Existence and Uniqueness of a Fixed Point

Myat Tha Zin Hlaing

Lecturer
Department of Mathematics
Univeristy of Mandalay
09 5400201
Abstract: It is collected some definitions and results. It is shown that if a function
from an interval into itself is continuous and its derivative exists except endpoints of
the given interval with its absolute value less than one, then it has exactly a fixed
point in a given interval. It is illustrated that there exists which a function has exactly
a fixed point unless it satisfy the assumption. To find this number it is used by fixed-
point iteration method. Finally, it is shown that there exists a sequence converges to a
unique fixed point.
Keywords: roots, fixed-point, existence, uniqueness, convergence
1. Introduction
Luitzen Egbertus Jan Brouwer (1881-1966), who, worked in topology, set
theory, measure theory and complex analysis. He was regarded as one of the greatest
mathematician of the 20th century. He is known as the founder of modern topology
and his famous theorem is fixed-point theorem: if a function g from a closed interval
into itself is continuous then it has at least one fixed point.
In this paper, first we collect some basic definitions such as continuous
function, convergence of sequence of real numbers and differentiable function. Also,
recall that some results such as Intermediate Value Theorem and Mean Value
Theorem. A fixed point for a function is a number which the image of a number
under the given function does not change. Adding the assumption of Brouwer's
theorem, derivative of the function g exist except endpoints of a given interval with
absolute value of it is less than one, it has exactly one fixed point (Theorem 3.2). But,
unless satisfying this assumption, a function has a fixed point (Example 3.5). It is not
easy to find this fixed point. It is used to find approximating method, called, Fixed-
Point Iteration. Finally, it is proved that for any number p 0 in a given interval, there is
a sequence {p n }n 0 in a given interval converges to a unique fixed point p, in in that
interval. To draw graphs it is used Maple 2020 software and online latex equation
editor, LateX4technics.
2. Basic Facts
In this section we collect some basic definitions and results to use them in next
section.

2.1 Definition Let f be a function on an interval [a, b] of real numbers and x 0  [a, b].
The function f is said to be continuous at x 0 if
lim f (x)  f (x 0 ).
x x 0

The function f is continuous on [a, b] if it is continuous on every number in [a, b] .

The set of all continuous functions on [a, b] is denoted by C[a, b].
2.2 Example The function f is defined by f (x)  x is continuous on , the set of all
real numbers but g is defined by
 2

 1 if x  0,
g(x)  
1 if x  0,
is not continuous at x  0. (See Figure 2.1.)

(a) (b)
Figure 2.1. (a) Graph of f(x) = x, (b) Graph of function g in Example 2.2

2.3 Definition Let {x n }n 1 be an infinite sequence of real numbers. The sequence is
said to be converge to x if, for any   0 there exists a positive number N such that if
n  N  | x n  x | .
In this case we write lim x n  x and x is called the limit of the sequence. The
n 
sequence is said to be diverge if it does not converge.
2.4 Example The {x n }n 1 where x n  1/ n converges to 0 but the sequence {y n }n 1
where y n  ( 1) n diverges. (See Figure 2.2.)

(a) (b)

Figure 2.2. (a) Graph of sequemce {x n }n 1 , (b) Graph of sequence {y n }n 1

2.5 Theorem If f is a function defined on [a, b] and x 0  [a, b] then the following
statements are equivalent:
(a) f is continuous at x 0 ;
(b) If {x n }n 1 is any sequence in [a, b] converging to x 0 then
limn  f (x n )  f (x 0 ).
 3

2.6 Definition Let f be a function defined in an open interval containing x 0 . The

function f is said to be differentiable at x 0 if
f (x)  f (x 0 )
f (x 0 )  lim
x x 0 x  x0
exists. The number f (x 0 ) is called the derivative of f at x 0 . A function that has
derivative at each number in an interval [a, b] is differentiable on [a, b] .
2.7 Theorem (Mean Value Theorem) If f  C[a, b] and f is differentiable on (a, b) ,
then a number c in (a, b) exists with
f (b)  f (a)
f (c)  .
ba
(See Figure 2.3.)

Figure 2.3. Mean Value Theorem for y  f (x)

2.8 Theorem ( Intermediate Value Theorem) If f  C[a, b] and L is any number
between f (a) and f (b) , then there exists a number in (a, b) such f (c)  L.
(See Figure 2.4.)

Figure 2.4. Intermediate Value Theorem for y  f (x)

3. Finding Fixed Point
A fixed point for a function is a number at which the value of a function does not
change when the function is applied. We find fixed points for some functions and
show that there exists a unique fixed-point for a continuous function from a closed
interval to itself with its derivative exists for each point in this interval except
endpoints.
 4

3.1 Definition A number p is a fixed point for a given function g if g(p)  p.

For given a root-finding problem f (p)  0, we can define functions g with a fixed
point at p in a number of ways, for example, as follows:
g(x)  x  f (x) or g(x)  x  4f (x) .
Conversely, if the function g has a fixed point at p, then the function f defined by
f (x)  x  g(x)
has a zero at p. When a function has fixed points, how to find them as follows.
3.2 Example We determine any fixed points of the function g(x)  x 2  2. Let p be a
fixed point for g. Then
p  g(p)  p2  2; p2  p  2  0; (p  2)(p  1)  0.
The fixed points for g are p  1 and p  2 .
Geometrically, a fixed point for g occurs precisely when the graph of y  g(x)
intersects the graph y  x. So, g has two fixed points at p  1 and p  2 . These are
shown in Figure 3.1.

Figure 3.1. Fixed points for g(x)  x 2  2

The following theorem gives sufficient conditions for the existence and
uniqueness of a fixed point.
3.3 Theorem (a) If g  C[a, b] and g(x) [a, b] for all x  [a, b] , then g has a least one
fixed point in [a, b].
(b) If, in addition, g(x) exists on (a, b) and a positive constant k  1 exists with
| g(x) | k for all x  (a, b), then there is exactly one fixed point in [a, b]. (See Figure
3.2.)

Figure 3.2. Fixed point for y  g(x)

Proof. (a) If g(a)  a or g(b)  b, then g has a fixed point at an endpoint. If not,
g(a)  a and g(b)  b. The function h(x)  g(x)  x is continuous on [a, b] , with
 5

h(a)  g(a)  a  0 and h(b)  g(b)  b  0.

The Intermediate Value Theorem implies that there exists p  (a, b) for which
h(p)  0. This number p is a fixed point for g because
0  h(p)  g(p)  p  g(p)  p.
(b) Suppose, in addition, that | g(x) | k  1 and that p and q both fixed points in
[a, b]. If p  q, the Mean Value Theorem implies that a number  exists between p and
q such that
g(p)  g(q)
 g().
pq
Thus
| p  q || g(p)  g(q) | g() | p  q | k | p  q || p  q |

which is impossible. Hence, p  q and the fixed point in [a, b] is unique. 

3.4 Example We show that the function g(x)  (x 2 1) / 3 has a unique fixed point on
the interval [1,1]. The maximum and minimum values of g(x) for x in [1,1] must
occur either when x is an endpoint of the interval or when the derivative is 0. Since
g(x)  2x / 3 , the function g is continuous and g(x) exists on [1,1]. The maximum
and minimum values of g(x) occur at x  1, x  0 or x  1. But g(1)  0, g(1)  0
and g(0)  1/ 3. So, an absolute maximum for g(x) on [1,1] at x  1 and x  1, and
an absolute minimum at x  1.
Moreover,
| g(x) |  | 2x 3|  2 3
for all x  (1,1). So, by Theorem 3.2, g has a unique fixed point in [1,1].
Now we find the unique fixed point p in the interval [1,1] can be determined
algebraically. If p  g(p)  (p2  1) / 3 then

3  13
p .
2
Since ((3  13) / 2,(3  13) / 2) is outside of the interval [1,1] , it is impossible.
Thus, ((3  13) / 2,(3  13) / 2). (See Figure 3.3(a).)

Also, g has a unique fixed point p  ((3  13) / 2,(3  13) / 2) for the interval
[3, 4]. In fact, we have g(4)  5 and g(4)  8 / 3  1. So, g does not satisfy the
hypothesis of Theorem 3.3 on [3, 4]. This demonstrates the hypotheses of Theorem
3.3 are sufficient to guarantee a unique fixed point but are not necessary. (See Figure
3.3(b).)
 6

(a) (b)
Figure 3.3(a) Fixed point of the function g on [1,1] and [3, 4]
The following example shows that Theorem 3.3 does not guarantee that a
function has a unique fixed point.
3.5 Example We consider the function g(x)  3 x on [0,1]. Since g(x)  3 x ln 3  0
on [0,1], the function g is strictly decreasing on [0,1]. So,
g(1)  1/ 3  g(x)  1  g(0), 0  x  1. Thus, for x  [0,1], we have g(x) [0,1].
Hence by Theorem 3.3(a), there is at least one fixed point in [0,1]. However,
g(0)   ln 3  1.09612289,
so | g(x) |/ 1 on (0,1). Thus, Theorem 3.3 cannot be used to determine uniqueness.
But g is always decreasing, and it must have the unique fixed point.(See Figure 3.4.)

Figure 3.4. Fixed point for the function g on [0,1]

3.6 Fixed-ThisPoint Iteration
We cannot explicity determine the fixed point in Example 3.5 because we have
now way to solve p  g(p)  3p. We can however determine approximations to this
fixed point to any specified degree of accuracy.
To approximate the fixed point of a function g, we choose an initial
approximation p 0 and generate the sequence {p n }n 0 by letting p n  g(p n 1 ) for each
n  1. If the sequence converges to p and g is continuous, then
p  lim pn  lim g(pn 1 )  g( lim pn 1 )  g(p)
n  n  n 
and a solution to x  g(x) is obtained. This technique is called fixed-point of
functional iteration. The procedure is illustrated in Figure 3.5.
 7

Figure 3.5. Illustration of fixed-point iteration

The following example illustrates to find fixed point for given function.
3.7 Example The equation of x3  4x 2 10  0 has a unique root in [1, 2]. (See Figure
3.6.) There are many ways to change the equation to the fixed-point form x  g(x)
using simple algebraic manipulation. For example, we can manipulate the equation
x3  4x 2 10  0 as follows:
1
4x 2  10  x3 ; x   (10  x3 )1/2 .
2

Figure 3.6. Graph of y  x3  4x 2  10

To obtain a positive solution, g 3 (x) is chosen. We should verify that the fixed point of
each is actually a solution to the origin equation, x3  4x 2 10  0 .
(a) x  g1 (x)  x  x 3  4x 2  10 (b) x  g 2 (x)  (10x  4x)1/2
3
 4x 2 10
(c) x  g3 (x)  12 (10  x 3 )1/2 (d) x  g 4 (x)  ( 410
x
)1/2 (e) x  g5 (x)  x  x
3x 2 8x

With p0  1.5, Table 3.1 lists the results of the fixed-point iteration for all five
choices of g.
Table 3.1
n (a) (b) (c) (d) (e)
0 1.5. 1.5 1.5 1.5 1.5
 8

1 −0.875 0.8165 1.286953768 1.348399725 1.373333333

2 6.732 2.9969 1.402540804 1.367376372 1.365262015

3 −469.7 (−8.65)1/2 1.345458374 1.364957015 1.365230014

1.03 10
8
4 1.375170253 1.365264748 1.365230013

5 1.360094193 1.365225594

6 1.367846968 1.365230576

7 1.363887004 1.365229942

8 1.365916734 1.365230022

9 1.364878217 1.365230012

10 1.365410062 1.365230014

15 1.365223680 1.365230013

20 1.365230236

25 1.365230006

30 1.365230013

The actual root is 1.365230013. It is interesting to note that choice (a) was divergent and
that (b) become undefined.
3.8 Theorem (Fixed-Point Theorem) Let g  C[a, b] be such that g(x) [a, b] for all
x in [a, b] . Suppose, in addition, that g exists on (a, b) and that a constant 0  k  1
exists with
| g(x) |  k
for all x  (a, b). Then for any number p 0 in [a, b] , the sequence defined by
pn  g(pn 1 ), n  1 ,
converges to the unique fixed point p in [a, b] .
Proof. By Theorem 3.3, there exists a unique fixed point p  [a, b] . Since g maps
[a, b] into itself, the sequence {p n }n 0 is defined for all n  0 and p n  [a, b] for all n.
By using the fact that | g(x) |  k and the Mean Value Theorem, we have for each n,
| pn  p |  | g(p n 1 )  g(p) |  | g(n ) | | p n 1  p |  k | p n 1  p | ,

where n  (a, b). Continuing in this way,

| p n  p |  k | p n 1  p |  k 2 | p n 2  p |   k n | p0  p |.
Since 0  k  1, we have
lim n  k n  0 and lim | p n  p |  lim k n | p0  p | 0.
n  n 
That is, for every   0 there is a positive integer N such that if
n  N  | pn  p |  | | pn  p |  0 |  .
Hence {p n }n 0 converges to p. 
3.9 Corollary If g satisfies the hypotheses of Theorem 3.8, then bounds for the error
involved in using pn to approximate p is given by
 9

| p n  p |  k n max{p0  a, b  p0 }
and
n
| pn  p |  1kk | p1  p0 }, for all n  1.
Proof. Because p  [a, b],
| pn  p |  k n | p0  p |  k n max{p0  a, b  p0 }.
For n  1,
| pn 1  pn |  | g(p n )  g(p n 1 ) |  k | p n  p n 1 |   k n | p1  p 0 | .
Thus, for m  n  1,
| pm  pn |  | pm  pm1  pm1   pn 1  pn |
 | pm  pm1 |  | pm1  pm2 |   | pn 1  pn |
 k m1 | p1  p0 |  k m2 | p1  p0 |   k n | p1  p0 |
 k n | p1  p0 | (1  k  k 2   k mn 1 ).
By Theorem 3.3, limm pm  p. Since we write
| pm  pn |  | pm  pn  p n  p |  | p m  p n  (p  p n ) |,
| pm  pn |  lim | p m  p n |
m
m  n 1
 lim k n | p1  p 0 |
m 
 ki
i 0

 k n | p1  p 0 |  k i
i 0

 k | p1  p0 | 11k .
n


3.10 Example We reconsider the various fixed-point schemes described in the
preceding Example 3.7 by using Theorem 3.8 and Corollary 3.9.
(a) For g1 (x)  x  x 3  4x 2  10, we have g1 (1)  6 and g1 (2)  12 . Thus g1 does not
map [1, 2] into itself. Moreover, g1 (x)  1  3x 2  8x, | g1 (x) |  1 for all x in [1, 2] . By
Theorem 3.8, the sequence does not converge.
(b) For g 2 (x)  (10x  4x)1/2 , we can see that g 2 does not map [1, 2] into itself. The
sequence {p n }n 0 is not defined in [1, 2] when p0  1.5 . By Theorem 3.8, the sequence
does not converge.
(c) For g3 (x)  12 (10  x 3 )1/2 , we have

g3 (x)   34 x 2 (10  x 3 )1/2  0

on [1, 2] . So, g 3 is strictly decreasing on [1, 2] . However, | g 3 (2) |  2.12, so the

condition | g 3 (x) |  k fails on [1, 2] . Since p0  1.5 , it suffices to consider the interval
[1.5, 2] instead of [1, 2] . On this interval | g 3 (x) |  0 and so g 3 is strictly decreasing.
Moreover,
 10

1  1.29  g3 (1.5)  g3 (x)  g3 (1)  1.5

for all x  [1.5, 2]. This shows that g 3 maps [1.5, 2] into itself. Also,
| g 3 (x) |  | g 3 (1.5) |  0.66 , on this interval. Hence, by Theorem 3.8, the sequence
converges.
(d) For g 4 (x)  ( 410
x
)1/2 , we have

| g 4 (x) | 5
| 5
 0.15,
10 (4 x)3/2 10 (5)3/2

for all x  [1, 2]. The bound on the magnitude of g 4 (x) is much smaller than the bound
on the magnitude of g3 (x) which explains the more rapid convergence using g 4 .
(e) The sequence defined by
3
 4x 2 10
g5 (x)  x  x
3x 2 8x
converges much more rapidly than our other choices.
4. Conclusion
We are familiar with solving an equation of one variable. In other words, it is
to find the roots of the given equation. To find a fixed point of a given function, g(x)
=x, we rewrite g  x   x  0 . Thus, finding a fixed point is to solve the rewriting
equation. Adding to the assumption of Brouwer's theorem, a given function has
derivative at each point in a given interval except endpoints of that interval, it has
exactly, fixed point. But it is found that a function has exactly fixed point unless it
satisfy assumption previous statement. But it is not easy to calculate it. So, it is used
to find Fixed-Point Iteration method. From this procedure, for any number in a given,
we construct a sequence in that interval converges to a fixed point. So, in this paper, it
gives a numerical approach to understand abstract subject, real analysis.
Acknowledgements
This paper has benefited from all of our teachers and colleagues who gave
generously of their time and expertise. My special thanks to parents and all brothers
and sisters who give patiently support throughout my life.
References
[1] R. L. Burden and J. D. Faires, " Numerical Analysis", Ninth Edition, Brooks/Cole, Cenguge
Learning, Boston, USA, 2011.
[2] W. A.J. Kosmala, "A Friendly Introduction to Analysis Simple and Multivariable",
Second Edition., Upper Sanddle River, Pearson Education Inc., 2004.
[3] W. Rundin, "Principles of Mathematical Analysis", Third Edition, McFraw-Hill Inc., New York,
1976.
[4] http://www.bu.edu>2013/08>tikzpgfmanual," TikZ and PGF"
[5] http://www. latex4technics.com,"LateX4technics"