Block 1 Vector Analysis

APPENDIX A2
DOUBLE INTEGRALS
In Unit 3, you have learnt how to integrate vector functions of single variables.
It is possible to extend the idea of a definite integral to calculate double and
triple integrals which are integrals of functions of two and three variable
respectively. Double and triple integrals have many applications in physics.
For example, we use these integrals to determine the volume of an object
bound by an arbitrary surface, its mass, its centre of mass or its moment of
inertia. In this appendix, we explain in brief how to evaluate a double integral,
which is an integral of a function of two variables.

A2.1 DOUBLE INTEGRALS

We first develop the geometrical concept of the double integral. Before we do
this, however, you should revise the concept of the definite integral of a
function.
y

f (x )

x
a b
f ( xi )
x i
b 
Fig. A2.1: Definite integral of the function f (x ) :  f ( x )dx  as area under a curve.

 
a 

We define the definite integral of a function f (x) over the interval [a, b] on the
b


x-axis, denoted by f ( x ) dx as the limit of a sum. We start by dividing the
a
interval [a, b] into n sub-intervals, the ith sub-interval having a width x i , as
shown in Fig. A2.1. The sum  f ( xi )xi is the total area of the n rectangles
we see in the figure. Then the sum of the areas of the rectangles is
approximately the area under the curve. It is also clear that if we increase the
number of sub-intervals, i.e., increase the value of n, the rectangles become
narrower, and the total area of the rectangles comes closer and closer to the
area under the curve. The exact area is then given by the limit of the sum as n
goes to infinity:
i n
Area under the curve f ( x )  lim
n 
 f ( x i )x i (A2.1a)
i 1

The expression on the right hand side of Eq. (A2.1a) is called the definite
136 integral of f(x) from a to b and denoted as follows:
Appendix A2 Double Integrals
b i n
 f ( x )dx  nlim

 f ( x i )x i (A2.1b)
a i 1

This definition of the definite integral holds even if f (x) has both positive and
negative values in the interval [a, b]. The integral exists if the function f is
continuous on [a, b] or has only a finite number of jump discontinuities.

Let us now explain the concept of double Integral of the function f (x,y) over
a bounded region R on the xy-plane denoted by

 f ( x, y ) dx dy
R
z
f ( x, y )

Box with base area

Ai  xi y i and height
f ( xi , y i )
Each rectangle has an
y
area Ai  xi y i

Projection of f in the xy
x xi plane
y i

Fig. A2.2: Double Integral  f ( x, y )dxdy as the volume under a surface

R
f ( x, y ) and above the region R in the xy plane.

The definition of the double integral is similar to that of a definite integral. We

divide the region R in the xy plane into n tiny rectangles by drawing lines
parallel to the x and y axes. Each rectangle has an area Ai (see Fig. A2.2).

We number the rectangles within R from i = 1 to i = n and choose a point

( xi , y i ) in each rectangle. Now consider the sum:
n n
Sn   f ( xi , y i )Ai   f ( xi , y i )xi y i (A2.2)
i 1 i 1

We can evaluate this sum for increasing values of n such that the maximum
diagonal of the rectangles goes to zero as the number of rectangles goes to
infinity. If f ( x, y ) is a continuous function in R, these sums (also called the
Reimann sums) converge to a limiting value which does not depend on either
the values of ( xi , y i ) or the choice of subdivision. This limit is the double
integral of the function f ( x, y ) over the region R.

The double integral of a function f(x, y), which is defined for all (x, y) in a
closed, bounded region R in the xy plane, is written as the limit of a sum as
follows:
n
lim
n 
 f ( xi , y i )Ai   f ( x, y )dA   f ( x, y )dxdy (A2.3)
i 1 R R
137
Block 1 Vector Analysis
For f ( x, y )  0, the double integral gives us the volume of the solid that lies
Just as the area
under the curve f (x) below the surface f ( x, y ) and above the region R in the xy plane (read the
in Fig. A2.1b is the margin remark).
area under the curve
above the x-axis. PROPERTIES OF THE DOUBLE INTEGRALS

For two functions f ( x, y ) and g ( x, y ), which are defined and continuous in

a region R:

1.  c f ( x, y )dxdy  c  f ( x, y )dxdy (A2.4)

R R

y where c is a constant.
R1
R3 2. Linearity

 f ( x, y )  g( x, y )dxdy   f ( x, y ) dxdy   f ( x, y ) dxdy (A2.5)

R R R

where  and  are constants.

x
3. Additivity

R2 If the region of integration R can be broken up into a finite number of

non overlapping regions R1 , R2 ……..Rn, (Fig. A2.3), then we can
Fig. A2.3: The region R
write:
is broken up into three
overlapping regions
R1, R2 and R3 .  f ( x, y )dxdy   f ( x, y )dxdy   f ( x, y )dxdy  ....   f ( x, y )dxdy
R R1 R2 Rn
(A2.6)
4. Area Property
If the function being integrated is f(x,y) = 1, then

 1 dxdy  Area of the region R (A2.7)

R

We briefly explain how to evaluate double integrals

y C1 p ( x0 )
A2.2 EVALUATION OF DOUBLE INTEGRALS

To evaluate a double integral  f ( x, y )dxdy over a region R, we have to carry

R

out two successive integrations over the variables x and y. How is this done?
You will see that there are actually two ways of doing this. Let us see what
x these are.
a x0 b
C2 First let us define the region of integration R shown in Fig. A2.4 as:
q ( x0 )
a  x  b; q( x )  y  p( x ) (A2.8)
Fig. A2.4: Evaluating
the double integral as As you can see from Fig. A2.4, the values of the x coordinate at the two
an iterated integral extremities of the region are x = a and x = b. Now in between these two values
(Eq. A2.9). of x, the region R is bound by the two curves C1 and C 2 . These two curves
are given by the equations y = p(x) and y = q(x), respectively. What does this
138 tell us?
Appendix A2 Double Integrals
It tells us that for each value of x in the interval [a, b], the value of y ranges y
C3
between q(x) and p(x) which are the points on the lower and upper curves d
bounding the region R. So for any value of x, for example x  x 0 in [a, b], the
values of y range from q( x 0 ) to p( x 0 ). Now if we were to first integrate the
0 y
function f(x, y) over the variable y, holding x as a constant, the limits in y
would be from y = q(x) to y = p(x). The result would be a function of only x. c
h( y 0 )
Next we integrate this function of x with respect to x from x = a to x = b. Thus,
x
we cover the entire region of R while integrating over the two variables.
g(y 0 )
Therefore: C4

 p( x )
b  Fig. A2.5: Evaluating the
f ( x, y )dxdy  
 
f ( x, y )dy dx

(A2.9) double integral as an
R x a  y q ( x )  iterated integral
(Eq. A2.11).
The quantity in the brackets, which is evaluated first is the integral of f(x, y)
over y alone, with the limits as specified. The result of this integral is a function
of x alone which is then integrated over x, over the limits shown.
We could have chosen to carry out this integration in another way. Refer to
Fig. A2.5. We can write down the limits on x and y for the same region R in a
different way as we describe below: Both these iterated
integrals defined in
c  y  d; g( y )  x  h( y ) (A2.10)
Eqs. (A2.9) and (A2.11)
Now for any value of y in the interval [c, d] the value of x is decided by the are equal if p(x), q(x),
function h(y) (curve C 3 ) and g(y) (curve C 4 ), which respectively now defined g(y), h(y) are continuous
functions, for the limits
the upper and lower boundaries of R. Now we can integrate the function f(x, y)
defined in Eqs. (A2.8)
over the variable x, holding y as a constant, the limits of the integral would be
and (A2.10). This is the
from x = g(y) to x = h(y) and the result would be a function of only y. We then consequence of a
integrate this function of y over x from y = c to y = d. So we get an alternative theorem in multivariable
expression for the evaluation of the double integral: calculus called the
Fubini’s theorem, which
d  h( y ) 
is beyond the scope of
 f ( x, y )dxdy   
  f ( x, y )dx dy

(A2.11)
this course.
R y c  x  g ( y ) 
As before, the integral within the brackets is carried out first. Both Eqs. (A2.9)
and (A2.11) are equivalent methods of determining the double integral. In
Eq. (A2.9), the integral over the variable y is carried out first. In Eq. (A2.11),
the integral over x is carried out first.
Suppose R cannot be represented by the inequalities shown in Eq. (A2.8) or
Eq. (A2.10), but can be subdivided into many parts that can be represented by
inequalities, then we evaluate the double integral over each part and sum up
to get the result as the double integral over R.
 h( y )
d  b  p( x ) 
The integrals of the form 
  
f ( x, y )dx dy and


  
f ( x, y )dy dx are

y c  x  g ( y )  x a  y q ( x ) 
called iterated (repeated) integrals because they are evaluated first by
integrating with respect to one variable, either x or y, as the case may be and
then integrating the result with respect to the second variable. Multiple
integrals are usually integrated as iterated integrals. We shall use the same
technique to evaluate triple integrals as you will see in Unit 4. Let us
summarise these results. 139
Block 1 Vector Analysis

EVALUATION OF A DOUBLE INTEGRAL

Suppose that f(x, y) is a continuous function on the region R. If R is
described by the inequalities a  x  b, q( x )  y  p( x ), then

 p( x )
b 
f ( x, y )dxdy  
 
f ( x, y )dy dx

(A2.9)
R x a  y q ( x ) 

If R is described by the inequalities c  y  d, g( y )  x  h( y ) , then

d  h( y ) 
 f ( x, y )dxdy   
  f ( x, y )dx dy

(A2.11)
R y c  x  g ( y ) 

In some textbooks that the iterated integrals are sometimes written without the
bracket as follows:
b d  bd

 f ( x, y )dy dx 

  
f ( x, y )dydx (A2.12a)
a c  ac
By convention, the
limits of integration on d b  db
the variable over
which the integration
and  f ( x, y )dx dy 
  
f ( x, y )dxdy (A2.12b)
c a  ca
is carried out first,
appears on the inner Note: The order of integration in Eqs. (A2.12a and b) is different. It is as
integral sign.
shown in the left hand side of each of these equations.

In Eq. (A2.12a), we write dy dx in the integrand. This means that we first

integrate with respect to y over the interval [c, d ] and then with respect to x
over [a, b] . In Eq. (A2.12b), we write dx dy in the integrand. So, the integration
is first with respect to x and then with respect to y.
y
A Special Case
d
An important special case is when we evaluate double integrals for which the
following is true:
c i) the region R is a rectangle defined by the limits, say
x
a  x  b, c  y  d (Fig. A2.6).
a b
ii) The function f(x, y) = h(x) g(y), that is f(x, y) is a product of two functions,
Fig. A2.6: A rectangular one of which is a function of only x, h(x) and the other a function of only y,
region of integration
that is g(y).
a  x  b, c  y  d.
Then the double integral can be evaluated as:

 b  d 
f ( x, y )dxdy   h( x ) dx   g ( y ) dy 
  (A2.13)
  
R  x a   y c 

As you can see, here we can integrate with respect to each variable
separately. We now evaluate some double integrals to illustrate these
140 methods.
Appendix A2 Double Integrals

XAMPLE A2.1: DOUBLE INTEGRAL OVER A

RECTANGULAR REGION

Evaluate the integral  sin x cos y dx dy where R is a square on the xy

R
plane defined by 0  x   / 2, 0  y   / 2.

SOLUTION  Using (Eq. A2.13), we solve the integral as:

 
 / 2  / 2 
sin x cos ydxdy   sin x dx   cos y dy    cos x 02  sin y 02  1
 
 x 0   y 0 
R  

Note that if the region R is rectangular, but f ( x, y ) cannot be written as the

product of two functions i.e., f ( x, y )  h( x )g( y ), we shall have to carry out an
iterated integral.
Let us now work out an example of an iterated integral. A B

XAMPLE A2.2: DOUBLE INTEGRAL

Fig. A2.7: Region of
Evaluate the integral  ( x  4y ) dx dy where R is the region bounded by the
R
integration for Example
A2.2. The two curves
2 intersect at x   1 and
curves y  2x and y  1  x 2.
x  1.
2 2
SOLUTION  In Fig. A2.7, we plot the two curves y  2x and y  x  1
which define R. Now, as you can see from the figure, the two curves
intersect at the points A and B. At the points of intersection of the two
curves, we have 2x 2  x 2  1. Solving for x we have:
2x 2  x 2  1  x 2  1  x  1, 1
So the points of intersection are x  1 and x  1. This marks the limits of
x for the region of integration. As you can see, for each value of x in the
range  1  x  1 , the value of y will vary in the range 2x 2  y  1  x 2 . Now
let us use Eq. (A2.9) to evaluate the integral with q( x )  2x 2 and
p( x )  x 2  1. Then we write:
 1 x 2
1 
( x  4y ) dx dy  

  ( x  4y )dy  dx

(i)
R x 1  y 2 x 2 
You can see that we have used a  1, b  1. Now we first carry out the
integration within the bracket, integrating over y and taking x as a constant.
x 2 1
( x  4y )dy   xy  2y 2 2   x(1  x 2  2x 2 )  2(1  x 2 )2  2(2x 2 )2
1 x 2
  2 x 
y  2x 2

 2  x  4x 2  x 3  6x 4 (ii)

Substituting the quantity in the bracket in Eq. (i) by the expression in

Eq. (ii) we get,
141
Block 1 Vector Analysis

 ( x  4y ) dx dy   2  x  4x 2  x 3  6x 4 dx
R x  1
1
 x2 x3 x 4 x5  64
 2x  4  6   (iii)
 2 3 4 5  1 15

SAQ - Double integrals over a rectangular region

1. Evaluate the following integrals:
22
a)   xye( x  y )dydx
00
0 2
x
b)   y sin 4 dxdy
1 0

 ( x
4
2. Evaluate  2y )dx dy , where R is the region defined by the
R
equations  1  x  1 and  x 2  y  x 2.

A2.3 SOLUTIONS AND ANSWERS

Self-Assessment Questions
1. a) Using Eq. (A2.13), we write the integral:
When we integrate 2 2
2  2 
 xye  
(x  y )
over y, x is a I dy dx   x e dx   y e y dy 
x

constant. x  0 y 0  0   0 

  
 x e x  e x 02  y e y  e y 02 e2  1
2
  
b) Using Eq. (A2.13), we write
0 2 0  2 
 x   x 
I   y sin   dx dy  y dy   sin  dx 
 4 

   4  
 
y  1 x 0 1  0 
0
y2    4  2
 x  2
      cos   
 2  1      4  0 

2. We use Eq. (A2.9) to write:

 x2 
    
1
 x  2y dy  dx 
4 4
I x  2y dx dy 
 
R x  1  y  x 2 
Carrying out the integral over y first and applying the limits of integration,
we get:
x2

 x   2x dx
1 1
I 4
y y 2
dx  6

1  x2 1

We now integrate over x to get:

1
 2x 7  4
I  
142  7  1 7
 TABLES OF DERIVATIVES AND INTEGRALS
Table A1.1: Derivatives of simple functions

S. df/dx S. df/dx
No. No.

d d  1 1
1. (c )  0, c constant 10.
    , x  0
dx dx  x  x2

d d n n
2. (x)  1 11. (x )  
dx dx n 1
x

d n
( x )  nx n  1
d d  d 
3. 12.  g ( x )  h( x )   g ( x )   h( x )
dx dx  dx   dx 

d d  df   dg 
4. (sin x )  cos x 13.  f ( x ) g ( x )   g  f 
dx dx  dx   dx 

df dg
d g  f
5. (cos x )   sin x 14. d  f (x)  dx dx , g  0
dx   
dx  g ( x )  g2

d d x
6. (tan x )  sec 2 x 15. (e )  e x
dx dx

d 1 d 1
7. (sin1 x )  16. ln x 
dx 1  x2 dx x

d 1 d x
8. (cos1 x )   17. (c )  c x ln c, c  0
dx 1  x2 dx

d 1 d 1
9. (tan1 x )  18. logc x  c  1, c  0
dx 1  x2 dx x ln c

143
Block 1 Vector Analysis

Table A1.2: Integrals of simple functions

S. Integral S. Integral
No. No.

 a dx  ax  c, a and c constants  sin x dx

1. 5.   cos x  c, c constant

xn 1
 cos x dx
2. 6.  sin x  c, c constant
 x n dx 
n 1
 c, c constant

 tan x dx
3. 1 7.
 x dx  ln x  c, c constant  ln sec x  c, c constant

e e
4. x
dx  e x  c, c constant 8. ax
dx  aeax  c, a and c constants

144