
certainly converge to c. Test each of the following functions
for this condition on the given interval:
a. f ðxÞ5 sin x; c 5 0; I 5 ð2π=4; π=4Þ
b. f ðxÞ5 x
2
2 x ; c 5 0; I 5 ð1=2; 2Þ
c. f ðxÞ5 x
3
1 8; c 522; I 5 ð23; 21Þ
d. f ðxÞ5 3x
2
2 12; c 5 2; I 5 ð1:5; 2:5Þ
16. It is interesting to observe that division can be accom-
plished numerically by an implementation of the Newton-
Raphson Method that does not involve any division at all.
Suppose that we wish to calculate α/β where β is nonzero.
The idea is to approximate 1/β andthenmultiplybyα.For
the approximation of 1/β, apply the Newton-Raphson
method to f (x) 5 β 2 1/x. What is the function Φ(x)of
formula (4.8.2)? (The formula involves multiplication and
subtraction, but not division.) If β 5 3andx
1
5 0.4, what
value does formula (4.8.3) yield for x
3
?
17. A Babylonian approximation to
ffiffiffi
c
p
from the second
millennium B.C. consists of starting with a first estimate
x
1
and then computing the subsequent approximations
that are defined iteratively by
x
j11
5
1
2
x
j
1
c
x
j
:
Show that this ancient algorithm is the Newton-Raphson
Method applied to f (x) 5 x
2
2 c. Of course there was no
notion of derivative when this algorithm was discovered.
Considering that x
j 1 1
is the average of x
j
and c/x
j
, what
might Babylonian mathematicians have had in mind when
they devised this algorithm?
18. Define
gðxÞ¼
ðx 4Þ
1=2
if x $ 4
ð4 xÞ
1=2
if x , 4
(
Prove that, for any initial guess other than x 5 4, the
Newton-Raphson Method will not converge.
19. If f is a function which is always concave up or always
concave down, if f
0
has no root, and if f has a root, then
the Newton-Raphson Method always converges to a root
no matter what the first estimate. Explain why.
c In doing Exercises 20223,
refer to Exercises 84289 of
Section 4.2 for background on discrete dynamical systems.
Assume that f
00
exists and is continuous. b
20. Let Φ(x) 5 x 2 f (x)/f
0
(x). Suppose that f
0
(x
*
) 6¼0. Show
that x
*
is an equilibrium point of the dynamical system
associated to Φ if and only if x
*
is a root of f.
21. Suppose that x
*
is a root of f and that |Φ
0
(x
*
)| , 1 and that
Φ
0
is continuous on an open interval centered at x
*
. Show
that the sequence {x
n
} of Newton-Raphson iterates will
converge to x
*
provided that x
1
is sufficiently close to x
*
.
22. Show that |Φ
0
(x
*
)| , 1 is equivalent to
f ðx
*
Þf
00
ðx
*
Þ
f ðx
*
Þ
2
, 1:
23. Deduce that under the hypotheses of the headnote, the
Newton-Raphson iterates will converge to the root x
*
of f
provided that x
1
is chosen sufficiently close to x
*
.
Calculator/Computer Exercises
24. Plot
f (x) 5 x
5
1 4x
4
1 6x
3
1 4x
2
1 1
in the viewing window [22.3, 0.5] 3 [23.25, 3]. You will
see that f has a root near 2 2.1. Apply the Newton-
Raphson Method using the initial estimate x
1
5 0. What
happens? Now try the initial estimate x
1
521.5 and cal-
culate x
2
through x
10
. Use the graph of f to explain
why 21.5, although not too distant from the root, is a poor
choice for x
1
. (In fact, when x
1
521.5 initializes the
Newton-Raphson process, x
54
is the first approximation
that is closer to the root than the initial estimate.) Now
use the initial estimate x
1
526. What approximation does
x
10
give? (Notice that, because of the geometry of the
graph of f, the number 26, though not as close to the root
as 21.5, is a more efficient choice for the initial estimate.)
c In each of Exercises 25228,
the given function has one real
root. Approximate it by making an initial estimate x
1
and
applying the Newton-Raphson Method until an integer n is found
such that |x
n
2 x
n 2 1
| , 5 3 10
27
. State x
1
, x
2
, x
3
, ...,x
n
. b
25. f (x) 5 x
5
1 x
3
1 2x 2 5
26. f (x) 5 2x 1 cos ( x)
27. f (x) 5 5x
5
2 9x
4
1 15x
3
2 27x
2
1 10x 2 18
28. f (x) 5 x
7
1 3x
4
2 x
2
1 3x 2 2
c In Exercises 29 and 30 find
the largest real root of the
given function f (x). b
29. f (x) 5 x 1 sin
(4.6x)/x
30. f (x) 5 x 1 tan ( x), x2 (2π,π).
31. Use the Newton-Raphson Method to approximate the
positive solution of
ffiffiffiffiffiffiffiffiffiffiffiffiffi
x
3
1 1
p
5
ffiffiffiffiffiffiffiffiffiffiffiffiffi
x
2
1 1
3
p
to 5 decimal
places.
32. The polynomial f (x) 5 x
11
1 x
7
2 1 has one root γ in the
interval [0, 1]. Find this root to five decimal places using
the Newton-Raphson Method.
c If (x 2 c)
2
is a factor of a polynomial p (x)but(x 2 c)
3
is not,
then c is a ro ot of p (x)ofmultiplicity 2. The graph of y 5 p (x)
touches the x-axis at a root of multiplicity 2 but does not cross
the x-axis there. In each of Exercises 33236,
plot the given
polynomial p (x) in the specified viewing rectangle. Identify a
rational number c that is a root of p with multiplicity 2. Use the
Newton-Raphson Method with initial estimate x
1
5 c 1 1/2 to
obtain iterates x
2
, x
3
, ...,x
n
. Terminate the process at the
smallest value of n for which |x
N
2 c| . 5 3 10
24
.WhatisN?
You will notice that the convergence is slow. Record the value
of N so that it can be used for comparison in Exercise 37. b
4.8 The Newton-Raphson Method 355