End Behavior

Learning Outcomes

  • Estimate the end behavior of a function as π‘₯ increases or decreases without bound
  • Recognize an oblique asymptote on the graph of a function

The behavior of a function as xβ†’Β±βˆžxβ†’Β±βˆž is called the function’s end behavior. At each of the function’s ends, the function could exhibit one of the following types of behavior:

  1. The function f(x)f(x) approaches a horizontal asymptote y=Ly=L.
  2. The function f(x)β†’βˆžf(x)β†’βˆž or f(x)β†’βˆ’βˆžf(x)β†’βˆ’βˆž.
  3. The function does not approach a finite limit, nor does it approach ∞∞ or βˆ’βˆžβˆ’βˆž. In this case, the function may have some oscillatory behavior.

Let’s consider several classes of functions here and look at the different types of end behaviors for these functions.

End Behavior for Polynomial Functions

Consider the power function f(x)=xnf(x)=xn where nn is a positive integer. From Figure 11 and Figure 12, we see that

limxβ†’βˆžxn=∞;n=1,2,3,β‹―limxβ†’βˆžxn=∞;n=1,2,3,β‹―

and

limxβ†’βˆ’βˆžxn={∞;n=2,4,6,β‹―βˆ’βˆž;n=1,3,5,β‹―limxβ†’βˆ’βˆžxn={∞;n=2,4,6,β‹―βˆ’βˆž;n=1,3,5,β‹―
The functions x2, x4, and x6 are graphed, and it is apparent that as the exponent grows the functions increase more quickly.

Figure 11. For power functions with an even power of nn, limxβ†’βˆžxn=∞=limxβ†’βˆ’βˆžxnlimxβ†’βˆžxn=∞=limxβ†’βˆ’βˆžxn.

 

The functions x, x3, and x5 are graphed, and it is apparent that as the exponent grows the functions increase more quickly.

Figure 12. For power functions with an odd power of nn, limxβ†’βˆžxn=∞limxβ†’βˆžxn=∞ and limxβ†’βˆ’βˆžxn=βˆ’βˆžlimxβ†’βˆ’βˆžxn=βˆ’βˆž.

Using these facts, it is not difficult to evaluate limxβ†’βˆžcxnlimxβ†’βˆžcxn and limxβ†’βˆ’βˆžcxnlimxβ†’βˆ’βˆžcxn, where cc is any constant and nn is a positive integer. If c>0c>0, the graph of y=cxny=cxn is a vertical stretch or compression of y=xny=xn, and therefore

limxβ†’βˆžcxn=limxβ†’βˆžxnlimxβ†’βˆžcxn=limxβ†’βˆžxn and limxβ†’βˆ’βˆžcxn=limxβ†’βˆ’βˆžxnlimxβ†’βˆ’βˆžcxn=limxβ†’βˆ’βˆžxn if c>0c>0

 

If c<0c<0, the graph of y=cxny=cxn is a vertical stretch or compression combined with a reflection about the xx-axis, and therefore

limxβ†’βˆžcxn=βˆ’limxβ†’βˆžxnlimxβ†’βˆžcxn=βˆ’limxβ†’βˆžxn and limxβ†’βˆ’βˆžcxn=βˆ’limxβ†’βˆ’βˆžxnlimxβ†’βˆ’βˆžcxn=βˆ’limxβ†’βˆ’βˆžxn if c<0c<0

 

If c=0,y=cxn=0c=0,y=cxn=0, in which case limxβ†’βˆžcxn=0=limxβ†’βˆ’βˆžcxnlimxβ†’βˆžcxn=0=limxβ†’βˆ’βˆžcxn.

Example: Limits at Infinity for Power Functions

For each function ff, evaluate limxβ†’βˆžf(x)limxβ†’βˆžf(x) and limxβ†’βˆ’βˆžf(x)limxβ†’βˆ’βˆžf(x).

  1. f(x)=βˆ’5x3f(x)=βˆ’5x3
  2. f(x)=2x4f(x)=2x4

Watch the following video to see the worked solution to Example: Limits at Infinity for Power Functions.

Try It

Let f(x)=βˆ’3x4f(x)=βˆ’3x4. Find limxβ†’βˆžf(x)limxβ†’βˆžf(x).

 

 

We now look at how the limits at infinity for power functions can be used to determine limxβ†’Β±βˆžf(x)limxβ†’Β±βˆžf(x) for any polynomial function ff. Consider a polynomial function

f(x)=anxn+anβˆ’1xnβˆ’1+β‹―+a1x+a0f(x)=anxn+anβˆ’1xnβˆ’1+β‹―+a1x+a0

of degree nβ‰₯1nβ‰₯1 so that anβ‰ 0anβ‰ 0. Factoring, we see that

f(x)=anxn(1+anβˆ’1an1x+β‹―+a1an1xnβˆ’1+a0an1xn)f(x)=anxn(1+anβˆ’1an1x+β‹―+a1an1xnβˆ’1+a0an1xn).

As xβ†’Β±βˆžxβ†’Β±βˆž, all the terms inside the parentheses approach zero except the first term. We conclude that

limxβ†’Β±βˆžf(x)=limxβ†’Β±βˆžanxnlimxβ†’Β±βˆžf(x)=limxβ†’Β±βˆžanxn.

For example, the function f(x)=5x3βˆ’3x2+4f(x)=5x3βˆ’3x2+4 behaves like g(x)=5x3g(x)=5x3 as xβ†’Β±βˆžxβ†’Β±βˆž as shown in Figure 13 and Table 1.

Both functions f(x) = 5x3 – 3x2 + 4 and g(x) = 5x3 are plotted. Their behavior for large positive and large negative numbers converges.

Figure 13. The end behavior of a polynomial is determined by the behavior of the term with the largest exponent.

Table 1. A polynomial’s end behavior is determined by the term with the largest exponent.
xx 10 100 1000
f(x)=5x3βˆ’3x2+4f(x)=5x3βˆ’3x2+4 4704 4,970,004 4,997,000,004
g(x)=5x3g(x)=5x3 5000 5,000,000 5,000,000,000
xx -10 -100 -1000
f(x)=5x3βˆ’3x2+4f(x)=5x3βˆ’3x2+4 -5296 -5,029,996 -5,002,999,996
g(x)=5x3g(x)=5x3 -5000 -5,000,000 -5,000,000,000

End Behavior for Algebraic Functions

The end behavior for rational functions and functions involving radicals is a little more complicated than for polynomials. In the example below, we show that the limits at infinity of a rational function f(x)=p(x)q(x)f(x)=p(x)q(x) depend on the relationship between the degree of the numerator and the degree of the denominator. To evaluate the limits at infinity for a rational function, we divide the numerator and denominator by the highest power of xx appearing in the denominator. This determines which term in the overall expression dominates the behavior of the function at large values of xx.

Note that this is not your first encounter with horizontal asymptotes. It may be helpful to recall what you already know about them.

Recall: Horizontal Asymptotes of Rational Functions

The horizontal asymptote of a rational function can be determined by looking at the degrees of the numerator and denominator.

  • Case 1: Degree of numerator is less than degree of denominator: horizontal asymptote at y=0y=0
  • Case 2: Degree of numerator is greater than degree of denominator by one: no horizontal asymptote; slant asymptote.
    • If the degree of the numerator is greater than the degree of the denominator by more than one, the end behavior of the function’s graph will mimic that of the graph of the reduced ratio of leading terms.
  • Case 3: Degree of numerator is equal to degree of denominator: horizontal asymptote at ratio of leading coefficients.

Example: Determining End Behavior for Rational Functions

For each of the following functions, determine the limits as xβ†’βˆžxβ†’βˆž and xβ†’βˆ’βˆžxβ†’βˆ’βˆž. Then, use this information to describe the end behavior of the function.

  1. f(x)=3xβˆ’12x+5f(x)=3xβˆ’12x+5 (Note: The degree of the numerator and the denominator are the same.)
  2. f(x)=3x2+2x4x3βˆ’5x+7f(x)=3x2+2x4x3βˆ’5x+7 (Note: The degree of numerator is less than the degree of the denominator.)
  3. f(x)=3x2+4xx+2f(x)=3x2+4xx+2 (Note: The degree of numerator is greater than the degree of the denominator.)

Watch the following video to see the worked solution to Example: Determining End Behavior for Rational Functions.

Try It

Evaluate limxβ†’Β±βˆž3x2+2xβˆ’15x2βˆ’4x+7limxβ†’Β±βˆž3x2+2xβˆ’15x2βˆ’4x+7 and use these limits to determine the end behavior of f(x)=3x2+2xβˆ’15x2βˆ’4x+7f(x)=3x2+2xβˆ’15x2βˆ’4x+7.

 

 

Before proceeding, consider the graph of f(x)=(3x2+4x)(x+2)f(x)=(3x2+4x)(x+2) shown in Figure 17. As xβ†’βˆžxβ†’βˆž and xβ†’βˆ’βˆžxβ†’βˆ’βˆž, the graph of ff appears almost linear. Although ff is certainly not a linear function, we now investigate why the graph of ff seems to be approaching a linear function. First, using long division of polynomials, we can write

f(x)=3x2+4xx+2=3xβˆ’2+4x+2f(x)=3x2+4xx+2=3xβˆ’2+4x+2

Since 4(x+2)β†’04(x+2)β†’0 as xβ†’Β±βˆžxβ†’Β±βˆž, we conclude that

limxβ†’Β±βˆž(f(x)βˆ’(3xβˆ’2))=limxβ†’Β±βˆž4x+2=0limxβ†’Β±βˆž(f(x)βˆ’(3xβˆ’2))=limxβ†’Β±βˆž4x+2=0

Therefore, the graph of ff approaches the line y=3xβˆ’2y=3xβˆ’2 as xβ†’Β±βˆžxβ†’Β±βˆž. This line is known as an oblique asymptote for ff (Figure 17).

The function f(x) = (3x2 + 4x)/(x + 2) is plotted as is its diagonal asymptote y = 3x – 2.

Figure 17. The graph of the rational function f(x)=(3x2+4x)/(x+2)f(x)=(3x2+4x)/(x+2) approaches the oblique asymptote y=3xβˆ’2y=3xβˆ’2 as xβ†’Β±βˆžxβ†’Β±βˆž.

We can summarize the results of the example above to make the following conclusion regarding end behavior for rational functions. Consider a rational function

f(x)=p(x)q(x)=anxn+anβˆ’1xnβˆ’1+β‹―+a1x+a0bmxm+bmβˆ’1xmβˆ’1+β‹―+b1x+b0f(x)=p(x)q(x)=anxn+anβˆ’1xnβˆ’1+β‹―+a1x+a0bmxm+bmβˆ’1xmβˆ’1+β‹―+b1x+b0,

where an≠0an≠0 and bm≠0bm≠0.

  1. If the degree of the numerator is the same as the degree of the denominator (n=m)(n=m), then ff has a horizontal asymptote of y=an/bmy=an/bm as xβ†’Β±βˆžxβ†’Β±βˆž.
  2. If the degree of the numerator is less than the degree of the denominator [latex](n
  3. If the degree of the numerator is greater than the degree of the denominator (n>m)(n>m), then ff does not have a horizontal asymptote. The limits at infinity are either positive or negative infinity, depending on the signs of the leading terms. In addition, using long division, the function can be rewritten as
    f(x)=p(x)q(x)=g(x)+r(x)q(x)f(x)=p(x)q(x)=g(x)+r(x)q(x),

    where the degree of r(x)r(x) is less than the degree of q(x)q(x). As a result, limxβ†’Β±βˆžr(x)/q(x)=0limxβ†’Β±βˆžr(x)/q(x)=0. Therefore, the values of [f(x)βˆ’g(x)][f(x)βˆ’g(x)] approach zero as xβ†’Β±βˆžxβ†’Β±βˆž. If the degree of p(x)p(x) is exactly one more than the degree of q(x)q(x) (n=m+1)(n=m+1), the function g(x)g(x) is a linear function. In this case, we call g(x)g(x) an oblique asymptote.
    Now let’s consider the end behavior for functions involving a radical.

Example: Determining End Behavior for a Function Involving a Radical

Find the limits as xβ†’βˆžxβ†’βˆž and xβ†’βˆ’βˆžxβ†’βˆ’βˆž for f(x)=3xβˆ’2√4x2+5f(x)=3xβˆ’2√4x2+5 and describe the end behavior of ff.

Try It

Evaluate limxβ†’βˆžβˆš3x2+4x+6limxβ†’βˆžβˆš3x2+4x+6.

Try It

Determining End Behavior for Transcendental Functions

The six basic trigonometric functions are periodic and do not approach a finite limit as xβ†’Β±βˆžxβ†’Β±βˆž. For example, sinxsinx oscillates between 1 and -1 (Figure 19). The tangent function xx has an infinite number of vertical asymptotes as xβ†’Β±βˆžxβ†’Β±βˆž; therefore, it does not approach a finite limit nor does it approach ±∞±∞ as xβ†’Β±βˆžxβ†’Β±βˆž as shown in (Figure 20).

The function f(x) = sin x is graphed.

Figure 19. The function f(x)=sinxf(x)=sinx oscillates between 1 and -1 as xβ†’Β±βˆžxβ†’Β±βˆž

 

The function f(x) = tan x is graphed.

Figure 20. The function f(x)=tanxf(x)=tanx does not approach a limit and does not approach ±∞±∞ as xβ†’Β±βˆžxβ†’Β±βˆž

Recall that for any base b>0,bβ‰ 1b>0,bβ‰ 1, the function y=bxy=bx is an exponential function with domain (βˆ’βˆž,∞)(βˆ’βˆž,∞) and range (0,∞)(0,∞). If b>1,y=bxb>1,y=bx is increasing over (βˆ’βˆž,∞)(βˆ’βˆž,∞). If [latex]01[/latex]. Therefore, f(x)=exf(x)=ex is increasing on (βˆ’βˆž,∞)(βˆ’βˆž,∞) and the range is (0,∞)(0,∞). The exponential function f(x)=exf(x)=ex approaches ∞ as xβ†’βˆž and approaches 0 as xβ†’βˆ’βˆž as shown in (Figure) and (Figure).

End behavior of the natural exponential function
x -5 -2 0 2 5
ex 0.00674 0.135 1 7.389 148.413
The function f(x) = ex is graphed.

Figure 21. The exponential function approaches zero as xβ†’βˆ’βˆž and approaches ∞ as xβ†’βˆž.

Recall that the natural logarithm function f(x)=ln(x) is the inverse of the natural exponential function y=ex. Therefore, the domain of f(x)=ln(x) is (0,∞) and the range is (βˆ’βˆž,∞). The graph of f(x)=ln(x) is the reflection of the graph of y=ex about the line y=x. Therefore, ln(x)β†’βˆ’βˆž as xβ†’0+ and ln(x)β†’βˆž as xβ†’βˆž as shown in (Figure) and (Figure).

End behavior of the natural logarithm function
x 0.01 0.1 1 10 100
ln(x) -4.605 -2.303 0 2.303 4.605
The function f(x) = ln(x) is graphed.

Figure 22. The natural logarithm function approaches ∞ as xβ†’βˆž.

example: Determining End Behavior for a Transcendental Function

Find the limits as xβ†’βˆž and xβ†’βˆ’βˆž for f(x)=(2+3ex)(7βˆ’5ex) and describe the end behavior of f.

Watch the following video to see the worked solution to Example: Determining End Behavior for a Transcendental Function.

Try It

Find the limits as xβ†’βˆž and xβ†’βˆ’βˆž for f(x)=(3exβˆ’4)(5ex+2).