You are on page 1of 24

17:26 o A O 4061%

SFU
Calculus Early Transcendentals:
Integral&Multi-Variable
Calculus for Social Sciences

2.7 Improper Integrals


Recall that the Fundamental Theorem of
Calculus says that if f is a continuous function
on the closed interval (a, b], then

Ha) de F()= Flb)- Fla),


=

where F is any antiderivative of f

Both the continuity condition and closed interval


must hold to use the Fundamental Theoremn of
Calculus, and in this case, f(a) dæ represents
Ja
the net area under f(c) from a to b:
f)
Area

4 D

We begin with an example where blindly


applying the Fundamental Theorem of Calculus
can give an incorrect result.

Example2.51. Using FTC.


Explain why.da is not equal to-2.
Solution
********************

Formalizing this example leads to the concept of


an improper integral. There are two ways to
extend the Fundamental Theorem of Calculus.
One is to use an infinite interval, i.e., [a, oo),
(-o, b] or (-o, oo). The second is to allow the
interval a, b] to contain an infinite discontinuity
of f(T). In either case, the integral is called an
improper integral. One of the most important
applications of this concept is probability
distributions because determining quantities
like the cumulative distribution or expected
value typically require integrals on infinite
intervals.
2.7.1 Improper Integrals:
Infinite Limits of Integration
To compute improper integrals, we use the
concept of limits along with the Fundamental
Theorem of Calculus.

Definition 2.52. Improper Integrals -

One Infinite Limit of Integration. If


fr) is continuous on (a, oo), then the
improper integral of f over la, co) is

cR
fe)de= = Ja
f(r) de.

If fe) is continuous on (-co, b], then the


improper integral of f over (-o, b| is

fe)da dar = _lim flc) da.


-oo R-0o.

Since we are dealing with limits, we are


interested in convergence and divergence of the
improper integral. If the limit exists and is a
finite number, we say the improper integral
converges. Otherwise, we say the improper
integral diverges, which we capture in the
following definition.
Definition 2.53. Convergence and
Divergence. If the limit exists and is a finite
number, we say the improper integral
converges.

If the limit is too or does not exist, we say


the improper integral diverges.

To intuitive
get an (though not
completely
correct) interpretation of improper integrals, we

attemptto analyze a
fr)dr graphically. Here
assume f(r) is continuous on la, oo):

fr) fa)
Area Area
R X

We let R be a fixed number in [a, co). Then by


taking the limit as R approaches oo, we get the
improper integral:
R
flr) da = lim f(z) da.
R+00 Ja

We can then apply the Fundamental Theorem of


Calculus to the last integral as f(«) is continuous
on the closed interval [a, R.
We next define the improper integral for the
interval (-co, co).
Definition 2.54. Improper Integrals
Two Infinite Limits of Integration. If
both f) da and
a fle) dr are

convergent, then the improper integral of f


over (-co, co) is

e) da=a)da+ e)da

The above definition requires both of the


integrals
f ) dr and fe) da

to be convergent for fT)dr to also be

convergent. f either of f(«)dr or

fo)de is divergent, then so i s f ( r ) dr.


Example2.55. Improper Integral-
One Infinite Limit of Integration.

Determine whether
de is
convergent or divergent.
V Solution
*************************

Using the definition for improper


integrals we write this as:

- dx =im d a = lin
R
=
lim In|R -

In|1|

Therefore, the integral is


divergent.

Example2.56. Improper Integral -

Two
Infinite Limits of Integration.

Determine whether sin(z") dz is

convergent or divergent.
VSolution
***********************

We must
compute both x
sin(r2) dx
and
-oo
r
sin(z*) dx. Note that we don't
have to split the integral up at 0, any
finite value a will work. First we
compute the indefinite integral. Let
u =
2, then du =
2ar dæ and hence,

zsin(e) dz =sin(u) du=-o


Using the definition of improper
integral gives:

sin() de = lim sin(r2) de


R+0 J0

lim
2 cos(H") * 2
R-

This limit does not exist since cos


oscillates between -1 and +1. In
particular, cos does not approach any
particular value as gets larger and
larger. Thus, sin(*) dr diverges,

and hence, the integral a sin(a*) dr


2.7.2 Improper Integrals:
Discontinuities
there is discontinuity
a in a, b) or at an
When
then the improper integral is as
endpoint,
follows.

Definition 2.57. {Improper Integrals


Discontinuities on Integration Bounds} If f(c)
is continuous on (a, b], then the improper
(a, b] is
integral of f over

flr) da = lim f ) da.


R>a JR

If f(r) is continuous on la, b), then the


improper integral of f over [a, b) is

b
pR
f(r) da = lim f ) da.
R+6

Definition 2.53 on convergence and divergence


***etesestm

of an improper integral holds here as well: If the


limit above exists and is a finite number, we
say
the improper integral converges. Otherwise, we
say the improper integral diverges.
Definition 2.58. Improper Integrals
Discontinuities Within Integration
Interval. Iff has a discontinuity at z = c

where c e la, b], and both f)dr and

f() da are convergent, then f over [a, b]

f(c) da =
fl)dr+ fa) dr.

Again, we can get an intuitive sense of this

concept by analyzing flz) de graphically.

Here assume f(z) is continuous on (a, b] but


discontinuous at c = a:

fo)
Area
X R F
We let R be a fixed number in (a,b). Then byy
taking the limit as R approaches a from the
right, we get the improper integral:

f) da =
lim
Ra RR fla) da.
Example2.59. A Divergent
Integral.

Determine if dz is convergent
or divergent.
V SolutiOn

The function f(r) =


1/* has a
discontinuity at r = 0, which lies
in [-1,1]. We must compute
. d and J0 d. Lets
start with d
J0

dx = lim da= linD


R+0T JR R

which diverges to +oo. Therefore,


1
da is divergent since one

ofd andd is
divergent.
Example2.60. Integral of the
Logarithm.

Determine if Inc dr is convergent or


Jo
divergent. Evaluate it if it is convergent.
VSolution
Note that f(r) = Inx is discontinuous
at the endpoint a =0. We first use
integration by Parts to compute
Inx dr. VWe let u = lnr and du = da.

Then du (1/2)dt, v
= =
z, giving
Inr dr = z - -

= rna -

1d
cnr - c+C

Now using the definition of improper

integral for Inzde

Inrdr = lim Inc dr = lim(


R+0 JR R +0
= -1 - lim (Rln R) + lim
R0 R0

Note that lim R =0. We next compute


R+0+

lim (Rln R). First, we rewrite the


R0
expression as follows:
In R
lim (Rn R) = lim
+0 R0 1/R

Now the limit is of the indeterminate


and I'Hôpital's Rule can
type (-o)/(oo)
be applied.

In R 1/R
lim (Rln R) R0
lim
=
lim
R0T 1/R R0-1/R

Thus, lim
R0T
(Rln R) =0. Thus

Ine dr =-1,

and the integral is convergent to -1.

Graphically, one might interpret this to


mean that the net area under lnc on
0,1 is -1 (the area in this case lies
below the r-axis).
f(x) = In(x)
Example2.61. Integral of a Square
Root.

is convergent or
Determine if A
it if it is convergent.
divergent. Evaluate
vSolution
******

Note that=is discontinuous at the


V4-
4. We use a u-substitution
endpoint z=

to compute J We let u = 4 - z ,
V4-
then du = -da, giving:

dr du
ul/2

-- -u1/2 du
=-2(u)/2 +C
- 2 4 z +C

Now using the definition of improper


da
integrals for j, y-
A 2 4 ) = lim
R4
Example2.62. Improper Integral
da
Determine if is convergent or
(x-1)
divergent. Evaluate it if it is convergent.
Solution

1
Note that f(a) =- is
(-1)1/3

discontinuous at the endpoint a = 1.

We first use substitution to find


do
We let u = * -

1. Then
( -1)1/3
du = dz, 8iving
da du
= - /u1u
ul/3
(-1)/3

Now using thedefinition of improper


dx
integral for
J1 (-1)
da p2
= lim
1)1/3 R1 JR (x - 1)
3

R1
lim R-1)

and the integral is convergent to.


Graphically, one might interpret this to

mean that the net area under


on [1,2] is
(-1)1/3
y--1)

A-

2.7.3 p-Integrals
1
Integrals of the form come up again in the
study of series. These integrals can be either
classified as an improper integral with an infinite
limit of integration, da, or as aan

improper integral with discontinuity at z = 0,

da. In asymptotic analysis, it is useful to


know when either of these intervals converge or
diverge.

Theorem 2.63. p-Test for Infinite Limit.


Fora>0:
1
i fp> 1, then d r COnverges.

i. Ifp<1,then dx diverges.
Theorem 2.63. p-Test for Infinite Limit.
Fora>0:
i. fp> 1, then d z converges.

i. fp1, then a
dax diverges

Proof

i. Ifp> 1, we have
R P

d=im lim -P
ii. Ifp1, the above tells us thatthe
resulting limit is infinite.

Theorem 2.64. p-Test for Discontinuity.


For a >0:

p<1,<1,then da converges.
ii. Ifp2 1, then da diverges.

Proof
* *******
Proof

i. If p<1, we have to
1
da = lim = lim a
Jo R+0+ 1-P

ii. If p 2 1, the above tells us that the


resulting limit is infinite.

With Example 2.55 and


Example 2.59. you have
already seen how the p-Test is applied. For good
measure, here is one more example.

Example2.65. p-Test.
Determine if the following integrals are
convergent or divergent.

1 da

da

Solution

1. This is ap-integral with an infinite


upper limit of and
integration
P = 3> 1. Therefore, by the p-
1
Test for Infinite Limit, da

converges.
We classify d as ap-
integral with a discontinuity at
a =0
and p =
42 1. Thus, by the
p-Test for Discontinuity, the
integral diverges.

2.7.4 Comparison Test


The following test allows us to determine
convergence/divergence information about
improper integrals that are hard to compute by
comparing them to easier ones. We state the
test for la, oo), but similar versions hold for the
other improper integrals.

Theorem 2.66. Comparison Test for


Improper Integrals. Assume that
flr) 2 g{a) 2 0fora 2a.
1. f f r ) dz \thmfontfconverges), then

g) dr also Ithmfontfconverges

2. f glr) da lthmfont{diverges}, then

Sa)dr also Ithmfont{diverges)


Informally, (i) says that if f(z) is larger than glz),
and the area under f(r) is finite (converges),
then the area g(z) must also be finite
under
(converges). Informally, (ii) says that f(a)
if is
is
larger than g(z), and the area under g(«)
infinite (diverges), then the area under f(c)
must also be infinite (diverges).

g)
Area underf
Area under g

Example2.67. Comparison Test.


CO da converges.
Show that 2
SolutioOn
We use the Comparison Test to show
that it converges. Note that
0cos a<1and hence
0 Cos
a2

Thus, taking f(z) =


1/r* and
g(a) cos z/a*
= we have
g(r) 2 0. One can easily see that
fl)2
d a converges. Therefore,
cos
dx also converges
Exercise 2.7.1.
******

Determine whether the following improper


integrals are convergent or divergent.
Evaluate those that are convergent.
a. d a Answer Solution

Converges to n/2

b. da Answer *************
Solution
J0 2+1
Divergent (to oo)

e (cos + sin z) dx, Answer


Solution
*********************

Converges to 1

d. T/2 e r Solution
sec d
Jo

Divergent (to co)

dz
Answer Solution
(4 )2/5

Converges to ( 4 )
da Answer Solution
dz

Converges to 1

8 dc Answer Solution

Divergent

h J0
3 da Answer Solution

Converges to 1/3

1 da Answer Solution
zln a)?

Divergent

sin?() da Answer

Solution
T
Converges to 2-277

do ASwer Solution

Converges to r
Converges to T

d Answer Solution
2+1

Divergent

Exercise 2.7.2.

Provethat the integral dx is

convergentif p> 1 and divergernt if


0 <p1.
Solution

Since

d r = lim
R0
lim-pli
R-P

So if p> 1, we see that the


integral
converges. However for 0< p<1,the
integral diverges.
Suppose that p > 0. Find all values of p for

which d converges.
Answer Solution
**************** *****************

We wish to find all values of p for which

theintegral dz converges. We
know that

JdiJ, d = lim
R+0t Inaln|
R+0 JR

and so for P = 1 the integral diverges


For p>0,41, we have

da 1 p
+ C.

We are left with two cases: Ifp> 1, then

o d T = lim da = lim
R0 1 -

which diverges. If < p <1, then

1
dr R0 JR
lim dn =lim
R0 1-p

which converges to Thus, the


1-p
integral converges for all p e (0, 1), and
diverges for all p2 1.
P
d=R0
lim JR
de R-0t
lin 1-p
which converges to Thus, the

integral converges for all p e (0,1), and


diverges for all p 2 1.

Exercise 2.7.4

Show that sin


aya+1) dx Converges

Solution

We now show that sin


dr
+1)
J1
converges using the comparison test.
First notice that 0 sin' z < 1 for all .
Therefore,

sin 1
al +1) (yT+1) 3/2 +

Since

daim
Ro0
dz
3/2 lim
Roo

we conclude that sin


do
a(T+1)
converges.

You might also like