim f (xk , yk )|Ωk |
Ω kPn k→0
k

The basic properties of the definite integral like integrability of f ± g, kf and domain
decomposition theorems holds in this case also.

We call a domain as y-regular if Ω = {(x, y) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)} for some

continuous functions g1 , g2 . Similarly, Ω is x-regular if Ω = {(x, y) : c ≤ y ≤ d, h1 (y) ≤ x ≤
h2 (y)}. A domain is called regular if it is either x-regular or y-regular. Then we have the
following Fubini’s theorem for regular domains.

Theorem 1.1.1 Let f (x, y) be continuous over Ω.

1. If Ω is y-regular, i.e.,Ω = {(x, y) : a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)}. Then

ZZ Z b Z g2 (x) !
f (x, y)dA = f (x, y)dy dx.
R a g1 (x)

2. If Ω is x-regular, i.e.,Ω = {(x, y) : c ≤ y ≤ d, h1 (y) ≤ x ≤ h2 (y)}. Then

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

This theorem basically says that if a function is integrable over a domain Ω, then the value
of integral is does not depend on order of integration. That is we can integrate with respect
to x first followed by y or vice versa.
ZZ
Example 1.1.1 Evaluate the integral (x + y + xy)dA where Ω is the triangle bounded by
Ω
y = 0, x = 1 and y = x.

Solution: The triangle is regular in both x and y. The given triangle is

{(x, y) : 0 ≤ x ≤ 1, 0 ≤ y ≤ x} = {(x, y) : 0 ≤ y ≤ x, y ≤ x ≤ 1}

Therefore, taking it as y− regular we see that the domain is bounded below by y = g1 (x) = 0
and above by y = g2 (x) = x over x = 0 to x = 1. Hence,
ZZ Z x=1 Z x 
(x + y + xy)dA = (x + y + xy)dy dx
Ω x=0 y=0
1
x2 x3
Z
15
= (x2 + + )dx =
0 2 2 24

2
Similarly, taking it as x− regular, we see that the domain is bounded below y x = h1 (y) = y
and above by x = h2 (y) = 1 over y = 0 to 1. Hence
ZZ Z 1 Z 1 
15
(x + y + xy)dA = (x + y + xy)dx dy =
Ω y=0 x=y 24
ZZ
Example 1.1.2 Evaluate the integral (2+4x)dA where Ω is the domain bounded by y = x
Ω
and y = x2 .

Solution:
ZZ Z 1 Z x 
(2 + 4x)dA = (2 + 4x)dy dx
Ω x=0 y=x2
Z 1
= (2x + 2x2 − 4x3 )dx = 2/3
0
√ !
Z 1 Z y
On the other hand, this is also equal to (2 + 4x)dx dx.
y=0 x=y
P
Remark 1.1.1 1. Whenf (x, y) = 1, then we approximate the area of Ω as A ∼ k Ωk =
P
k f (xk , yk )|Ωk | where f = 1. By the definition of Riemann integral this sum converges
RR
to Ω dA as kPn k → 0.
RR
2. As discussed in the beginning, when f (x, y) ≥ 0, the Ω f (x, y)dA is the volume of the
solid bounded above by z = f (x, y) and below by Ω.

Problem 1.1.1 Find the area bounded by y = 2x2 and y 2 = 4x.

Solution: The two parabola’s intersect at (0, 0) and (1, 1). Hence the area is
√
Z 1Z 2 x Z 1 √ 2
A= dydx = (2 x − 2x2 )dx = .
0 2x2 0 3

Problem 1.1.2 Find the volume of the solid under the paraboloid z = x2 + y 2 over the
bounded domain R bounded by y = x, x = 0 and x + y = 2.

Solution: The domain of integration R is Y -regular bounded above by x + y = 2 and below

by y = x with x varying over (0, 1).
ZZ Z 1 Z y=2−x 
V = (x2 + y 2 )dA = (x2 + y 2 )dy dx
R 0 y=x
1 1
y3
Z Z
1
= + yx2 y=x
y=2−x
dx = ( ((2 − x)3 − x3 ) + x2 (2 − 2x))dx
0 3 0 3

Problem 1.1.3 Find the volume of the solid bounded above by the surface z = x2 and below
by the plane region R bounded by the parabola y = 2 − x2 , y = x.

3
Solution: The points of intersection of y = x, y = 2 − x2 are x = −2, 1. So R = {(x, y) :
−2 ≤ x ≤ 1, x ≤ y ≤ 2 − x2 }. Therefore,
Z 1 Z 2−x2 Z 1
2
V = x dydx = x2 (2 − x2 − x)dx
x=−2 y=x −2

Change of order
Consider the evaluation of integral R sinx x dA over the triangle formed by y = 0, x = 1 and
RR

y = x. Since the can be extended as continuous function over R, by the basic properties of
Riemann integral the function is integrable. Now by Fubini’s theorem, the value of integral
does not depend on the order of integration. As we noted earlier R is regular in x and y. If
we take it as x regular, then R = {(x, y) : 0 ≤ y ≤ 1, y ≤ x ≤ 1} and try to evaluate the
integral, then ZZ Z 1 Z 1 
sin x sin x
dA = dx dy.
R x 0 x=y x

This is singular integral and difficult to evaluate.

But when we consider R to be y-regular, we see that R = {(x, y) : 0 ≤ x ≤ 1, 0 ≤ y ≤ x}.
Then the given integral is
ZZ Z 1Z x Z 1
sin x sin x
dA = dydx = sin xdx = 1 − cos 1
R x 0 y=0 x 0

At times this technique can be used to evaluate some complicated definite integrals, for
example,
R∞ e−ax −e−bx
Problem 1.1.4 Evaluate the integral 0 x dx, a, b > 0.

Solution: This integral is equivalent to

∞ ∞ Z b
e−ax − e−bx
Z Z 
−xy
dx = e dy dx
0 x 0 a

The domain of integration is the infinite strip {(x, y) : 0 ≤ x ≤ ∞, a ≤ y ≤ b}. Changing the
order of integration, we get
Z ∞ Z b  Z b Z ∞ 
−xy −xy
e dy dx = e dx dy
0 a y=a 0
b
= ln
a

1.2 Double integrals in Polar form

Suppose we are given a bounded region whose boundaries are given by polar equations, say
f1 (r, θ) = 0, f2 (r, θ) = 0. Then we divide the region into smaller ”polar rectangles” whose
sides have constant r, θ values.

4
Suppose f (r, θ) is defined over a region R defined using the polar equations, R : α ≤ θ ≤
β, g1 (θ) ≤ r ≤ g2 (θ). Then we divide the r range by ∆r, 2∆r, ...., m∆r and α, α + ∆θ, .....α +
m0 ∆θ = β. Let ∆A be the polar rectangle with sides rk − ∆r/2, rk + ∆r/2 and α + k∆θ, α +
(k + 1)∆θ. Then we define the Riemann sum as
X
Sn = f (rk , θk )∆Ak .
k

The area of small ”polar rectangle” Ak is

∆Ak = area of outer sector − area of inner sector = rk ∆r∆θ.

As kPn k → 0,we get

X ZZ
Sn = f (rk , θ)rk ∆r∆θ → f (r, θ)rdrdθ.
k R

Problem 1.2.1 Find the area common to the cardioids r = 1 + cos θ and r = 1 − cos θ.

Solution: Since the region is symmetric with respect to x -axis and y-axis, the required area
is
Z π/2 Z 1−cos θ
A =4 rdrdθ
θ=0 r=0
Zπ/2 Z π/2
1 2 π
=4 (1 − 2 cos θ + cos θ)dθ = 2( − 2 + cos2 θdθ)
0 2 2 0
ZZ
Problem 1.2.2 Evaluate 3ydA where R is the region bounded below by x-axis and above
R
by the cardioid r = 1 − cos θ.

Solution: The given integral is equivalent to

ZZ Z π Z 1−cos θ
3ydA = r sin θrdrdθ.
R θ=0 r=0
Z ∞
2
Problem 1.2.3 Evaluate I = e−x dx.
0

Solution: Recall that 2I = Γ( 12 ). Using the Fubini’s theorem, we may write

Z ∞  Z ∞  Z ∞Z ∞
−x2 −y 2 2 2
2
I = e dx e dy = e−(x +y ) dA
0 0 0 0

Where the integration is over the first quadrant (0, ∞) × (0, ∞). So representing this in polar
form we integrate over {(r, θ) : 0 ≤ r < ∞, 0 ≤ θ ≤ π2 }. Therefore, the above integral becomes
Z π/2 Z ∞
2 π
e−r rdrdθ = .
0 0 4