Vous êtes sur la page 1sur 32

(Section 2.3: Limits and Infinity I) 2.3.

SECTION 2.3: LIMITS AND INFINITY I


PART A: INFINITY
The topic of infinity is a very rich one, and it has different interpretations.
(See Footnote 1 and Wikipedia, for example.) We will not consider infinity to be a
real number, although the real number system can be extended to include infinity.
The symbol for infinity is  . We write negative infinity as   . (See Footnote 2.)
When something approaches  along the real number line, it (generally) increases
without bound. If you give me any real number, this thing will exceed that number.
We will make this idea more precise in a later section.
Something approaching   (generally) decreases without bound.

PART B: NOTATION FOR NONEXISTENT LIMITS


We say that a limit exists  it exists as a single real number.
If it is appropriate to write that a limit is  or   , then we will do so, even
though the limit still technically does not exist. (However, we can say that the limit
exists in the extended reals; See Footnote 1.)
If a limit does not exist, and  and  are inappropriate, then we will write
DNE. That is, the limit does not even exist in the extended reals.

(Section 2.3: Limits and Infinity I) 2.3.2


PART C: HORIZONTAL ASYMPTOTES (HAs) and LONG-RUN LIMITS
Example 1 (Reciprocal Function: A Graph with One HA)

()

Let f x =

()

1
. Consider the graph of y = f x .
x

(Figure 2.3.a)
Lets experiment with a table:

 

()

f x =

1
x

0

100
1

100

10
1

10

()

1

1

10
1
10



100
1
100

0

()

We will refer to lim f x and lim f x as long-run limits.


x

x 

()

Here, we can write: lim f x = 0 .


x

()

In words: the limit of f x as x approaches infinity is 0.

()

That is, as x increases without bound, f x approaches 0.

()

We will give a more rigorous definition of lim f x in a later


x

section.

()

Similarly, lim f x = 0 .
x 

()

In words: the limit of f x as x approaches negative infinity is 0.

()

That is, as x decreases without bound, f x approaches 0.

(Section 2.3: Limits and Infinity I) 2.3.3


As a consequence, the graph of y = 0 (i.e., the x-axis) is a

()

horizontal asymptote (HA) for the graph of y = f x .


An asymptote is a line that a graph approaches; it is usually drawn as a
dashed line.

1
is a rotated hyperbola, which is a type of conic
x
section with two branches. The asymptotes of the hyperbola are the
coordinate axes (the x- and y-axes).

Note: The graph of y =

Using Long-Run Limits to Find Horizontal Asymptotes (HAs)

()

The graph of y = f x has a horizontal asymptote (HA) at y = L


(for a real constant L) 

()
lim f ( x ) = L ).
x 

( lim f x = L , or
x

The graph of a function can have at most two HAs.


Example 2 (A Graph with Two HAs)

()

Let f x =

x
x

. We saw this in Example 13 in Section 2.1.

(Figure 2.3.b)

()
lim f ( x ) = 1 .
x 

lim f x = 1 , and

x

There are two HAs, at y = 1 and y = 1 .


(Usually, when a graph exhibits this kind of flatness, we dont
even bother drawing in the dashed lines.)

(Section 2.3: Limits and Infinity I) 2.3.4


Example 3 (A Graph with No HAs)
Let f x = x + 3 .

()

(Figure 2.3.c)

()

We can write: lim f x =  .


x

()

This is because, as x increases without bound, the f x

function values also (generally) increase without bound.


Note: The use of = here may be technically inappropriate,
but it is commonly accepted.

()

Similarly, lim f x =   .
x 

()

This is because, as x decreases without bound, the f x

function values also (generally) decrease without bound.


Because neither long-run limit exists, the graph has no HAs.
Example 4 (Another Way for a Long-Run Limit to Not Exist)
Let f x = sin x .

()

(Figure 2.3.d)

()

()

lim f x does not exist (DNE). Neither does lim f x . No HAs here.

x

()

x 

This is because the f x function values do not approach a single real


constant as x approaches infinity, nor as x approaches negative
infinity. The function values oscillate between 1 and 1. We cannot
even say that the limit is  or   , because the function values are
safely bounded between 1 and 1 for all real values of x.

(Section 2.3: Limits and Infinity I) 2.3.5


Example 5 (A Graph that Crosses Over Its HA)

()

Let f x =

sin x
.
x

(Figure 2.3.e)

()

lim f x = 0 .

x

We will prove this using the Squeeze (Sandwich) Theorem in a later


section.
If a graph has one or two HAs, the graph may or may not intersect with
1
(touch) a HA. In Example 1, with f x = , the graph didnt.
x
x
, and in this example, the graph did.
In Example 2, with f x =
x
In this example, the graph unceasingly weaves above and below the HA at
y = 0 as x   .

()

()

( )

Note: We will explain the hole at 0, 1 in Chapter 3. It turns out to be quite


important!

(Section 2.3: Limits and Infinity I) 2.3.6

PART D : THE LIMIT FORMS

()

1
1
AND



Review Example 1 on f x =

1
:
x

(Figure 2.3.f)

 

()

f x =

1
x

0

100
1

100

10
1

10

1

10

100



1

1
10

1
100

0

1
1
= 0 , and lim
= 0.
x x
x  x
1
We can also write:  0 as x   , and also as x    .
x

We concluded: lim

More generally, it is true that, for functions N and D,

( )  0 if  N x  1, and D x    .
() 
 ( )
D ( x)

N x

This is true whether we are considering all of the indicated limits as x   ,


as x    , or as x  a (or as x  a + , or as x  a  ) for some real constant a.
Also,
N x
 0 if  N x  1, and D x     .
D x

()
()

()

()

1
1
and
yield 0 as a limit.


1
1
We will write: Limit Form  0 . Also, Limit Form
 0.



We will say that the Limit Forms

(Section 2.3: Limits and Infinity I) 2.3.7


1
Example 6  Limit Form 



Evaluate lim

x  3

, also written as lim

x 

1
.
x1/3

Solution Method
Observe that 1 1 as x   .
(See Section 2.1, Example 3; also, Example 2 in this section.)
Also,

x   as x   .
As x increases without bound, so does its cube root.
3

It helps to consider the graph of y = 3 x :

(Figure 2.3.g)
Therefore, the Limit Form

1
1
applies here, and lim
= 0.
x 3

x

Solution

lim

x 3

1
x


1
=0
Limit
Form

 

A graph can help demonstrate (?) this.


Consider the graph of y =

1
3

(Figure 2.3.h)
Unfortunately, the HA at y = 0 might not have been clear in the
absence of the dashed line there. Be careful about living by
graphing calculators alone!

(Section 2.3: Limits and Infinity I) 2.3.8


1 
Example 7  Limit Form
  

lim

x  3


1 
=0
Limit
Form

  

Example 8 (No Limit Form)

lim

x 

, also written as lim

x
This is because

x  

1
, does not exist (DNE).
x1/ 2

x is not defined as a real quantity whenever x < 0 .

Note: We do not call

1
a Limit Form. (See Footnote 3.)
DNE

The graph of y = x :

The graph of y =

(Figure 2.3.i)

(Figure 2.3.j)

Limit Forms and the Limit Properties from Section 2.2


Remember that the sum of (a finite number of) limits equals the limit of the
sum, provided the limits exist as real constants.

For example, the Limit Form 2 + 3  5 .


Also,

( )
the Limit Form ( 2  3)  6 , and

the Limit Form 2  3  1 ,

 2
2
the Limit Form    .
3
 3

(Section 2.3: Limits and Infinity I) 2.3.9


PART E: VARIATIONS ON LIMIT FORMS
Approaching 0 in Different Ways
More precisely, we can write: Limit Form

1
 0+


This is because, for functions N and D,


N x
if  N x  1, and D x    , then
approaches 0
D x

()

()
()

()

from the right (or from above, since were dealing with
y function values).
The 0+ notation can help us if we want to be more descriptive
about how the limit is approaching 0, or if this limit analysis is
part of a larger limit problem.
Also:

Limit Form

1
 0


()

See Example 5 on f x =
written as 0+ , nor as 0 .

sin x
for a case where a limit is 0, but it cannot be
x

Signs are Handled as Expected in Products and Quotients

1
 0

1
Limit Form
 0+


Limit Form

The graph of y =

1
; see what happens as x   and as x    :
x

(Figure 2.3.k)

(Section 2.3: Limits and Infinity I) 2.3.10


The Rescaling Property of Some Limit Forms
This property applies to Limit Forms that do not yield a limit that is a real
nonzero constant; that is, the property applies to Limit Forms that yield:
0,  , or   (or, for that matter, DNE).
If such a Limit Form is rescaled by multiplying it by a positive real
constant, then the resulting Limit Form will yield the same limit as the
original.

1
yields 0 (or, more precisely, 0+ ) as a limit,

2 3
1/ 5
so the Limit Forms , , and
all yield 0 (or 0+ ) as a limit, also.
 


For example, the Limit Form

1
yields 0 (or, more precisely, 0 ) as a limit, so

2 3
1 / 5
,
, and
all yield 0 (or 0 ) as a limit, also.
the Limit Forms
 


Likewise, the Limit Form

The Rescaling Property does not apply to Limit Forms that yield a limit that
1
is a real nonzero constant. For example, the Limit Form yields 1, but the
1
2
Limit Form yields 2. We will discuss similar Limit Form properties in a
1
later section.

Limit Form 1    .
As a consequence, by our rules for signs and rescaling, we can also
write:

Limit Form c     for any positive real constant c.

( ( ))

Limit Form c       for any positive real constant c.

Limit Form c      for any negative real constant c.

( ( ))

Limit Form c      for any negative real constant c.

(Section 2.3: Limits and Infinity I) 2.3.11


Some Limit Forms Involving Infinity
Remember our rules for signs and rescaling.
Go ahead and experiment with numbers!
Limit Form
Limit Form
Limit Form
Limit Form
Limit Form

1
 0+

1
  (We will see this in Section 2.4.)
0+


1


0+
0+
 0+


( )
Limit Form (    )  
Limit Form 1   

( )
Limit Form (   + c )    for any real constant c

Limit Form  + c   for any real constant c

(
)
Limit Form (     )   
Limit Form  +   

Limit Form    
Limit Form     0+
Limit Form 0  0
Classic Indeterminate Limit Forms
(We will discuss these in later sections.)

0 
, , 0   ,    ,  0 , 00 , 1
0 
Note: Limit Form 2   ; it is not indeterminate.


 1
Limit Form    0 ; it is not indeterminate.
 2

(Section 2.3: Limits and Infinity I) 2.3.12

PART F : "LONG - RUN" LIMIT RULES FOR

In Example 1, we saw that: lim

x

c
xk

1
1
= 0 , and lim
= 0.
x  x
x

In Examples 6 and 7, we saw that: lim

x 

1
1
=
0
lim
= 0.
,
and
x   x1/3
x1/3

In the graphs in Example 8, we see that: lim

x 

lim

x  

1
= 0 , but
x1/ 2

1
does not exist (DNE). Warning 1: Be aware of this issue!
x1/ 2

More generally:

If c is a real constant, and k is a positive rational constant,


c
then: lim k = 0 .
x x
c
Also, lim k = 0 if x k is defined as a real quantity whenever x < 0 ;
x   x
c
otherwise, lim k does not exist (DNE).
x   x
(See Footnote 4 for the justification. See also Footnote 5.)
Think About It: What if k < 0 ?
Example 9

lim

x 

2
2
=
0
lim
= 0.
,
and
x   x 3
x3

Example 10


= 0 , but
x  x 3/ 4

lim 3/ 4 , also written as lim
x   x
x 
lim

( x)
4

, does not exist (DNE).

(Section 2.3: Limits and Infinity I) 2.3.13


PART G: LONG-RUN LIMITS FOR POLYNOMIALS
Example 11 (Constant Function)

lim 2 = 2 .

x

A graph can demonstrate this.

()

Think: f x = 2 . Consider the graph of y = 2 .

(Figure 2.3.letter l)
There is a HA at y = 2 , but we will omit the dashed line here.
(See Example 3 in Section 2.1.)
More generally:
If c is a real constant, then:

lim c = c , and lim c = c .


x 

x

The graph of y = c has itself as its sole HA.


On the other hand:
Graphs of nonconstant polynomial functions have no HAs.
These graphs explode in the long run to the right and to the left.
For such a function f,

()
Also, lim f ( x ) =  or   .
x 
lim f x =  or   .

x

Granted, explode may be too strong a term for our linear function
given by f x = x + 3 in Example 3.

()

(Section 2.3: Limits and Infinity I) 2.3.14

How can we tell if a limit is  or   ?


A key tool is something we will call Dominant Term Substitution.
When analyzing a long-run limit, if a and b are integers, then
x a dominates x b  a > b .
(See Footnotes 6 and 7 for more on dominance.)
When analyzing a long-run limit,
the dominant term of a polynomial is the leading term.
Dominant Term Substitution
The limit of an expression is the same as the limit of its
dominant term. (By limit here, we include  and   .)
Note: Do not use dominance if an expression is not defined
as a real quantity when considering the limit. This is never
an issue with polynomials.
Why does this work? The Factoring Principle of Dominance:
If the dominant term is factored out of an expression,
we obtain the dominant term times something approaching 1.
(This something can then be safely removed from the limit
analysis).
We will see this in Example 12.
Example 12 (A Question of Dominance)

Evaluate lim x 8  x 6 .
x

Solution Method
There is a tension between the two terms, x 8 and  x 6 , because
x 8 approaches  as x   , while  x 6 approaches   .
(Review graphs of polynomials such as these in Precalculus.)
Is someone in charge here? In the long run, yes! The dominant
term here is x 8 .
Warning 2: x 8 is the leading term because it is the term of
higher degree, not because it is written first.

(Section 2.3: Limits and Infinity I) 2.3.15


In the long run, its impact renders the other term,  x 6 ,
negligible by comparison. x 8 acts like a bully. It explodes
more dramatically  x 6 than does.
Solution

lim x 8  x 6 = lim x 8

x

x

=
A More Rigorous Solution: The Factoring Principle of Dominance
We can factor out the dominant term, x 8 (not the GCF), and
c
we can apply the Long-Run Limit Rules for k from Part F.
x

x6 
8
6
8
lim x  x = lim x  1  8 
x
x
x



1
= lim x 8  1  2 
x

x
Now, as x   , x 8   , and
1
1
 0 , and thus 1  2  1 , so we
2
x
x
have the Limit Form  1   .


 1
Essentially, 
x8  1   2  .
x 
 


0



1


=



A graph can demonstrate this.


Consider the graph of y = x 8  x 6 :

(Figure 2.3.m)

(Section 2.3: Limits and Infinity I) 2.3.16


PART H: LONG-RUN LIMITS FOR RATIONAL FUNCTIONS
Twin Limits Property of Rational Functions

()

If f is a rational function with its implied domain, then lim f x = L


x

()

(for some real constant L) implies lim f x = L , and vice-versa.


x 

()

That is, if there is a long-run limit value L for f x as x


explodes in one direction along the x-axis, then that value must
also be the long-run limit value for f x as x explodes in the other

()

direction.
Therefore, the graph of f can have either no HAs or one HA, but
not two. (See Examples 1, 3, and 11.)
If f is not rational, then its graph could have zero, one, or two HAs.
Example 13

4x 3 + x  1
Evaluate lim f x , where f x =
.
x
5x 3  2x

()

()

Solution Method 1 (A Rigorous Solution)

()
Let D ( x ) = 5x

Let N x = 4x 3 + x  1, the numerator, which is a polynomial.


3

 2x , the denominator, which is a polynomial that is

not the 0 polynomial.

()

Then, f x =

( ).
D ( x)

N x

()

()

We will divide (each term of) N x and D x by the dominant term

()

in D x . Here, it is x 3 . This will ensure that the new denominator


will approach a nonzero real constant, namely the leading coefficient
of D x . Here, it is 5.

()

We then use the Long-Run Limit Rules for

c
from Part F.
xk

(Section 2.3: Limits and Infinity I) 2.3.17


Warning 3: This technique is intended for use in evaluating
long-run limits as x   or x    , not as x approaches a
constant. (Note: Another variable besides x could be involved.)
Solution 1 (A Rigorous Solution)

4x 3 + x  1
lim f x = lim
x
x 5x 3  2x

()

x
1
4x 3
+ 3 3
3
x
= lim x 3 x
x
2x
5x

x3
x3
0
0


1
1
4+ 2  3
x
x
= lim
x
2
5 2
x


(Warning 4: If you use "  0",


remember the arrows!)

0

4
, or 0.8
5

A graph can demonstrate this.

()

Because lim f x =
x 

4
4
, we have a HA at y = .
5
5

Because f is a rational function, it is the only HA.


4
We can also show that lim f x = using a similar solution.
x  
5

()

(Figure 2.3.n)

(Section 2.3: Limits and Infinity I) 2.3.18


Solution Method 2 (The Short Cut: Dominant Term Substitution)

()

()

Again, N x = 4x 3 + x  1, and D x = 5x 3  2x .

()

Then, f x =

( ).
D ( x)

N x

()

In our limit analysis, we may replace N x with its dominant term,

()

4x 3 , and we may replace D x with its dominant term, 5x 3 . This is


justified by the Factoring Principle of Dominance (which also gives
rise to a rigorous method, but Solution 1 is probably easier):

1

1
1 
3 
4x
1
+


4x 2 4x 3 
4x 3 + x  1
lim
= lim
x 5x 3  2x
x

2 
5x 3  1  2 
 5x


1

Solution 2 (The Short Cut: Dominant Term Substitution)

4x 3 + x  1
lim f x = lim
x
x 5x 3  2x
4x 3
= lim
x 5x 3
4
=
5

()

Super Short Cut: This is because the numerator, 4x 3 + x  1, and


the denominator, 5x 3  2x , are nonzero polynomials of equal degree,
4
namely 3. The ratio of their leading coefficients is .
5

(Section 2.3: Limits and Infinity I) 2.3.19


Example 14

x2  3
Evaluate lim f x , where f x = 3
.
x
x + 4x 2 + 1

()

()

Solution 1 (A Rigorous Solution): Left to the reader! Refer to Example 13.


Solution 2 (The Short Cut: Dominant Term Substitution)

x2  3
lim f x = lim 3
x
x x + 4x 2 + 1
x2
= lim 3
x x

()

= lim

x

1
x

=0
Super Short Cut: This is because the numerator, x 2  3 , has a
degree (i.e., 2) that is less than the degree (i.e., 3) of the denominator,
x 3 + 4x 2 + 1. We say that f is a proper rational function.
(The fraction is bottom heavy.)
How can graphs demonstrate this?

()

The graph of y = f x approaches that of y =


common HA at y = 0 in the long run.

(Figure 2.3.o)

1
and also the
x

(Figure 2.3.p)

(Section 2.3: Limits and Infinity I) 2.3.20


Example 15

 5 + 3x 2 + 6x 3
.
Evaluate lim f x and lim f x , where f x =
x
x 
1 + 3x 2

()

()

()

Solution 1 (A Rigorous Solution): Left to the reader! Refer to Example 13.


Solution 2 (The Short Cut: Dominant Term Substitution)
The numerator,  5 + 3x 2 + 6x 3 , has a degree (i.e., 3) that is greater
than the degree (i.e., 2) of the denominator, 1 + 3x 2 .
(The fraction is top heavy.) We know from this alone that the graph
of f will have no HA. However, because 3 is exactly one more than
2, the graph will have a slant asymptote (SA), also known as an
oblique asymptote.

 5 + 3x 2 + 6x 3
6x 3
lim
= lim
x
x 3x 2
1 + 3x 2
= lim 2x
x

=
 5 + 3x 2 + 6x 3
6x 3
lim
= lim
x  
x   3x 2
1 + 3x 2
= lim 2x
x  

= 
A graph can demonstrate all of this.

(Figure 2.3.q)

(Section 2.3: Limits and Infinity I) 2.3.21


How can we find an equation for the slant asymptote (SA)?
We can use the Long Division technique from Precalculus to
 5 + 3x 2 + 6x 3
reexpress f x =
as:
1 + 3x 2

()

()

f x = 2x + 1 +

 

polynomial
part, p( x )

= 2x + 1 

 2x  6
2
1
3x
+

proper rational
part, r ( x )

2x + 6
3x 2 + 1

Heres the work:

We can now stop the division process here, because the degree of the
new dividend is less than the degree of the divisor. The degree of
 2x  6 is 1, which is less than the degree of 3x 2 + 0x + 1 , which is 2;
 2x  6
this means that  2x  6 is our remainder, and r x =
3x 2 + 1
corresponds to a proper rational function. As a result, r x  0 as

()

()

x   and as x    ; see Case 1 in our upcoming summary.


(The proper rational part decays in the long run.) Therefore, the
graph of y = f x approaches the graph of y = p x as x   and
as x    .

()

()

()

Here, the graph of y = f x approaches the graph of y = 2x + 1 ,


which is an equation for our slant asymptote (SA) (or oblique
asymptote) for the graph of f.

(Section 2.3: Limits and Infinity I) 2.3.22


Zoom Out Property of HAs and SAs
If we keep expanding the scope of the window of a grapher, then a
graph with a HA or a SA will generally look more and more like the
HA or SA.
Example 16
Analyze the long-run behavior of the graph of
 4x 7 + 12x 6 + 5x 4  23x 3 + 11
f x =
.
4x 3  5

()

Solution
By Dominant Term Substitution, we find that:

 4x 7
lim f x = lim
x
x 4x 3

 4x 7
lim f x = lim
x 
x  4x 3

()

()

= lim  x 4

= lim  x 4

= 

= 

x 

x

If we perform Long Division, we obtain:

()

1
.
4x  5
The graph of y =  x 4 + 3x 3  2 is a nonlinear asymptote that the
graph of f approaches in the long run. Observe that the leading
f x =  x 4 + 3x 3  2 +

term is  x 4 , which was the star of our Dominant Term Substitution.


Based on this alone, we know that the graph approaches the shape of a
downward-opening bowl if we zoom out. (Review long-run
behaviors of polynomial graphs in Precalculus.)

()

Graph of y = f x

(Figure 2.3.r)

()

Graph of p x =  x 4 + 3x 3  2

(Figure 2.3.s)

(Section 2.3: Limits and Infinity I) 2.3.23


In summary:
Super Short Cuts for Evaluating Long-Run Limits for Rational Functions,
and for Finding HAs for their Graphs

()

Assume f x =

( ) on its implied domain, where


D ( x)

N x

()
()
D ( x ) is not the zero polynomial.
Let deg ( N ) and deg ( D ) be their respective degrees.
N x and D x are polynomials, and

Case 1

( )

( )
the x-axis ( y = 0 ) is the only HA of its graph.
lim f ( x ) = 0 , and lim f ( x ) = 0 .
x
x 

If deg N < deg D , then f is a proper rational function, and

()

See Example 14, where f x =

x2  3
.
x 3 + 4x 2 + 1

(Think: Bottom heavy.)


Case 2

( )

( )

If deg N = deg D , then y = L is the only HA of the graph of f,


where L =

( ).
the leading coefficient of D ( x )

the leading coefficient of N x

(Think: The ratio of the leading coefficients.)

()

()

lim f x = L , and lim f x = L .

x

x 

4x 3 + x  1
.
See Example 13, where f x =
5x 3  2x

()

(to be continued)

(Section 2.3: Limits and Infinity I) 2.3.24

Case 3

( )

( )

If deg N > deg D , then the graph of f has no HAs.


In particular,

()
lim f ( x ) =  or   .
x 

lim f x =  or   , and

x

( )

( )

If deg N = deg D + 1, then the graph does have a


slant asymptote (SA), also known as an oblique asymptote.

 5 + 3x 2 + 6x 3
See Example 15, where f x =
. (Think: Top heavy.)
1 + 3x 2

()

PART I: LONG-RUN LIMITS FOR ALGEBRAIC FUNCTIONS


If we are considering a long-run limit as x   , then we will say that,
for rational constants a and b, x a dominates x b  a > b .
The same is true as x    , provided that x a and x b are defined as real
quantities whenever x < 0 .
Example 17

lim 5x 7 / 2  2x 3 + 1 + x  2 = lim 5x 7 / 2

x

x

=
This analysis is appropriate, because 5x 7 / 2 is the dominant term as x   .
(It helps to think of 1 as x 0 , even though 00 is typically considered to be
undefined.)
Example 18

lim 5x 7 / 2  2x 3 + 1 + x  2 does not exist (DNE), not even as  or   .

x 

This is because 5x
whenever x < 0 .

7/2

( x ) , is undefined as a real quantity


7

, also written as 5

(Section 2.3: Limits and Infinity I) 2.3.25


When taking the limit of an algebraic function, Dominant Term Substitution
may also be applied to radicands and bases of powers.
A radical is a kind of power. For example,

x = x1/ 2 .

(See Footnotes 8 and 9 on how dominance can fail us for non-algebraic


functions.)
Example 19
Evaluate lim

4x 3  x10  5

( x + 3)

x

Solution Method (The Short Cut: Dominant Term Substitution)


Note: Although the radicand, x10  5 , can be negative in value, we
have a different situation from the previous example. It eventually
stays nonnegative as x   , in the sense that x10  5  0 on the
x-interval c,  for some real constant c. Therefore, the radical

eventually yields real values as x   .


In the radicand, x10  5 , x10 dominates  5 .
In the power-base, x + 3, x dominates 3.
Solution

lim

x

4x 3  x10  5

( x + 3)

= lim

x

4x 3  x10

( x)

4x 3  x 5
= lim
x
x2
 x5
= lim 2
x x
= lim  x 3
x

= 

(Section 2.3: Limits and Infinity I) 2.3.26


A More Rigorous Solution
We can apply the Factoring Principle of Dominance.

lim

x

4x 3  x10  5

( x + 3)


5

4x 3  x10  1  10

x

1

= lim



 
3

x 1 + 
x 

 





1


x

= lim

4x 3  x10
2

 x 
4x 3  x 5
= lim
x
x2
 4x 3

 x 5  5 + 1
x

= lim
x
x2
1

 4

 x 5   2 + 1
 x

= lim
x
x2
 x5
= lim 2
x x
= lim  x 3
x

x

= 
Note: If were to consider lim

x 

x10 =

(x )
5

4x 3  x10  5

( x + 3)

, then remember that

= x 5 =  x 5 whenever x < 0 . When considering the limit as

x    , one may assume that x < 0 in the limit analysis. (Here, we


technically assume x <  10 5 due to the domain issue.)

(Section 2.3: Limits and Infinity I) 2.3.27


PART J: A WORD PROBLEM
Example 20 (Salt water mixtures)
A deep well contains 10 gallons of pure water at noon. Starting at noon, a
salt-water mixture is poured into the well at the rate of 2 gallons per
minute. The mixture has a salt concentration of 0.3 pounds of salt per gallon.
(Assume that the salt content is well mixed with the water in the incoming
mixture and in the well at all times.)

()

a) Find an expression for C t , the concentration of salt in the salt-water


mixture in the well t minutes after noon, where t  0 .
Solution

()

Let V t be the volume (in gallons) of the salt-water mixture in


the well t minutes after noon, where t  0 .
t minutes after noon, 2t gallons of the mixture have been
poured into the well. The well started with 10 gallons of
pure water, so the total volume in the well is given by:

()

V t = 10 + 2t (in gal).

()

Let S t be the weight (in pounds) of the salt in the salt-water


mixture in the well t minutes after noon, where t  0 .
All of the salt in the well has been poured in. We have:

 0.3 lbs   2 gal 


S t =

 t min = 0.6t (in lbs).
 gal   min 

()

(Section 2.3: Limits and Infinity I) 2.3.28


Then,

()

C t =

()
V (t )
S t

 Multiply by 10, though 5 is better.


0.6t
10 + 2t  Multiply by 10, though 5 is better.
6t
=
100 + 20t
 lbs 
3t
=
in
50 + 10t  gal 
=

()

b) Find lim C t , and interpret the result. Discuss the realism of this
t

problem.
We will use Case 2 of the Short Cuts for Evaluating LongRun Limits for Rational Functions on Page 2.3.23.

()

3t
50 + 10t
3 lbs
lbs
=
, or 0.3
10 gal
gal

lim C t =

t

The concentration of salt in the salt-water mixture in the well


lbs
in the long run, the same as for the
approaches 0.3
gal
incoming mixture. However, this calculation assumes that the
well has infinite capacity, which is probably not the case in
practice. Also, it assumes an unlimited supply of salt water. It
might be a good approximation for a filled up well of high
capacity, though.

()

Here is the graph of y = C t :

(Figure 2.3.t)

(Section 2.3: Limits and Infinity I) 2.3.29


FOOTNOTES
1. Infinity.
Infinity is not a number in the usual real number system that we will study in calculus.
The affinely extended real number system, denoted by R or   ,   , includes two points
of infinity, one referred to as  (or +  ), and the other referred to as   . (We are
adjoining them to the real number system.) We obtain the two-point compactification of
the real numbers. We never refer to  and   as real numbers, though.
Sometimes,  and   are treated as the same (we collapse them together and identify
them with one another as  ), and we then obtain the one-point compactification of the real
1
numbers, also known as the real projective line. Then, we can write =  , and we can say
0
that the slope of a vertical line is  .
A point at infinity is sometimes added to the complex plane, and it typically corresponds to
the north pole of a Riemann sphere that the complex plane can be thought of as wrapping
around.
There are different ways of looking at having infinitely many numbers. The sets of
integers and rational numbers are called countable sets, while the set of real numbers is called
uncountable. The elements in a countable set can be organized in a one-to-one
correspondence with the positive integers (or counting numbers) 1, 2, 3, or some subset
thereof. You can learn about this in a discrete math class. 19th-century German
mathematician Georg Cantor was a key figure.
See Wikipedia, Encyclopedia Britannica, Rudy Ruckers Infinity and the Mind, William F.
Trench, Introduction to Real Analysis (free online), pp.7-9.
2.  . This symbol is sometimes called a lemniscate, a reference to a type of polar curve we
will analyze much later. It was invented by English mathematician John Wallis in 1657.
See Wikipedia and Encyclopedia Britannica.
3.

1
.
DNE
Although we do not call this a Limit Form, we can at least conclude that the desired limit is
not a real nonzero constant, nor can it be said to be  or   . In other words, the desired
limit must be either 0 or nonexistent (DNE), not even in the sense of  or   .
The desired limit cannot be said to be  or   . (Assume that D is a function.)
As, for example, x   , it is possible that
Likewise, it is possible that

()

1
  only if D x  0 as x   .
D x

()

()

1
   only if D x  0 .
D x

()

The desired limit cannot be a real nonzero constant.


1
1
 L for some real nonzero constant L as, for example, x   , then D x 
If
L
D x

()

(which is a real nonzero constant) as x   .

()

(Section 2.3: Limits and Infinity I) 2.3.30


The following variation of the contrapositive of the preceding if-then statement is also true:
1
If D x does not approach a real nonzero constant, then
does not approach a real
D x

()

()

nonzero constant, either.


1
1
= lim csc x does not exist (DNE), so the
lim
limit situation can yield a
x  sin x x 
DNE
nonexistent limit (DNE).
It is possible to obtain a limit that is 0, even if D x is not approaching  , nor   .

()

 x, if x is a rational value
For example, let: D x = 
.
 x, if x is an irrational value

()

()

Then, lim D x does not exist (DNE). However, lim


x 

x 

1
= 0 . If we extend the real
D x

()

number system to the real projective line, in which we collapse together and identify  and
1
  , then we are effectively dealing with the generic Limit Form , which yields 0 as a

limit.

c
. Assume that c is a real constant (i.e.,
xk
any arbitrary real constant), and k is a positive rational constant.
c
c
Then, x k   as x   . lim k has Limit Form , which implies that the limit is 0
x  x

(regardless of whether c is positive, negative, or 0).
Also, x k    as x    , if x k is defined as a real quantity whenever x < 0 .

4. Justification for Long-Run Limit Rules for

c
c
has Limit Form
, which implies that the limit is 0. If x k is undefined as
k
x   x

c
a real quantity whenever x < 0 , then lim x k does not exist (DNE), and lim k does
x   x
x  
not exist (DNE), even if c = 0 .
Then, lim

5. Irrational Exponents; Roots of Negative Real Numbers. It is true that x k   and


c
lim k = 0 (for any real constant c) for any positive real constant k. But what if k is
x  x
irrational? For example, if k =  , then how do we define something like 2 when x = 2 ?
Remember that  = 3.14159... . Consider the corresponding sequence:
23 = 8
31

23.1 = 210 =
314

10

231  8.57419
157

23.14 = 2 100 = 2 50 =


50

2157  8.81524

(Section 2.3: Limits and Infinity I) 2.3.31


The limit of this sequence (as the number of digits of  approaches  ) is taken to be 2 .

( )

It turns out that 2  8.82498 . However, defining  2

(  2)
(  2)

( )

31
10

( )

is more problematic. For example,

3.1

= 2

31

has ten distinct 10th roots in C, the set of complex numbers. Refer to DeMoivres

. We are looking for a 10th root of  2

31

. From Precalculus, we know that

Theorem for the complex roots of a complex number. See The Math Forum @ Drexel: Ask
Dr. Math, Meaning of Irrational Exponents.
6. Dominant terms. We will say that x d dominates x n as x   for real constants d and n
 d > n . This is because the growth of the (absolute value of) x d makes the growth of the
(absolute value) of x n seem negligible by comparison in the long run. More precisely,
xn
xn
lim d = lim x n d = 0  d > n . (If d > n , the denominator of d explodes more
x  x
x 
x
dramatically than the numerator does., and the limit is 0 as x   .)
xn
Also, lim d = 0  ( d > n , and x n and x d are defined as real quantities
x   x
whenever x < 0 ).
This dominant term analysis can be extended to non-algebraic expressions. For example,
we will see the algebraic expression e x in a later chapter. However, the identification of a
dominant term in a long-run analysis may well depend on whether we are considering a
1
limit as x   or a limit as x    . It turns out that lim x = 0 , so e x dominates 1 if we
x  e
ex
= 0 , so 1
are considering a limit as x   . However, it also turns out that lim
x   1
dominates e x if we are considering a limit as x    .
In sin x + cos x , neither term dominates in the long run. Any nonconstant polynomial will
dominate either term.
7. Dominance and computer science. The issue of dominance is the foundation of the big O
notation used in algorithm analysis, which deals with the running time and space and
memory requirements of computer algorithms. There is particular interest in the long run
behavior of these algorithms as, say, the input size approaches infinity.
8. When dominance fails us, I. We will discuss exponential expressions such as 2 x in a later

2 x +3
2 x +3
2 x  23
2 x  23
.
We
obtain:
lim
=
lim
=
lim
= 8 . If we
x
x  2 x
x  2 x
x  2 x
x 
2
1

chapter. Consider lim

had replaced x + 3 with x in the exponent of 2 x +3 , we would have obtained:


?
2 x +3 
2x
lim
= lim x = 1 , which is incorrect. What went wrong? The Factoring Principle of
x  2 x
x  2
Dominance should not be applied to exponents. Instead, we can consider factoring 2 x out of
2 x +3
2 x +3 : 2 x +3 = 2 x 23 = 2 x 8 . In the numerator of x , we have 2 x times something
2
approaching 8, not 1, and the 8 influences the limit of the overall expression.

( )

()

(Section 2.3: Limits and Infinity I) 2.3.32.


9. When dominance fails us, II. Consider lim

sin x + 

)=

x 

sin x + 
sin x

) . We obtain:

 sin x
(by Sum ID or Unit Circle) = 1 . If we had replaced x + 
x 
x  sin x
sin x
with x in the argument of sin x +  , we would have obtained:
lim

lim

?
sin x +  
sin x
lim
= lim
= 1 , which is incorrect. What went wrong? The Factoring
x 
x

sin x
sin x
Principle of Dominance should not be applied to arguments of non-algebraic functions.
Here, sin x +  is not equivalent to sin x times something approaching 1.

Vous aimerez peut-être aussi