Skip to main content
Logo image

Active Calculus 2nd Ed

Section 8.1 Extending local linearization

Subsection 8.1.1 Introduction

Early in our study of calculus in Section 1.8, we learned that if a function f has a derivative at a fixed value x=a, when we zoom in on its graph near (a,f(a)), the function looks linear. Indeed, such a function is differentiable, and we know that near a fixed input value a,
f(x)L(x)=f(a)+f(a)(xa),
where L is the tangent line approximation to f at a.
In this section, we use the function f(x)=ex as a case study to investigate how we can use other basic functions to better approximate the value of f(x) near a=0.

Preview Activity 8.1.1.

Consider the function f(x)=ex near a=0. We know that f(x)=ex, so f(0)=1; along with the fact that f(0)=1, it follows that the tangent line approximation is
L(x)=f(0)+f(0)(x0)=1+1(x0)=1+x.
Here is the function f(x)=ex and its tangent line L(x)=1+x near the point (0,f(0)).
(a)
Build a spreadsheet that computes the difference between f(x) and L(x) for x-values between 1 and 1, spaced 0.1 units apart. Note: we will revisit this spreadsheet for additional computations in Activity 8.1.4, so it would be ideal if you save your work electronically in place you can find it later.
Your spreadsheet should start like the following table:
Row Δx x f(x)=ex L(x)=1+x
0 0.1 1.0 0.36787 0.0
1 0.1 0.9 0.40656 0.1
2 0.1 0.8 0.44932 0.2
Then Row 14 is:
Row Δx x f(x)=ex L(x)=1+x
14
(b)
Next, add a column to your spreadsheet that computes |f(x)L(x)|. What is the new entry in Row 14?
How do the differences between f(x) and L(x) change as you move across the interval from 1 to 1?
  • No changes.
  • Differences get smaller and smaller.
  • Differences get larger and larger.
  • Differences get larger the farther away x is from 0.
(c)
For about what values of x is it true that |f(x)L(x)|<0.1?
(d)
Notice that the curvature in f(x)=ex is what makes the linear approximation L(x) lose accuracy. What kind of simple function might do a better job approximating ex than a linear one?

Subsection 8.1.2 Finding a quadratic approximation

In Preview Activity 8.1.1, we found that the error in the tangent line approximation of f(x)=ex at a=0 grows significantly as we consider x-values further and further from 0. This is due to the fact that the tangent line is straight while the function f(x)=ex has some curvature. To hopefully improve the approximation, we are going to try to find a quadratic function whose curvature matches that of f(x)=ex at the point of tangency.
While we have usually used the notation “L(x)” for the tangent line, in what follows we will instead write “T1(x)”, and think of this as “the degree 1 approximation”. In a similar way, we will write “T2(x)” for the quadratic (degree 2) approximation.
Recall that for any function f that has a derivative at a=0, its tangent line approximation at a=0 is
T1(x)=f(0)+f(0)(x0).
Moreover, the functions T1 and f have two exact values in common. First, their function values agree at the point of tangency: T1(0)=f(0). And second, since T1(x) is a linear function whose slope is f(0), it is also true that their derivative values agree at the point of tangency: T1(0)=f(0).
To generate a quadratic function that approximates f near a=0, we choose to have this quadratic function not only share the same function value and derivative value as f at a=0, but also the same second derivative value
 1 
Here we are implicitly assuming that the function f(x) has a second derivative at a=0, which is a property that holds for f(x)=ex.
at a=0 in order to match the concavity or curvature of f. In other words, we are adding a term to the linear approximation that gives the same amount of curvature as the function f.
We can state these requirements more formally as follows.

The quadratic approximation of f(x)=ex.

To extend the linear approximation of f(x)=ex to a quadratic approximation, we seek a function T2(x) of the form
T2(x)=b0+b1x+b2x2
that satisfies
  • T2(0)=f(0), so T2 and f share the same height at a=0;
  • T2(0)=f(0), so T2 and f share the same slope at a=0;
  • T2(0)=f(0), so T2 and f share the same concavity at a=0.
In Activity 8.1.2, we explore how these three requirements determine b0, b1, and b2 in T2(x) for the function f(x)=ex.

Activity 8.1.2.
Let f(x)=ex and T2(x)=b0+b1x+b2x2. We seek numerical values for the constants b0, b1, and b2 so that f(0)=T2(0), f(0)=T2(0), and f(0)=T2(0).
(a)
Note that since b0, b1, and b2 are constants, if we take the derivative of the quadratic function T2 using the sum and constant multiple rules, it follows that T2(x)=b1+2b2x.
What is T2(x)?
(b)
Recall that f(x)=ex. Determine f(x) and f(x).
(c)
Enter the formulas you’ve determined for f(x), f(x), and T2(x) to fill in the blanks below.
f(x)=exT2(x)=b0+b1x+b2x2f(x)=XXXT2(x)=b1+2b2xf(x)=XXXT2(x)=XXX
(d)
Next, observe that since T2(x)=b0+b1x+b2x2, it follows that T2(0)=b0. Reason similarly to determine the values of T2(0) and T2(0), as well as those of f(0), f(0), and f(0) and enter these values appropriately in the blanks below.
f(0)=XXXT2(0)=b0f(0)=XXXT2(0)=XXXf(0)=XXXT2(0)=XXX
(e)
Now, recall that we want the function values, first derivative values, and second derivative values of f and T2 to match at a=0. What does T2(0)=f(0) tell us about the value of b0, and what is its value? What does T2(0)=f(0) imply the value of b1 is? How can we reason similarly to find b2?
(f)
Having now determined the numerical values of b0, b1, and b2, use appropriate computing technology to plot the function T2(x)=b0+b1x+b2x2 along with f(x)=ex and T1(x)=1+x in the same window as that shown in Figure 8.1.3.
Figure 8.1.3. The function f(x)=ex and its tangent line T1(x)=1+x near the point (0,f(0)).
What do you notice? For about which values of x is |f(x)T2(x)|<0.1?

Subsection 8.1.3 Over and over again

A remarkable feature of mathematics is that when a process effectively generates an approximation, doing that same process again (perhaps with some slight modifications) often improves the approximation. In Activity 8.1.2, we found a quadratic approximation of f(x)=ex near the point (0,f(0)) that results in an improvement over the linear approximation of f. It is reasonable to hope that a degree 3 polynomial approximation of f(x)=ex will be even better.
To investigate, we seek a degree 3 polynomial T3(x) of the form
T3(x)=c0+c1x+c2x2+c3x3
that satisfies
  • the same conditions we imposed on T2(x):
    T3(0)=f(0),T3(0)=f(0), and T3(0)=f(0)
    so that T3 and f share the same function value, first derivative value, and second derivative value at a=0,
  • plus the additional condition that
    T3(0)=f(0),
    so T3 and f share the same third derivative value
     2 
    Here we are assuming that the original function f has a third derivative at a=0, which is valid since f(x)=ex.
    at a=0.
We are using the unknown constants c0, c1, c2, and c3 for T3(x) instead of the constants b0, b1, and b2 that we considered for T2(x) since we don’t yet know whether the first three values of ci will be the same as those of bi or not.
Like in our work with T2, we observe that since T3 is a polynomial, its derivatives are straightforward to compute. For instance,
T3(x)=c1+2c2x+3c3x2.
We continue our investigation of this new approximation of f(x)=ex in Activity 8.1.3, where we work to determine the values of c0, c1, c2, and c3 plus explore how well T3(x) approximates f(x) near a=0.

Activity 8.1.3.
Let f(x)=ex and T3(x)=c0+c1x+c2x2+c3x3.
(a)
By computing the third derivative of f(x) and the second and third derivatives of T3(x) and evaluating the relevant functions at x=0, fill in the blanks below.
f(x)=exT3(x)=c0+c1x+c2x2+c3x3f(x)=exT3(x)=c1+2c2x+3c3x2f(x)=exT3(x)=XXXXXf(x)=XXXT3(x)=XXXXXf(0)=1T3(0)=XXXf(0)=1T3(0)=XXXf(0)=1T3(0)=XXXf(0)=XXXT3(0)=XXX
(b)
Next, recall that we want f and T3 to share the same function and derivative values at a=0 up to and including the third derivative. For instance, one of the four needed equations is T3(0)=f(0). Use the four equations your work in the preceding question to determine the values of c0, c1, c2, and c3.
(c)
Having now determined the numerical values of c0, c1, c2, and c3, use appropriate computational technology to plot the function T3(x)=c0+c1x+c2x2+c3x3 along with f(x)=ex, T1(x)=1+x, and T2(x)=1+x+12x2 in the same window as shown in Figure 8.1.7.
Figure 8.1.7. The function f(x)=ex, its tangent line T1(x)=1+x, and the quadratic approximation T2(x)=1+x+12x2 near the point (0,f(0)).
What do you notice? For approximately which values of x is |f(x)T3(x)|<0.1?
(d)
What if we wanted a degree-4 polynomial approximation to f(x)=ex near a=0? Based on the patterns you’ve observed in T1, T2, and T3, conjecture values for the constants d0,,d4 for a function T4 of the form
T4(x)=d0+d1x+d2x2+d3x3+d4x4
that satisfies T4(0)=f(0), T4(0)=f(0), , T4(4)(0)=f(4)(0). Add this function T4 to your plot in part (c) that includes f(x) and the lower-degree polynomial approximations. What do you notice?

Subsection 8.1.4 As the degree of the approximation increases

Our work so far with the case study function f(x)=ex suggests that as we find degree n polynomial approximations, Tn, that satisfy
Tn(0)=f(0),Tn(0)=f(0),Tn(0)=fn(0),,Tn(n)(0)=f(n)(0),
increasing the value of n improves the accuracy of the approximation.
In the next activity, we introduce the idea of the error of a polynomial approximation and investigate explicitly how the error varies for approximations of f(x)=ex as we vary n and vary x.

Activity 8.1.4.
We continue to work with f(x)=ex and the four approximations of degree 1, 2, 3, and 4 given by T1(x), T2(x), T3(x), and T4(x).
(a)
In Preview Activity 8.1.1, we built a spreadsheet that computed the differences between f(x) and T1(x) for x-values between 1 and 1, spaced 0.1 units apart. Your spreadsheet started like the one shown in the table in Preview Activity 8.1.1.
Next, we build an updated version of this spreadsheet that computes similar differences between f and the three higher degree approximations we have found. In particular, we now want to have columns for Δx, x, f(x), T1(x), T2(x), T3(x), and T4(x), plus the absolute differences |f(x)T1(x)|, |f(x)T2(x)|, |f(x)T3(x)|, and |f(x)T4(x)|. Hint: when building your entries, note that you can think of T2(x) as T2(x)=T1(x)+12x2, and similarly view T3(x) as “T2(x) plus one more term”.
Include at least 5 digits of accuracy beyond the decimal. The first seven columns of your spreadsheet might start like this:
Table 8.1.9. Comparing f(x)=ex and its degree 1, 2, 3, 4 and approximations near a=0.
Δx x f(x) T1(x) T2(x) T3(x) T4(x)
0.1 1.0 0.36787 0.00000 0.50000 0.33333 0.37500
0.1 0.9 0.40657 0.10000 0.50500 0.38350 0.41083
The next four columns of your spreadsheet should begin as follows:
Table 8.1.10. The absolute error between f(x)=ex and its degree 1, 2, 3, and 4 approximations at x=1 and x=0.9.
|f(x)T1(x)| |f(x)T2(x)| |f(x)T3(x)| |f(x)T4(x)|
0.36787 0.13212 0.03454 0.00712
0.30657 0.09843 0.02307 0.00426
(b)
We call the value of |f(x)T2(x)| the absolute error of the quadratic approximation of f at the value x. What is the absolute error of the quadratic approximation at x=1? at x=1?
(c)
What is the absolute error of the cubic (degree 3) approximation, T3(x), at x=1? at x=1?
(d)
What is the absolute error of the quartic (degree 4) approximation, T4(x), at x=1? at x=1?
(e)
Study your spreadsheet for trends that you notice as the value of x changes or the degree n of the approximation changes. What are your observations?
(f)
Investigate the errors in the various approximations for a wider interval of x-values. For example, you might consider starting at x=2 with Δx=0.2. What do you notice?
One important application of our work so far is that these polynomial approximations provide a way to approximate values of the function f(x)=ex. For example, since we’ve shown that
ex1+x+12x2+16x3+124x4,
it follows that
e121+12+12(12)2+16(12)3+124(12)4=211128=1.6484375.
In fact, this approach through polynomial approximation is one way that computers determine the value of e12, which is approximately 1.64872127, to whatever accuracy is needed: by using even better polynomial approximations than the degree-4 one that we found, computers are able to generate the approximate value 1.64872127 simply by the basic computations of addition and multiplication with enough terms.
Throughout this section, we have focused on f(x)=ex. One of the characteristics that makes f(x)=ex special is the fact that its derivative is itself; indeed, the nth derivative of f is f(n)(x)=ex for every natural number n, which in turn implies that f(n)(0)=1 for every value of n. This will ultimately help to find patterns in the coefficients of the degree n polynomial approximation, Tn(x), and be able to easily write down a formula for any value of n.
It is natural to think that we can find similar approximations of other functions, especially ones such as sin(x) and cos(x) that also exhibit repeating patterns in their derivatives. In Section 8.2, we will develop a general approach to finding the coefficient of xn in the degree n approximation of any function with n derivatives and learn how to find a general expression for the degree n approximation.

Subsection 8.1.5 Summary

  • For the function f(x)=ex, which bends considerably as we move away from a=0 (especially for x>0), the tangent line, T1(x), is not a very good approximation for x-values that satisfy |x|>0.5. For example, |e0.5T1(0.5)|0.148721, so the linear approximation has an absolute error of more than 0.1 at x=0.5.
  • Using the strategy of finding a higher degree polynomial whose function and derivative values match at the selected point of tangency, we are able to find higher degree polynomials that much more effectively approximate f(x)=ex near a=0 than the approximation generated by the tangent line. For example, using the degree 3 approximation T3(x)=1+x+12x2+16x3, we see that |f(x)T3(x)|<0.01 for all x that satisfy |x|<0.6.
  • It appears that the degree of the polynomial impacts the accuracy of the approximation of f(x)=ex in at least two ways: if we fix an x-value, the higher the degree of the polynomial, the more accurate the approximation. In addition, raising the degree of the polynomial approximation appears to widen the interval on which the approximation is effective.

Exercises 8.1.6 Exercises

1.

Let f(x)=cos(x) and let T2(x)=b0+b1x+b2x2. We find the values of b0,b1 and b2 that make the functions cos(x) and T2(x) have the same height, slope, and concavity at x=0.
(a)
Complete the table below by computing derivatives of T2(x) and cos(x) and evaluating them at x=0 as directed.
f(x)= cos(x) T2(x)= b0+b1x+b2x2
f(x)= T2(x)=
f(x)= T2(x)=
f(0)= T2(0)= b0
f(0)= T2(0)=
f(0)= T2(0)=
(b)
Now, since we want T2(x) and cos(x) to have the same height at x=0, set T2(0)=f(0).
This implies b0=.
Since we want T2(x) and cos(x) to have the same slope at x=0, set T2(0)=f(0).
This implies b1=.
Finally, since we want T2(x) and cos(x) to have the same concavity at x=0, set T2(0)=f(0).
This implies b2=.
Putting everything together, what is the resulting formula for T2(x)=b0+b1x+b2x2?
T2(x)=

2.

Let f(x)=ln(1+x) and let T2(x)=b0+b1x+b2x2. We find the values of b0,b1 and b2 that make the functions ln(1+x) and T2(x) have the same height, slope, and concavity at x=0.
(a)
Complete the table below by computing derivatives of T2(x) and ln(1+x) and evaluating them at x=0 as directed.
f(x)= ln(1+x) T2(x)= b0+b1x+b2x2
f(x)= T2(x)=
f(x)= T2(x)=
f(0)= T2(0)= b0
f(0)= T2(0)=
f(0)= T2(0)=
(b)
Now, since we want T2(x) and ln(1+x) to have the same height at x=0, set T2(0)=f(0).
This implies b0=.
Since we want T2(x) and ln(1+x) to have the same slope at x=0, set T2(0)=f(0).
This implies b1=.
Finally, since we want T2(x) and ln(1+x) to have the same concavity at x=0, set T2(0)=f(0).
This implies b2=.
Putting everything together, what is the resulting formula for T2(x)=b0+b1x+b2x2?
T2(x)=

3.

Let f(x)=e2x and let T3(x)=c0+c1x+c2x2+c3x3. We find the values of c0,,c3 that make the functions e2x and T3(x) have the same height, slope, concavity, and third derivative value at x=0.
(a)
Complete the table below by computing derivatives of T3(x) and e2x and evaluating them at x=0 as directed.
f(x)= e2x T3(x)= c0+c1x+c2x2+c3x3
f(x)= T3(x)=
f(x)= T3(x)=
f(x)= T3(x)=
f(0)= T3(0)= c0
f(0)= T3(0)=
f(0)= T3(0)=
f(0)= T3(0)=
(b)
Now, since we want T3(x) and e2x to have the same height at x=0, set T3(0)=f(0).
This implies c0=.
Since we want T3(x) and e2x to have the same slope at x=0, set T3(0)=f(0).
This implies c1=.
Since we want T3(x) and e2x to have the same concavity at x=0, set T3(0)=f(0).
This implies c2=.
Finally, since we want T3(x) and e2x to have the same third derivative value at x=0, set T3(0)=f(0).
This implies c3=.
Putting everything together, what is the resulting formula for T3(x)=c0+c1x+c2x2+c3x3?
T3(x)=

4.

Let f(x)=3x2+x+5 and let T2(x)=b0+b1x+b2x2. We find the values of b0,b1 and b2 that make the functions f(x) and T2(x) have the same height, slope, and concavity at x=0.
(a)
Complete the table below by computing derivatives of T2(x) and 3x2+x+5 and evaluating them at x=0 as directed.
f(x)= 3x2+x+5 T2(x)= b0+b1x+b2x2
f(x)= T2(x)=
f(x)= T2(x)=
f(0)= T2(0)= b0
f(0)= T2(0)=
f(0)= T2(0)=
(b)
Now, since we want T2(x) and 3x2+x+5 to have the same height at x=0, set T2(0)=f(0).
This implies b0=.
Since we want T2(x) and 3x2+x+5 to have the same slope at x=0, set T2(0)=f(0).
This implies b1=.
Finally, since we want T2(x) and 3x2+x+5 to have the same concavity at x=0, set T2(0)=f(0).
This implies b2=.
Putting everything together, what is the resulting formula for T2(x)=b0+b1x+b2x2?
T2(x)=
What does your answer mean about the quadratic approximation to a quadratic function?

5.

In the following problem, note that “second degree Taylor polynomial” is the same as the quadratic approximation we’ve been studying, and we’ll see this terminology again in the following section.
Suppose that T2(x)=c0+c1x+c2x2 is the second degree Taylor polynomial for the function f about x=0. What can you say about the signs of c0, c1, c2 if f has the graph given below?
(For each, enter + if the term is positive, and if it is negative. Note that because this is essentially multiple choice problem it will not show which parts of your answer are correct or incorrect.)
c0 is
c1 is
c2 is

6.

Throughout our work in Section 8.1, we have focused on approximating the function f(x)=ex. In this exercise, we change the function of interest to f(x)=13x3+14x22x1, and consider the linear and quadratic approximations to f near a=0.
  1. Determine f(x) and f(x) and enter their formulas below.
    f(x)=13x3+14x22x1f(x)=XXXXXXXXXXf(x)=XXXXXXXXXX
  2. Next, compute f(0), f(0), and f(0) and enter those values below.
    f(0)=XXXXXf(0)=XXXXXf(0)=XXXXX
  3. Use your work so far to determine the formula for T1(x), the tangent line approximation to f(x) at a=0 (which satisfies T1(0)=f(0) and T1(0)=f(0)).
  4. Let T2(x) be the quadratic approximation to f(x) near a=0 that satisfies T2(0)=f(0), T2(0)=f(0), and T2(0)=f(0). You might start by letting T2(x)=c0+c1x+c2x2, and creating an updated table like the one shown below.
    f(x)=x3/3+x2/42x1f(x)=XXXXXXXXXXf(x)=XXXXXXXXXXf(0)=XXXXXf(0)=XXXXXf(0)=XXXXX
    T2(x)=c0+c1x+c2x2T2(x)=c1+2c2xT2(x)=XXXXXXXXXXT2(0)=XXXXXT2(0)=XXXXXT2(0)=XXXXX
    Use your work in in the table above to find the formula for T2(x) that satisfies T2(0)=f(0), T2(0)=f(0), and T2(0)=f(0).
  5. Plot f(x), T1(x), and T2(x) on the same axes, centered at a=0. What do you notice?

7.

In this exercise, we extend our work in Exercise 8.1.6.6. We continue to consider the function f(x)=13x3+14x22x1, but now build the cubic (degree 3) approximation to f near a=0.
  1. Let T3(x)=k0+k1x+k2x2+k3x3 and determine T3(x), T3(x), and T3(x). In addition, determine f(x). Record your results below, along with the values of each of these functions at a=0.
    f(x)=x3/3+x2/42x1f(x)=XXXXXXXXXXf(x)=XXXXXXXXXXf(x)=XXXXXXXXXXf(0)=XXXXXf(0)=XXXXXf(0)=XXXXXf(0)=XXXXX
    T3(x)=k0+k1x+k2x2+k3x3T3(x)=XXXXXXXXXXT3(x)=XXXXXXXXXXT3(x)=XXXXXXXXXXT3(0)=XXXXXT3(0)=XXXXXT3(0)=XXXXXT3(0)=XXXXX
  2. Use your work in above to find the formula for T3(x) that satisfies T3(0)=f(0), T3(0)=f(0), T3(0)=f(0), and T3(0)=f(0).
  3. Plot f(x), T1(x), T2(x), and T3(x) on the same axes, centered at a=0. What do you notice?
  4. What do you expect the degree 4 approximation to f(x), T4(x) to be? Why?
  5. What do you expect will happen if we find the degree 6 approximation to any function f(x) that is itself a degree 6 polynomial?
You have attempted 1 of 7 activities on this page.