Plovers Solution Series

Volume 1:
Rudin, Principles of Mathematical Analysis
SYLee
Meng-Gen Tsai
2
i
Preface
This solution is written for one special book named Principles of Mathemat-
ical Analysis. It is written by Rudin. The purpose I wrote this solution is
just for fun; surely, you will nd a lot of errors containing in my solution :pp
Finally, I dedicate this solution to SYLee, wwli, ljl, and many other friends.
Meng-Gen, Tsai.
ii
Contents
Preface i
1 The Real and Complex Number Systems 3
2 Basic Topology 9
3 Numerical Sequences and Series 31
4 Continuity 35
5 Dierentiation 41
6 Sequences and Series of Functions 67
1
2 CONTENTS
Chapter 1
The Real and Complex
Number Systems
3
4 CHAPTER 1. THE REAL AND COMPLEX NUMBER SYSTEMS
Written by Men-Gen Tsai
email: b89902089@ntu.edu.tw
1.
2.
3.
4.
5.
6. Fix b > 1.
(a) If m, n, p, q are integers, n > 0, q > 0, and r = m/n = p/q, prove
that
(b
m
)
1/n
= (b
p
)
1/q
.
Hence it makes sense to dene b
r
= (b
m
)
1/n
.
(b) Prove that b
r+s
= b
r
b
s
if r and s are rational.
(c) If x is real, dene B(x) to be the set of all numbers b
t
, where t is
rational and t x. Prove that
b
r
= sup B(r)
where r is rational. Hence it makes sense to dene
b
x
= sup B(x)
for every real x.
(d) Prove that b
x+y
= b
x
b
y
for all real x and y.
Proof: For (a): mq = np since m/n = p/q. Thus b
mq
= b
np
.
By Theorem 1.21 we know that (b
mq
)
1/(mn)
= (b
np
)
1/(mn)
, that is,
(b
m
)
1/n
= (b
p
)
1/q
, that is, b
r
is well-dened.
For (b): Let r = m/n and s = p/q where m, n, p, q are integers, and
n > 0, q > 0. Hence (b
r+s
)
nq
= (b
m/n+p/q
)
nq
= (b
(mq+np)/(nq)
)
nq
=
b
mq+np
= b
mq
b
np
= (b
m/n
)
nq
(b
p/q
)
nq
= (b
m/n
b
p/q
)
nq
. By Theorem 1.21
5
we know that ((b
r+s
)
nq
)
1/(nq)
= ((b
m/n
b
p/q
)
nq
)
1/(nq)
, that is b
r+s
=
b
m/n
b
p/q
= b
r
b
s
.
For (c): Note that b
r
B(r). For all b
t
B(r) where t is rational and
t r. Hence, b
r
= b
t
b
rt
b
t
1
rt
since b > 1 and r t 0. Hence b
r
is an upper bound of B(r). Hence b
r
= sup B(r).
For (d): b
x
b
y
= sup B(x) sup B(y) b
t
x
b
t
y
= b
t
x
+t
y
for all rational
t
x
x and t
y
y. Note that t
x
+ t
y
x + y and t
x
+ t
y
is rational.
Therefore, sup B(x) sup B(y) is a upper bound of B(x + y), that is,
b
x
b
y
sup B(x + y) = b
(
x + y).
Conversely, we claim that b
x
b
r
= b
x+r
if x R
1
and r Q. The
following is my proof.
b
x+r
= sup B(x + r) = sup{b
s
: s x + r, s Q}
= sup{b
sr
b
r
: s r x, s r Q}
= b
r
sup{b
sr
: s r x, s r Q}
= b
r
sup B(x)
= b
r
b
x
.
And we also claim that b
x+y
b
x
if y 0. The following is my proof:
(r Q)
B(x) = {b
r
: r x} {b
r
: r x + y} = B(x + y),
Therefore, sup B(x + y) sup B(x), that is, b
x+y
b
x
.
Hence,
b
x+y
= sup B(x + y)
= sup{b
r
: r x + y, r Q}
= sup{b
s
b
rs
: r x + y, s x, r Q, s Q}
sup{sup B(x)b
rs
: r x + y, s x, r Q, s Q}
= sup B(x) sup{b
rs
: r x + y, s x, r Q, s Q}
= sup B(x) sup{b
rs
: r s x + y s, s x, r s Q}
6 CHAPTER 1. THE REAL AND COMPLEX NUMBER SYSTEMS
= sup B(x) sup B(x + y s)
sup B(x) sup B(y)
= b
x
b
y
Therefore, b
x+y
= b
x
b
y
.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
7
26.
27.
28.
29.
30.
8 CHAPTER 1. THE REAL AND COMPLEX NUMBER SYSTEMS
Chapter 2
Basic Topology
9
10 CHAPTER 2. BASIC TOPOLOGY
Written by Men-Gen Tsai
email: b89902089@ntu.edu.tw
1. Prove that the empty set is a subset of every set.
Proof: For any element x of the empty set, x is also an element of
every set since x does not exist. Hence, the empty set is a subset of
every set.
2. A complex number z is said to be algebraic if there are integers a
0
, ..., a
n
,
not all zero, such that
a
0
z
n
+ a
1
z
n1
+ ... + a
n1
z + a
n
= 0.
Prove that the set of all algebraic numbers is countable. Hint: For
every positive integer N there are only nitely many equations with
n +|a
0
| +|a
1
| + ... +|a
n
| = N.
Proof: For every positive integer N there are only nitely many equa-
tions with
n +|a
0
| +|a
1
| + ... +|a
n
| = N.
(since 1 n N and 0 |a
0
| N). We collect those equations as
C
N
. Hence

C
N
is countable. For each algebraic number, we can form
an equation and this equation lies in C
M
for some M and thus the set
of all algebraic numbers is countable.
3. Prove that there exist real numbers which are not algebraic.
Proof: If not, R
1
= { all algebraic numbers } is countable, a contra-
diction.
4. Is the set of all irrational real numbers countable?
Solution: If RQ is countable, then R
1
= (RQ)

Q is countable,
a contradiction. Thus R Q is uncountable.
11
5. Construct a bounded set of real numbers with exactly three limit points.
Solution: Put
A = {1/n : n N}
_
{1 + 1/n : n N}
_
{2 + 1/n : n N}.
A is bounded by 3, and A contains three limit points - 0, 1, 2.
6. Let E

be the set of all limit points of a set E. Prove that S

is
closed. Prove that E and E have the same limit points. (Recall that
E = E

.) Do E and E

always have the same limit points?

Proof: For any point p of X E

, that is, p is not a limit point E,

there exists a neighborhood of p such that q is not in E with q = p for
every q in that neighborhood.
Hence, p is an interior point of X E

, that is, X E

is open, that
is, E

is closed.
Next, if p is a limit point of E, then p is also a limit point of E since
E = E

. If p is a limit point of E, then every neighborhood N

r
(p)
of p contains a point q = p such that q E. If q E, we completed
the proof. So we suppose that q E E = E

E. Then q is a limit
point of E. Hence,
N
r
(q)
where r

=
1
2
min(rd(p, q), d(p, q)) is a neighborhood of q and contains
a point x = q such that x E. Note that N
r
(q) contains in N
r
(p){p}.
That is, x = p and x is in N
r
(p). Hence, q also a limit point of E. Hence,
E and E have the same limit points.
Last, the answer of the nal sub-problem is no. Put
E = {1/n : n N},
and E

= {0} and (E

= .
7. Let A
1
, A
2
, A
3
, ... be subsets of a metric space. (a) If B
n
=

n
i=1
A
i
,
prove that B
n
=

n
i=1
A
i
, for n = 1, 2, 3, ... (b) If B =

i=1
, prove that
B

i=1
A
i
. Show, by an example, that this inclusion can be proper.
Proof of (a): (Method 1) B
n
is the smallest closed subset of X that
contains B
n
. Note that

A
i
is a closed subset of X that contains B
n
,
thus
B
n

n
_
i=1
A
i
.
12 CHAPTER 2. BASIC TOPOLOGY
If p B
n
B
n
, then every neighborhood of p contained a point q = p
such that q B
n
. If p is not in A
i
for all i, then there exists some
neighborhood N
r
i
(p) of p such that (N
r
i
(p) p)

A
i
= for all i. Take
r = min{r
1
, r
2
, ..., r
n
}, and we have N
r
(p)

B
n
= , a contradiction.
Thus p A
i
for some i. Hence
B
n

n
_
i=1
A
i
.
that is,
B
n
=
n
_
i=1
A
i
.
(Method 2) Since

n
i=1
A
i
is closed and B
n
=

n
i=1
A
i


n
i=1
A
i
,
B
n

n
i=1
A
i
.
Proof of (b): Since B is closed and B B A
i
, B A
i
for all i.
Hence B

A
i
.
Note: My example is A
i
= (1/i, ) for all i. Thus, A
i
= [1/i, ), and
B = (0, ), B = [0, ). Note that 0 is not in A
i
for all i. Thus this
inclusion can be proper.
8. Is every point of every open set E R
2
a limit point of E? Answer
the same question for closed sets in R
2
.
Solution: For the rst part of this problem, the answer is yes.
(Reason): For every point p of E, p is an interior point of E. That
is, there is a neighborhood N
r
(p) of p such that N
r
(p) is a subset of
E. Then for every real r

, we can choose a point q such that d(p, q) =

1/2 min(r, r

). Note that q = p, q N
r
(p), and q N
r
(p). Hence, every
neighborhood N
r
(p) contains a point q = p such that q N
r
(p) E,
that is, p is a limit points of E.
For the last part of this problem, the answer is no. Consider A =
{(0, 0)}. A

= and thus (0, 0) is not a limit point of E.

9. Let E
o
denote the set of all interior points of a set E.
(a) Prove that E
o
is always open.
(b) Prove that E is open if and only if E
o
= E.
13
(c) If G is contained in E and G is open, prove that G is contained in
E
o
.
(d) Prove that the complement of E
o
is the closure of the complement
of E.
(e) Do E and E always have the same interiors?
(f) Do E and E
o
always have the same closures?
Proof of (a): If E is non-empty, take p E
o
. We need to show that
p (E
o
)
o
. Since p E
o
, there is a neighborhood N
r
of p such that
N
r
is contained in E. For each q N
r
, note that N
s
(q) is contained in
N
r
(p), where s = min{d(p, q), r d(p, q)}. Hence q is also an interior
point of E, that is, N
r
is contained in E
o
. Hence E
o
is always open.
Proof of (b): () It is clear that E
o
is contained in E. Since E is
open, every point of E is an interior point of E, that is, E is contained
in E
o
. Therefore E
o
= E.
() Since every point of E is an interior point of E (E
o
(E) = E), E
is open.
Proof of (c): If p G, p is an interior point of G since G is open. Note
that E contains G, and thus p is also an interior point of E. Hence
p E
o
. Therefore G is contained in E
o
. (Thus E
o
is the biggest open
set contained in E. Similarly, E is the smallest closed set containing
E.)
Proof of (d): Suppose p X E
o
. If p X E, then p X E
clearly. If p E, then N is not contained in E for any neighborhood
N of p. Thus N contains an point q X E. Note that q = p, that
is, p is a limit point of X E. Hence X E
o
is contained in X E.
Next, suppose p X E. If p X E, then p X E
o
clearly. If
p E, then every neighborhood of p contains a point q = p such that
q X E. Hence p is not an interior point of E. Hence X E is
contained in X E
o
. Therefore X E
o
= X E.
Solution of (e): No. Take X = R
1
and E = Q. Thus E
o
= and
E
o
= (R
1
)
o
= R
1
= .
Solution of (f ): No. Take X = R
1
and E = Q. Thus E = R
1
, and
E
o
= = .
14 CHAPTER 2. BASIC TOPOLOGY
10. Let X be an innite set. For p X and q X, dene
d(p, q) =
_
1 (if p = q)
0 (if p = q)
Prove that this is a metric. Which subsets of the resulting metric space
are open? Which are closed? Which are compact?
Proof: (a) d(p, q) = 1 > 0 if p = q; d(p, p) = 0. (b) d(p, q) = d(q, p)
since p = q implies q = p and p = q implies q = p. (c) d(p, q)
d(p, r) + d(r, q) for any r X if p = q. If p = q, then either r = p or
r = q, that is, r = p or r = q. Thus, d(p, q) = 1 d(p, r) + d(r, q). By
(a)-(c) we know that d is a metric.
Every subset of X is open and closed. We claim that for any p X,
p is not a limit point. Since d(p, q) = 1 > 1/2 if q = p, there exists an
neighborhood N
1/2
(p) of p contains no points of q = p such that q X.
Hence every subset of X contains no limit points and thus it is closed.
Since X S is closed for every subset S of X, S = X (X S) is
open. Hence every subset of X is open.
Every nite subset of X is compact. Let S = {p
1
, ..., p
n
} be nite.
Consider an open cover {G

} of S. Since S is covered by G

, p
i
is
covered by G

i
, thus {G

1
, ..., G

n
} is nite subcover of S. Hence S is
compact. Next, suppose S is innite. Consider an open cover {G
p
} of
S, where
G
p
= N1
2
(p)
for every p S. Note that q is not in G
p
if q = p. If S is compact,
then S can be covered by nite subcover, say
G
p
1
, ..., G
p
n
.
Then there exists q such that q = p
i
for all i since S is innite, a con-
tradiction. Hence only every nite subset of X is compact.
11. For x R
1
and y R
1
, dene
d
1
(x, y) = (x, y)
2
,
d
2
(x, y) =
_
|x y|,
15
d
3
(x, y) = |x
2
y
2
|,
d
4
(x, y) = |x 2y|,
d
5
(x, y) =
|x y|
1 +|x y|
.
Determine, for each of these, whether it is a metric or not.
Solution: (1) d
1
(x, y) is not a metric. Since d
1
(0, 2) = 4, d
1
(0, 1) = 1,
and d
1
(1, 2) = 1, d
1
(0, 2) > d
1
(0, 1) + d
1
(1, 2). Thus d
1
(x, y) is not a
metric.
(2) d
2
(x, y) is a metric. (a) d(p, q) > 0 if p = q; d(p, p) = 0. (b)
d(p, q) =
_
|p q| =
_
|q p| = d(q, p). (c) |p q| |p r| + |r q|,
_
|p q|
_
|p r| +|r q|
_
|p r| +
_
|r q|. That is, d(p, q)
d(p, r) + d(r, q). (3) d
3
(x, y) is not a metric since d
3
(1, 1) = 0.
(4) d
4
(x, y) is not a metric since d
4
(1, 1) = 1 = 0.
(5) d
5
(x, y) is a metric since |x y| is a metric.
Claim: d(x, y) is a metric, then
d

(x, y) =
d(x, y)
1 + d(x, y)
is also a metric.
Proof of Claim: (a) d

(p, q) > 0 if p = q; d(p, p) = 0. (b) d

(p, q) =
d

(q, p). (c) Let x = d(p, q), y = d(p, r), and z = d(r, q). Then x y+z.
d

(p, q) d

(p, r) + d

(r, q)

x
1 + x

y
1 + y
+
z
1 + z
x(1 + y)(1 + z) y(1 + z)(1 + x) + z(1 + x)(1 + y)
x + xy + xz + xyz (y + xy + yz + xyz) + (z + xz + yz + xyz)
x y + z + 2yz + xyz
x y + z
Thus, d

is also a metric.
16 CHAPTER 2. BASIC TOPOLOGY
12. Let K R
1
consist of 0 and the numbers 1/n, for n = 1, 2, 3, ....
Prove that K is compact directly from the denition (without using
the Heine-Borel theorem).
Proof: Suppose that {O

} is an arbitrary open covering of K. Let

E {O

} consists 0. Since E is open and 0 E, 0 is an interior point

of E. Thus there is a neighborhood N = N
r
(0) of 0 such that N E.
Thus N contains
1
[1/r] + 1
,
1
[1/r] + 2
, ...
Next, we take nitely many open sets E
n
{O

} such that 1/n E

n
for n = 1, 2, ..., [1/r]. Hence {E, E
1
, ..., E
[1/r]
is a nite subcover of K.
Therefore, K is compact.
Note: The unique limit point of K is 0. Suppose p = 0 is a limit point
of K. Clearly, 0 < p < 1. (p cannot be 1). Thus there exists n Z
+
such that
1
n + 1
< p <
1
n
.
Hence N
r
(p) where r = min{
1
n
p, p
1
n+1
} contains no points of K,
a contradiction.
13. Construct a compact set of real numbers whose limit points form a
countable set.
Solution: Let K be consist of 0 and the numbers 1/n for n = 1, 2, 3, ...
Let xK = {xk : k K} and x + K = {x + k : k K} for x R
1
. I
take
S
n
= (1
1
2
n
) +
K
2
n+1
,
S =

_
n=1
S
n
_
{1}.
Claim: S is compact and the set of all limit points of S is K

{1}.
Clearly, S lies in [0, 1], that is, S is bounded in R
1
. Note that S
n

[1
1
2
n
, 1
1
2
n+1
]. By Exercise 12 and its note, I have that all limit
points of S

[0, 1) is
0,
1
2
, ...,
1
2
n
, ...
17
Clearly, 1 is also a limit point of S. Therefore, the set of all limit points
of S is K

{1}. Note that K

{1} S, that is, K is compact. I com-

pleted the proof of my claim.
14. Give an example of an open cover of the segment (0, 1) which has no
nite subcover.
Solution: Take {O
n
} = {(1/n, 1)} for n = 1, 2, 3, .... The following is
my proof. For every x (0, 1),
x (
1
[1/x] + 1
, 0) {O
n
}
Hence {O
n
} is an open covering of (0, 1). Suppose there exists a nite
subcovering
(
1
n
1
, 1), ..., (
1
n
k
, 1)
where n
1
< n
2
< ... < n
k
, respectively. Clearly
1
2n
p
(0, 1) is not in
any elements of that subcover, a contradiction.
Note: By the above we know that (0, 1) is not compact.
15. Show that Theorem 2.36 and its Corollary become false (in R
1
, for ex-
ample) if the word compact is replaced by closed or by bounded.
Theorem 2.36: If {K

} is a collection of compact subsets of a metric

space X such that the intersection of every nite subcollection of {K

}
is nonempty, then

K

is nonempty.
Corollary: If {K
n
} is a sequence of nonempty compact sets such that
K
n
contains K
n+1
(n = 1, 2, 3, ...), then

K
n
is not empty.
Solution: For closed: [n, ). For bounded: (1/n, 1/n) {0}.
16. Regard Q, the set of all rational numbers, as a metric space, with
d(p, q) = |p q|. Let E be the set of all p Q such that 2 < p
2
< 3.
Show that E is closed and bounded in Q, but that E is not compact.
Is E open in Q?
18 CHAPTER 2. BASIC TOPOLOGY
Proof: Let S = (

2,

3)

3,

2). Then E = {p Q : p S}.

Clearly, E is bounded in Q. Since Q is dense in R, every limit point of
Q is in Q. (I regard Q as a metric space). Hence, E is closed in Q.
To prove that E is not compact, we form a open covering of E as
follows:
{G

} = {N
r
(p) : p E and (p r, p + r) S}
Surely, {G

} is a open covering of E. If E is compact, then there are

nitely many indices
1
, ...,
n
such that
E G

1
_
...
_
G

n
.
For every G

i
= N
r
i
(p
i
), take p = max
1in
p
i
. Thus, p is the nearest
point to

3. But N
r
(p) lies in E, thus [p + r,

3) cannot be covered
since Q is dense in R, a contradiction. Hence E is not compact.
Finally, the answer is yes. Take any p Q, then there exists a
neighborhood N(p) of p contained in E. (Take r small enough where
N
r
(p) = N(p), and Q is dense in R.) Thus every point in N(p) is also
in Q. Hence E is also open.
17. Let E be the set of all x [0, 1] whose decimal expansion contains only
the digits 4 and 7. Is E countable? Is E dense in [0, 1]? Is E compact?
Is E perfect?
Solution:
E =
_

n=1
a
n
10
n
: a
n
= 4 or a
n
= 7
_
Claim: E is uncountable.
Proof of Claim: If not, we list E as follows:
x
1
= 0.a
11
a
12
...a
1n
...
x
2
= 0.a
21
a
22
...a
2n
...
... ...
x
k
= 0.a
k1
a
k2
...a
kn
...
... ...
19
(Prevent ending with all digits 9) Let x = 0.x
1
x
2
...x
n
... where
x
n
=
_
4 if a
nn
= 7
7 if a
nn
= 4
By my construction, x / E, a contradiction. Thus E is uncountable.
Claim: E is not dense in [0, 1].
Proof of Claim: Note that E

(0.47, 0.74) = . Hence E is not dense

in [0, 1].
Claim: E is compact.
Proof of Claim: Clearly, E is bounded. For every limit point p of E,
I show that p E. If not, write the decimal expansion of p as follows
p = 0.p
1
p
2
...p
n
...
Since p / E, there exists the smallest k such that p
k
= 4 and p
k
= 7.
When p
k
= 0, 1, 2, 3, select the smallest l such that p
l
= 7 if possible.
(If l does not exist, then p < 0.4. Thus there is a neighborhood of p
such that contains no points of E, a contradiction.) Thus
0.p
1
...p
l1
4p
l+1
...p
k1
7 < p < 0.p
1
...p
k1
4.
Thus there is a neighborhood of p such that contains no points of E, a
contradiction.
When p
k
= 5, 6,
0.p
1
...p
k1
47 < p < 0.p
1
...p
k1
74.
Thus there is a neighborhood of p such that contains no points of E, a
contradiction.
When p
k
= 8, 9, it is similar. Hence E is closed. Therefore E is
compact.
Claim: E is perfect.
Proof of Claim: Take any p E, and I claim that p is a limit point
of E. Write p = 0.p
1
p
2
...p
n
... Let
x
k
= 0.y
1
y
2
...y
n
...
20 CHAPTER 2. BASIC TOPOLOGY
where
y
n
=
_

_
p
k
if k = n
4 if p
n
= 7
7 if p
n
= 4
Thus, |x
k
p| 0 as k . Also, x
k
= p for all k. Hence p is a limit
point of E. Therefore E is perfect.
18. Is there a nonempty perfect set in R
1
which contains no rational num-
ber?
Solution: Yes. The following claim will show the reason.
Claim: Given a measure zero set S, we have a perfect set P contains
no elements in S.
Proof of Claim: (due to SYLee). Since S has measure zero, there
exists a collection of open intervals {I
n
} such that
S
_
I
n
and

|I
n
| < 1.
Consider E = R
1

I
n
. E is nonempty since E has positive mea-
sure. Thus E is uncountable and E is closed. Therefore there exists
a nonempty perfect set P contained in E by Exercise 28. P

S = .
Thus P is our required perfect set.
19. (a) If A and B are disjoint closed sets in some metric space X, prove
that they are separated.
(b) Prove the same for disjoint open sets.
(c) Fix p X, > 0, dene A to be the set of all q X for which
d(p, q) < , dene B similarly, with > in place of <. Prove that A and
B are separated.
(d) Prove that every connected metric space with at least two points
is uncountable. Hint: Use (c).
Proof of (a): Recall the denition of separated: A and B are sep-
arated if A

B and A

B are empty. Since A and B are closed sets,

A = A and B = B. Hence A

B = A

B = A

B = . Hence A and
B are separated.
21
Proof of (b): Suppose A

B is not empty. Thus there exists p such

that p A and p B. For p A, there exists a neighborhood N
r
(p) of
p contained in A since A is open. For p B = B

, if p B, then
p A

B. Note that A and B are disjoint, and its a contradiction.

If p B

, then p is a limit point of B. Thus every neighborhood of p

contains a point q = p such that q B. Take an neighborhood N
r
(p)of
p containing a point q = p such that q B. Note that N
r
(p) A,
thus q A. With A and B are disjoint, we get a contradiction. Hence
A

(B) is empty.
Similarly, A

B is also empty. Thus A and B are separated.

Proof of (c): Suppose A

B is not empty. Thus there exists x such

that x A and x B. Since x A, d(p, x) < . x B = B

, thus
if x B, then d(p, x) > , a contradiction. The only possible is x is a
limit point of B. Hence we take a neighborhood N
r
(x) of x contains y
with y B where r =
d(x,p)
2
. Clearly, d(y, p) > . But,
d(y, p) d(y, x) + d(x, p)
< r + d(x, p)
=
d(x, p)
2
+ d(x, p)
=
+ d(x, p)
2
<
+
2
= .
A contradiction. Hence A

B is empty. Similarly, A

B is also empty.
Thus A and B are separated.
Note: Take care of > 0. Think a while and you can prove the next
sub-exercise.
Proof of (d): Let X be a connected metric space. Take p X, q X
with p = q, thus d(p, q) > 0 is xed. Let
A = {x X : d(x, p) < }; B = {x X : d(x, p) > }.
Take =
t
= td(p, q) where t (0, 1). Thus 0 < < d(p, q). p A
since d(p, p) = 0 < , and q B since d(p, q) > . Thus A and B are
non-empty.
22 CHAPTER 2. BASIC TOPOLOGY
By (c), A and B are separated. If X = A

B, then X is not connected,

a contradiction. Thus there exists y
t
X such that y / A

B. Let
E = E
t
= {x X : d(x, p) =
t
} y
t
.
For any real t (0, 1), E
t
is non-empty. Next, E
t
and E
s
are disjoint
if t = s (since a metric is well-dened). Thus X contains a uncount-
able set {y
t
: t (0, 1)} since (0, 1) is uncountable. Therefore, X is
uncountable.
Note: It is a good exercise. If that metric space contains only one
point, then it must be separated.
Similar Exercise Given by SYLee: (a) Let A = {x : d(p, x) < r}
and B = {x : d(p, x) > r} for some p in a metric space X. Show that
A, B are separated.
(b) Show that a connected metric space with at least two points must
be uncountable. [Hint: Use (a)]
Proof of (a): By denition of separated sets, we want to show A

B =
, and B

A = . In order to do these, it is sucient to show A

B =
. Let x A

B = , then we have:
(1) x A d(x, p) r(2) x B d(x, p) > r
It is impossible. So, A

B = .
Proof of (b): Suppose that C is countable, say C = a, b, x
3
, .... We
want to show C is disconnected. So, if C is a connected metric space
with at least two points, it must be uncountable. Consider the set
S = {d(a, x
i
) : x
i
C}, and thus let r R S and inf S < r < sup S.
And construct A and B as in (a), we have C = A

B, where A and B
are separated. That is C is disconnected.
Another Proof of (b): Let a C, b C, consider the continuous
function f from C into R dened by f(x) = d(x, a). So, f(C) is con-
nected and f(a) = 0, f(b) > 0. That is, f(C) is an interval. Therefore,
C is uncountable.
20. Are closures and interiors of connected sets always connected? (Look
at subsets of R
2
.)
23
Solution: Closures of connected sets is always connected, but interiors
of those is not. The counterexample is
S = N
1
(2, 0)
_
N
1
(2, 0)
_
{x axis} R
2
.
Since S is path-connected, S is connect. But S
o
= N
1
(2)

N
1
(2) is
disconnected clearly.
Claim: If S is a connected subset of a metric space, then S is con-
nected.
Pf of Claim: If not, then S is a union of two nonempty separated set
A and B. Thus A

B = A

B = . Note that
S = S T
= A
_
B T
= (A
_
B)

T
c
= (A

T
c
)
_
(B

T
c
)
where T = S S. Thus
(A

T
c
)

T
c
(A

T
c
)

T
c
A

B
= .
Hence (A

T
c
)

T
c
= . Similarly, A

T
c

(B

T
c
) = .
Now we claim that both A

T
c
and B

T
c
are nonempty. Suppose
that B

T
c
= . Thus
A

T
c
= S A

(S S)
c
= S
A

(A
_
B S)
c
= S
A

((A
_
B)

S
c
)
c
= S
A

((A
c

B
c
)
_
S) = S
(A

S)
_
(A

A
c

B
c
) = S
A

S = S.
24 CHAPTER 2. BASIC TOPOLOGY
Thus B is empty, a contradiction. Thus B

T
c
is nonempty. Similarly,
A

T
c
nonempty. Therefore S is a union of two nonempty separated
sets, a contradiction. Hence S is connected.
21. Let A and B be separated subsets of some R
k
, suppose a A, b B,
and dene
p(t) = (1 t)a + tb
for t R
1
. Put A
0
= p
1
(A), B
0
= p
1
(B). [Thus t A
0
if and only
if p(t) A.]
(a) Prove that A
0
and B
0
are separated subsets of R
1
.
(b) Prove that there exists t
0
(0, 1) such that p(t
0
) / A

B.
(c) Prove that every convex subset of R
k
is connected.
Proof of (a): I claim that A
0

B
0
is empty. (B
0

A
0
is similar). If
not, take x A
0

B
0
. x A
0
and x B
0
. x B
0
or x is a limit
point of B
0
. x B
0
will make x A
0

B
0
, that is, p(x) A

B, a
contradiction since A and B are separated.
Claim: x is a limit point of B
0
p(x) is a limit point of B. Take any
neighborhood N
r
of p(x), and p(t) lies in B for small enough t. More
precisely,
x
r
|b a|
< t < x +
r
|b a|
.
Since x is a limit point of B
0
, and (x r/|b a|, x + r/|b a|) is a
neighborhood N of x, thus N contains a point y = x such that y B
0
,
that is, p(y) B. Also, p(y) N
r
. Therefore, p(x) is a limit point of
B. Hence p(x) A

B, a contradiction since A and B are separated.

Hence A
0

B
0
is empty, that is, A
0
and B
0
are separated subsets of
R
1
.
Proof of (b): Suppose not. For every t
0
(0, 1), neither p(t
0
) A
nor p(t
0
) B (since A and B are separated). Also, p(t
0
) A

B for
all t
0
(0, 1). Hence (0, 1) = A
0

B
0
, a contradiction since (0, 1) is
connected. I completed the proof.
Proof of (c): Let S be a convex subset of R
k
. If S is not connected,
then S is a union of two nonempty separated sets A and B. By (b),
there exists t
0
(0, 1) such that p(t
0
) / A

B. But S is convex, p(t

0
)
25
must lie in A

B, a contradiction. Hence S is connected.

22. A metric space is called separable if it contains a countable dense sub-
set. Show that R
k
is separable. Hint: Consider the set of points which
have only rational coordinates.
Proof: Consider S = the set of points which have only rational coor-
dinates. For any point x = (x
1
, x
2
, ..., x
k
) R
k
, we can nd a rational
sequence {r
i
j
} x
j
for j = 1, ..., k since Q is dense in R
1
. Thus,
r
i
= (r
i
1
, r
i
2
, ..., r
i
k
) x
and r
i
S for all i. Hence S is dense in R
k
. Also, S is countable, that
is, S is a countable dense subset in R
k
, R
k
is separable.
23. A collection {V

} of open subsets of X is said to be a base for X if the

following is true: For every x X and every open set G X such that
x G, we have x V

G for some . In other words, every open set

in X is the union of a subcollection of {V

}.
Prove that every separable metric space has a countable base. Hint:
Take all neighborhoods with rational radius and center in some count-
able dense subset of X.
Proof: Let X be a separable metric space, and S be a countable dense
subset of X. Let a collection {V

} = { all neighborhoods with rational

radius and center in S }. We claim that {V

} is a base for X.
For every x X and every open set G X such that x G, there
exists a neighborhood N
r
(p) of p such that N
r
(p) G since x is an
interior point of G. Since S is dense in X, there exists {s
n
} x. Take
a rational number r
n
such that r
n
<
r
2
, and {V

} N
r
n
(s
n
) N
r
(p)
for enough large n. Hence we have x V

G for some . Hence

{V

} is a base for X.
24. Let X be a metric space in which every innite subset has a limit
point. Prove that X is separable. Hint: Fix > 0, and pick x
1

X. Having chosen x
1
, ..., x
j
X, choose x
j+1
, if possible, so that
d(x
i
, x
j+1
) for i = 1, ..., j. Show that this process must stop after
26 CHAPTER 2. BASIC TOPOLOGY
nite number of steps, and that X can therefore be covered by nite
many neighborhoods of radius . Take = 1/n(n = 1, 2, 3, ...), and
consider the centers of the corresponding neighborhoods.
Proof: Fix > 0, and pick x
1
X. Having chosen x
1
, ..., x
j
X,
choose x
j+1
, if possible, so that d(x
i
, x
j+1
) for i = 1, ..., j. If
this process cannot stop, then consider the set A = {x
1
, x
2
, ..., x
k
}. If
p is a limit point of A, then a neighborhood N
/3
(p) of p contains a
point q = p such that q A. q = x
k
for only one k N. If not,
d(x
i
, x
j
) d(x
i
, p) + d(x
j
, p) /3 + /3 < , and it contradicts the
fact that d(x
i
, x
j
) for i = j. Hence, this process must stop after
nite number of steps.
Suppose this process stop after k steps, and X is covered by N

(x
1
),
N

(x
2
), ..., N

(x
k
), that is, X can therefore be covered by nite many
neighborhoods of radius .
Take = 1/n(n = 1, 2, 3, ...), and consider the set A of the centers of
the corresponding neighborhoods.
Fix p X. Suppose that p is not in A, and every neighborhood
N
r
(p). Note that N
r/2
(p) can be covered by nite many neighborhoods
N
s
(x
1
), ..., N
s
(x
k
) of radius s = 1/n where n = [2/r] + 1 and x
i
A
for i = 1, ..., k. Hence, d(x
1
, p) d(x
1
, q) +d(q, p) r/2 +s < r where
q N
r/2
(p)

N
s
(x
1
). Therefore, x
1
N
r
(p) and x
1
= p since p is not
in A. Hence, p is a limit point of A if p is not in A, that is, A is a
countable dense subset, that is, X is separable.
25. Prove that every compact metric space K has a countable base, and
that K is therefore separable. Hint: For every positive integer n, there
are nitely many neighborhood of radius 1/n whose union covers K.
Proof: For every positive integer n, there are nitely many neighbor-
hood of radius 1/n whose union covers K (since K is compact). Collect
all of them, say {V

}, and it forms a countable collection. We claim

{V

} is a base.
For every x X and every open set G X, there exists N
r
(x) such that
N
r
(x) G since x is an interior point of G. Hence x N
m
(p) {V

}
for some p where m = [2/r] + 1. For every y N
m
(p), we have
d(y, x) d(y, p) + d(p, x) < m + m = 2m < r.
27
Hence N
m
(p) G, that is, V

G for some , and therefore {V

} is a
countable base of K. Next, collect all of the center of V

, say D, and
we claim D is dense in K (D is countable since V

is countable). For all

p K and any > 0 we can nd N
n
(x
n
) {V

} where n = [1/] + 1.
Note that x
n
D for all n and d(p, x
n
) 0 as n . Hence D is
dense in K.
26. Let X be a metric space in which every innite subsets has a limit
point. Prove that X is compact. Hint: By Exercises 23 and 24, X has
a countable base. It follows that every open cover of X has a countable
subcover {G
n
}, n = 1, 2, 3, .... If no nite subcollection of {G
n
} covers
X, then the complement F
n
of G
1

...

G
n
is nonempty for each n,
but

F
n
is empty. If E is a set contains a point from each F
n
, consider
a limit point of E, and obtain a contradiction.
Proof: By Exercises 23 and 24, X has a countable base. It follows that
every open cover of X has a countable subcover {G
n
}, n = 1, 2, 3, ....
If no nite subcollection of {G
n
} covers X, then the complement F
n
of
G
1

...

G
n
is nonempty for each n, but

F
n
is empty. If E is a set
contains a point from each F
n
, consider a limit point of E.
Note that F
k
F
k+1
... and F
n
is closed for all n, thus p lies in F
k
for all k. Hence p lies in

F
n
, but

F
n
is empty, a contradiction.
27. Dene a point p in a metric space X to be a condensation point of a set
E X if every neighborhood of p contains uncountably many points
of E.
Suppose E R
k
, E is uncountable, and let P be the set of all conden-
sation points of E. Prove that P is perfect and that at most countably
many points of E are not in P. In other words, show that P
c

E is at
most countable. Hint: Let {V
n
} be a countable base of R
k
, let W be
the union of those V
n
for which E

V
n
is at most countable, and show
that P = W
c
.
Proof: Let {V
n
} be a countable base of R
k
, let W be the union of
those V
n
for which E

V
n
is at most countable, and we will show that
P = W
c
. Suppose x P. (x is a condensation point of E). If
x V
n
for some n, then E

V
n
is uncountable since V
n
is open. Thus
28 CHAPTER 2. BASIC TOPOLOGY
x W
c
. (If x W, then there exists V
n
such that x V
n
and E

V
n
is uncountable, a contradiction). Therefore P W
c
.
Conversely, suppose x W
c
. x / V
n
for any n such that E

V
n
is
countable. Take any neighborhood N(x) of x. Take x V
n
N(x),
and E

V
n
is uncountable. Thus E

N(x) is also uncountable, x is a

condensation point of E. Thus W
c
P. Therefore P = W
c
. Note that
W is countable, and thus W W

E = P
c

E is at most countable.
To show that P is perfect, it is enough to show that P contains no
isolated point. (since P is closed). If p is an isolated point of P, then
there exists a neighborhood N of p such that N

E = . p is not a
condensation point of E, a contradiction. Therefore P is perfect.
28. Prove that every closed set in a separable metric space is the union
of a (possible empty) perfect set and a set which is at most countable.
(Corollary: Every countable closed set in R
k
has isolated points.) Hint:
Use Exercise 27.
Proof: Let X be a separable metric space, let E be a closed set on
X. Suppose E is uncountable. (If E is countable, there is nothing
to prove.) Let P be the set of all condensation points of E. Since X
has a countable base, P is perfect, and P
c

E is at most countable by
Exercise 27. Since E is closed, P E. Also, P
c

E = E P. Hence
E = P

(E P).
For corollary: if there is no isolated point in E, then E is perfect. Thus
E is uncountable, a contradiction.
Note: Its also called Cauchy-Bendixon Theorem.
29. Prove that every open set in R
1
is the union of an at most countable
collection of disjoint segments. Hint: Use Exercise 22.
Proof: (due to H.L.Royden, Real Analysis) Since O is open, for each x
in O, there is a y > x such that (x, y) O. Let b = sup{y : (x, y) O}.
Let a = inf{z : (z, x) O}. Then a < x < b, and I
x
= (a, b) is an open
interval containing x.
Now I
x
O, for if w I
x
, say x < w < b, we have by the denition of
b a number y > w such that (x, y) O, and so w O).
29
Moreover, b / O, for if b O, then for some > 0 we have (b, b+)
O, whence (x, b + ) O, contradicting the denition of b. Similarly,
a / O.
Consider the collection of open intervals {I
x
}, x O. Since each x O
is contained in I
x
, and each I
x
O, we have O =

I
x
.
Let (a, b) and (c, d) be two intervals in this collection with a point in
common. Then we must have c < b and a < d. Since c / O, it does not
belong to (a, b) and we have c a. Since a / O and hence not to (c, d),
we have a c. Thus a = c. Similarly, b = d, and (a, b) = (c, d). Thus
two dierent intervals in the collection {I
x
} must be disjoint. Thus
O is the union of the disjoint collection {I
x
} of open intervals, and it
remains only to show that this collection is countable. But each open
interval contains a rational number since Q is dense in R. Since we
have a collection of disjoint open intervals, each open interval contains
a dierent rational number, and the collection can be put in one-to-one
correspondence with a subset of the rationals. Thus it is a countable
collection.
30. Imitate the proof of Theorem 2.43 to obtain the following result:
If R
k
=

1
F
n
, where each F
n
is a closed subset of R
k
, then
at least one F
n
has a nonempty interior.
Equivalent statement: If G
n
is a dense open subset of R
k
, for
n = 1, 2, 3, ..., then

1
G
n
is not empty (in fact, it is dense
in R
k
).
(This is a special case of Baires theorem; see Exercise 22, Chap. 3, for
the general case.)
Proof: I prove Baires theorem directly. Let G
n
be a dense open
subset of R
k
for n = 1, 2, 3, .... I need to prove that

1
G
n
intersects
any nonempty open subset of R
k
is not empty.
Let G
0
is a nonempty open subset of R
k
. Since G
1
is dense and G
0
is nonempty, G
0

G
1
= . Suppose x
1
G
0

G
1
. Since G
0
and G
1
are open, G
0

G
1
is also open, that is, there exist a neighborhood V
1
such that V
1
G
0

G
1
. Next, since G
2
is a dense open set and V
1
is a nonempty open set, V
1

G
2
= . Thus, I can nd a nonempty
30 CHAPTER 2. BASIC TOPOLOGY
open set V
2
such that V
2
V
1

G
2
. Suppose I have get n nonempty
open sets V
1
, V
2
, ..., V
n
such that V
1
G
0

G
1
and V
i+1
V
i

G
n+1
for all i = 1, 2, ..., n 1. Since G
n+1
is a dense open set and V
n
is a
nonempty open set, V
n

G
n+1
is a nonempty open set. Thus I can nd
a nonempty open set V
n+1
such that V
n+1
V
n

G
n+1
. By induction, I
can form a sequence of open sets {V
n
: n Z
+
} such that V
1
G
0

G
1
and V
i+1
V
i

G
i+1
for all n Z
+
. Since V
1
is bounded and V
1

V
2
... V
n
..., by Theorem 2.39 I know that

n=1
V
n
= .
Since V
1
G
0

G
1
and V
n+1
G
n+1
, G
0

n=1
G
n
) = . Proved.
Note: By Baires theorem, Ive proved the equivalent statement. Next,
F
n
has a empty interior if and only if G
n
= R
k
F
n
is dense in R
k
.
Hence we completed all proof.
Chapter 3
Numerical Sequences and
Series
31
32 CHAPTER 3. NUMERICAL SEQUENCES AND SERIES
Written by Men-Gen Tsai
email: b89902089@ntu.edu.tw
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13. Prove that the Cauchy product of two absolutely convergent series con-
verges absolutely.
Note: Given

a
n
and

b
n
, we put c
n
=

n
k=0
a
k
b
nk
and call

c
n
the Cauchy product of the two given series.
Proof: Put A
n
=

n
k=0
|a
k
|, B
n
=

n
k=0
|b
k
|, C
n
=

n
k=0
|c
k
|. Then
C
n
= |a
0
b
0
| +|a
0
b
1
+ a
1
b
0
| + ... +|a
0
b
n
+ a
1
b
n1
+ ... + a
n
b
0
|
|a
0
||b
0
| + (|a
0
||b
1
| +|a
1
||b
0
|) + ...
+(|a
0
||b
n
| +|a
1
||b
n1
| + ... +|a
n
||b
0
|)
= |a
0
|B
n
+|a
1
|B
n1
+ ... +|a
n
|B
0
33
|a
0
|B
n
+|a
1
|B
n
+ ... +|a
n
|B
n
= (|a
0
| +|a
1
| + ... +|a
n
|)B
n
= A
n
B
n
AB
where A = limA
n
and B = limB
n
. Hence {C
n
} is bounded. Note
that {C
n
} is increasing, and thus C
n
is a convergent sequence, that
is, the Cauchy product of two absolutely convergent series converges
absolutely.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23. Suppose {p
n
} and {q
n
} are Cauchy sequences in a metric space X.
Show that the sequence {d(p
n
, q
n
)} converges. Hint: For any m, n,
d(p
n
, q
n
) d(p
n
, p
m
) + d(p
m
, q
m
) + d(q
m
, q
n
);
it follows that
|d(p
n
, q
n
) d(p
m
, q
m
)|
is small if m and n are large.
Proof: For any > 0, there exists N such that d(p
n
, p
m
) < and
d(q
m
, q
n
) < whenever m, n N. Note that
d(p
n
, q
n
) d(p
n
, p
m
) + d(p
m
, q
m
) + d(q
m
, q
n
).
34 CHAPTER 3. NUMERICAL SEQUENCES AND SERIES
It follows that
|d(p
n
, q
n
) d(p
m
, q
m
)| d(p
n
, p
m
) + d(q
m
, q
n
) < 2.
Thus {d(p
n
, q
n
)} is a Cauchy sequence in X. Hence {d(p
n
, q
n
)} con-
verges.
24.
25.
Chapter 4
Continuity
35
36 CHAPTER 4. CONTINUITY
Written by Men-Gen Tsai
email: b89902089@ntu.edu.tw
1. Suppose f is a real function dene on R
1
which satises
lim
h0
[f(x + h) f(x h)] = 0
for every x R
1
. Does this imply that f is continuous?
Solution: No. Take f(x) = 1, if x Z; f(x) = 0. otherwise.
2.
3.
4.
5. If f is a real continuous function dened on a closed set E R
1
, prove
that there exist continuous real function g on R
1
such that g(x) = f(x)
for all x E. (Such functions g are called continuous extensions of f
from E to R
1
.) Show that the result becomes false if the word closed
is omitted. Extend the result to vector valued functions. Hint: Let the
graph of g be a straight line on each of the segments which constitute
the complement of E (compare Exercise 29, Chap. 2). The result re-
mains true if R
1
is replaced by any metric space, but the proof is not
so simple.
Proof: Note that the following fact:
Every open set of real numbers is the union of a countable
collection of disjoint open intervals.
Thus, consider E
c
=

(a
i
, b
i
), where i Z, and a
i
< b
i
< a
i+1
< b
i+1
.
We extend g on (a
i
, b
i
) as following:
g(x) = f(a
i
) + (x a
i
)
f(b
i
) f(a
i
)
b
i
a
i
(g(x) = f(x) for x E). Thus g is well-dened on R
1
, and g is contin-
uous on R
1
clearly.
37
Next, consider f(x) = 1/x on a open set E = R0. f is continuous on
E, but we cannot redene f(0) = any real number to make new f(x)
continue at x = 0.
Next, consider a vector valued function
f(x) = (f
1
(x), ..., f
n
(x)),
where f
i
(x) is a real valued function. Since f is continuous on E, each
component of f, f
i
, is also continuous on E, thus we can extend f
i
, say
g
i
, for each i = 1, ..., n. Thus,
g(x) = (g
1
(x), ..., g
n
(x))
is a extension of f(x) since each component of g, g
i
, is continuous on
R
1
implies g is continuous on R
n
.
Note: The above fact only holds in R
1
. If we change R
1
into any
metric spaces, we have no the previous fact.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
38 CHAPTER 4. CONTINUITY
17.
18.
19. Suppose f is a real function with domain R
1
which has the interme-
diate value property: If f(a) < c < f(b), then f(x) = c for some x
between a and b.
Suppose also, for every rational r, that the set of all x with f(x) = r
is closed. Prove that f is continuous.
Hint: If x
n
x
0
, but f(x
n
) > r > f(x
0
) for some r and all n, then
f(t
n
) = r for some t
n
between x
0
and x
n
; thus t
n
x
0
. Find a contra-
diction. (N. J. Fine, Amer. Math. Monthly, vol. 73, 1966, p. 782.)
Proof: Let S = {x : f(x) = r}. If x
n
x
0
, but f(x
n
) > r > f(x
0
)
for some r and all n since Q is dense in R
1
, then f(t
n
) = r for some t
n
between x
0
and x
n
; thus t
n
x
0
. Hence x
0
is a limit point of S. Since
S is closed, f(x
0
) = r, a contradiction. Hence, limsup f(x
n
) f(x
0
).
Similarly, liminf f(x
n
) f(x
0
). Hence, limf(x
n
) = f(x
0
), and f is
continuous at x
0
.
Note: Original problem is stated as follows:
Let f be a function from the reals to the reals, dierentiable
at every point. Suppose that, for every r, the set of points
x, where f

(x) = r, is closed. Prove that f

is continuous.
If we replace Q into any dense subsets of R
1
, the conclusion also holds.
20.
21.
22. Let A and B be disjoint nonempty closed sets in a metric space X, and
dene
f(p) =

A
(p)

A
(p) +
B
(p)
, (p X).
39
Show that f is a continuous function on X whose range lies in [0, 1],
that f(p) = 0 precisely on A and f(p) = 1 precisely on B. This
establishes a converse of Exercise 3: Every closed set A contained in X
is Z(f) for some continuous real f on X. Setting
V = f
1
([0, 1/2)), W = f
1
((1/2, 1]),
show that V and W are open and disjoint, and that A is contained in
V , B is contained in W. (Thus pairs of disjoint closed set in a metric
space can be covered by pairs of disjoint open sets. This property of
metric spaces is called normality.)
Proof: Note that
A
(p) and
B
(p) are (uniformly) continuous on X,
and
A
(p) +
B
(p) > 0. (Clearly,
A
(p) +
B
(p) 0 by the denition.
If
A
(p) +
B
(p) = 0, then p A

B by Exercise 20, a contradiction).

Thus f(p) =
A
(p)/(
A
(p)+
B
(p)) is continuous on X. Next, f(p) 0,
and f(p) 1 since
A
(p)
A
(p) +
B
(p). Thus f(X) lies in [0, 1].
Next, f(p) = 0
A
(p) = 0 p A precisely, and f(p) = 1

B
(p) = 0 p B precisely by Exercise 20.
Now we prove a converse of Exercise 3: Every closed set A X is Z(f)
for some continuous real f on X. If Z(f) = , then f(x) = 1 for all
x X satises our requirement. If Z(f) = , we consider two possible
cases: (1) Z(f) = X; (2) Z(f) = X. If Z(f) = X, then f(x) = 0 for
all x X. If Z(f) = X, we can choose p X such that f(p) = 0.
Note that Z(f) and {p} are one pair of disjoint closed sets. Hence we
let
f(x) =

Z(f)
(x)

Z(f)
(x) +
{p}
(x)
.
By the previous result, we know that f(x) satises our requirement.
Hence we complete the whole proof.
Note that [0, 1/2) and (1/2, 1] are two open sets of f(X). Since f is
continuous, V = f
1
([0, 1/2)) and W = f
1
((1/2, 1]) are two open sets.
f
1
({0}) f
1
([0, 1/2)), and f
1
({1}) f
1
((1/2, 1]). Thus, A V
and B W. Thus a metric space X is normal.
40 CHAPTER 4. CONTINUITY
23.
24.
25.
26.
Chapter 5
Dierentiation
41
42 CHAPTER 5. DIFFERENTIATION
Written by Men-Gen Tsai
email: b89902089@ntu.edu.tw
1. Let f be dened for all real x, and suppose that
|f(x) f(y)| (x y)
2
for all real x and y. Prove that f is constant.
Proof: |f(x) f(y)| (xy)
2
for all real x and y. Fix y, |
f(x)f(y)
xy
|
|x y|. Let x y, therefore,
0 lim
xy
f(x) f(y)
x y
lim
xy
|x y| = 0
It implies that (f(x) f(y))/(x y) 0 as x y. Hence f

(y) = 0,
f = const.
2. Suppose f

(x) > 0 in (a, b). Prove that f is strictly increasing in (a, b),
and let g be its inverse function. Prove that g is dierentible, and that
g

(f(x)) =
1
f

(x)
(a < x < b).
Proof: For every pair x > y in (a, b), f(x) f(y) = f

(c)(xy) where
y < c < x by Mean-Value Theorem. Note that c (a, b) and f

(x) > 0
in (a, b), hence f

(c) > 0. f(x) f(y) > 0, f(x) > f(y) if x > y, f is

strictly increasing in (a, b).
Let g = g(x
0
+ h) g(x
0
). Note that x
0
= f(g(x
0
)), and thus,
(x
0
+ h) x
0
= f(g(x
0
+ h)) f(g(x
0
)),
h = f(g(x
0
) + g) f(g(x
0
)) = f(g + g) f(g).
Thus we apply the fundamental lemma of dierentiation,
h = [f

(g) + (g)]g,
1
f

(g) + (g)
=
g
h
43
Note that f

(g(x)) > 0 for all x (a, b) and (g) 0 as h 0, thus,

lim
h0
g/h = lim
h0
1
f

(g) + (g)
=
1
f

(g(x))
.
Thus g

(x) =
1
f

(g(x))
, g

(f(x)) =
1
f

(x)
.
3. Suppose g is a real function on R
1
, with bounded derivative (say
|g

| M). Fix > 0, and dene f(x) = x + g(x). Prove that f

is one-to-one if is small enough. (A set of admissible values of can
be determined which depends only on M.)
Proof: For every x < y, and x, y R, we will show that f(x) = f(y).
By using Mean-Value Theorem:
g(x) g(y) = g

(c)(x y) where x < c < y,

(x y) + ((x) g(y)) = (g

(c) + 1)(x y),

that is,
f(x) f(y) = (g

(c) + 1)(x y). ()

Since |g

(x)| M, M g

(x) M for all x R. Thus 1 M

g

(c) +1 1 +M, where x < c < y. Take c =

1
2M
, and g

(c) +1 > 0
where x < c < y for all x, y. Take into equation (*), and f(x)f(y) < 0
since xy < 0, that is, f(x) = f(y), that is, f is one-to-one (injective).
4. If
C
0
+
C
1
2
+ ... +
C
n1
n
+
C
n
n + 1
= 0,
where C
0
, ..., C
n
are real constants, prove that the equation
C
0
+ C
1
x + ... + C
n1
x
n1
+ C
n
x
n
= 0
has at least one real root between 0 and 1.
Proof: Let f(x) = C
0
x + ... +
C
n
n+1
x
n+1
. f is dierentiable in R
1
and f(0) = f(1) = 0. Thus, f(1) f(0) = f

(c) where c (0, 1) by

Mean-Value Theorem. Note that
f

(x) = C
0
+ C
1
x + ... + C
n1
x
n1
+ C
n
x
n
.
44 CHAPTER 5. DIFFERENTIATION
Thus, c (0, 1) is one real root between 0 and 1 of that equation.
5. Suppose f is dened and dierentiable for every x > 0, and f

(x) 0
as x +. Put g(x) = f(x + 1) f(x). Prove that g(x) 0 as
x +.
Proof: f(x + 1) f(x) = f

(c)(x + 1 x) where x < c < x + 1

by Mean-Value Theorem. Thus, g(x) = f

(c) where x < c < x + 1,

that is,
lim
x+
g(x) = lim
x+
f

(c) = lim
c+
f

(c) = 0.
6. Suppose
(a) f is continuous for x 0,
(b) f

(x) exists for x > 0,

(c) f(0) = 0,
(d) f

is monotonically increasing.
Put
g(x) =
f(x)
x
(x > 0)
and prove that g is monotonically increasing.
Proof: Our goal is to show g

(x) > 0 for all x > 0

g

(x) =
xf

(x)f(x)
x
2
> 0 f

(x) >
f(x)
x
.
Since f

(x) exists, f(x) f(0) = f

(c)(x 0) where 0 < c < x by

Mean-Value Theorem. f

(c) =
f(x)
x
where 0 < c < x. Since f

is
monotonically increasing, f

(x) > f

(c), that is, f

(x) >
f(x)
x
for all
x > 0.
7. Suppose f

(x), g

(x) exist, g

(x) = 0, and f(x) = g(x) = 0. Prove that

lim
tx
f(t)
g(t)
=
f

(x)
g

(x)
.
45
(This holds also for complex functions.)
Proof:
f

(t)
g

(t)
=
lim
tx
f(t)f(x)
tx
lim
tx
g(t)g(x)
tx
=
lim
tx
f(t)
lim
tx
g(t)
= lim
tx
f(t)
g(t)
Surely, this holds also for complex functions.
8. Suppose f

(x) is continuous on [a, b] and > 0. Prove that there exists

> 0 such that
|
f(t) f(x)
t x
f

(x)| <
whenever 0 < |t x| < , a x b, a t b. (This could be
expressed by saying f is uniformly dierentiable on [a, b] if f

is contin-
uous on [a, b].) Does this hold for vector-valued functions too?
Proof: Since f

(x) is continuous on a compact set [a, b], f

(x) is uni-
formly continuous on [a, b]. Hence, for any > 0 there exists > 0
such that
|f

(t) f

(x)| <
whenever 0 < |t x| < , a x b, a t b. Thus, f

(c) =
f(t)f(x)
tx
where c between t and x by Mean-Value Theorem. Note that
0 < |c x| < and thus |f

(c) f

(x)| < , thus,

|
f(t) f(x)
t x
f

(x)| <
whenever 0 < |t x| < , a x b, a t b.
Note: It does not hold for vector-valued functions. If not, take
f(x) = (cos x, sin x),
[a, b] = [0, 2], and x = 0. Hence f

(x) = (sin x, cos x). Take any

1 > > 0, there exists > 0 such that
|
f(t) f(0)
t 0
f

(0)| <
46 CHAPTER 5. DIFFERENTIATION
whenever 0 < |t| < by our hypothesis. With calculating,
|(
cos t 1
t
,
sin t
t
) (0, 1)| <
|(
cos t 1
t
,
sin t
t
1)| <
(
cos t 1
t
)
2
+ (
sin t
t
1)
2
<
2
<
2
t
2
+ 1
2(cos t + sin t)
t
<
since 1 > > 0. Note that
2
t
2
+ 1
4
t
<
2
t
2
+ 1
2(cos t + sin t)
t
But
2
t
2
+ 1
4
t
+ as t 0. It contradicts.
9. Let f be a continuous real function on R
1
, of which it is known that
f

(x) exists for all x = 0 and that f

(x) 0 as x 0. Dose it follow

that f

(0) exists?
Note: We prove a more general exercise as following.
Suppose that f is continuous on an open interval I containing x
0
, sup-
pose that f

is dened on I except possibly at x

0
, and suppose that
f

(x) L as x x
0
. Prove that f

(x
0
) = L.
Proof of the Note: Using LHospitals rule:
lim
h0
f(x
0
+ h) f(x
0
)
h
= lim
h0
f

(x
0
+ h)
By our hypothesis: f

(x) L as x x
0
. Thus,
lim
h0
f(x
0
+ h) f(x
0
)
h
= L,
Thus f

(x
0
) exists and
f

(x
0
) = L.
47
10. Suppose f and g are complex dierentiable functions on (0, 1), f(x) 0,
g(x) 0, f

(x) A, g

(x) B as x 0, where A and B are complex

numbers, B = 0. Prove that
lim
x0
f(x)
g(x)
=
A
B
.
Compare with Example 5.18. Hint:
f(x)
g(x)
= (
f(x)
x
A)
x
g(x)
+ A
x
g(x)
.
Apply Theorem 5.13 to the real and imaginary parts of
f(x)
x
and
g(x)
x
.
Proof: Write f(x) = f
1
(x) +if
2
(x), where f
1
(x), f
2
(x) are real-valued
functions. Thus,
df(x)
dx
=
df
1
(x)
dx
+ i
df
2
(x)
dx
,
Apply LHospitals rule to
f
1
(x)
x
and
f
2
(x)
x
, we have
lim
x0
f
1
(x)
x
= lim
x0
f

1
(x)
lim
x0
f
2
(x)
x
= lim
x0
f

2
(x)
Combine f
1
(x) and f
2
(x), we have
lim
x0
f
1
(x)
x
+ i lim
x0
f
2
(x)
x
= lim
x0
f
1
(x)
x
+ i
f
2
(x)
x
= lim
x0
f(x)
x
or
lim
x0
f
1
(x)
x
+ i lim
x0
f
2
(x)
x
= lim
x0
f

1
(x) + i lim
x0
f

2
(x) = lim
x0
f

(x)
Thus, lim
x0
f(x)
x
= lim
x0
f

(x). Similarly, lim

x0
g(x)
x
= lim
x0
g

(x).
Note that B = 0. Thus,
lim
x0
f(x)
g(x)
= lim
x0
(
f(x)
x
A)
x
g(x)
+ A
x
g(x)
= (A A)
1
B
+
A
B
=
A
B
.
48 CHAPTER 5. DIFFERENTIATION
Note: In Theorem 5.13, we know g(x) + as x 0. (f(x) = x,
and g(x) = x + x
2
e
i
x
2
).
11. Suppose f is dened in a neighborhood of x, and suppose f(x) exists.
Show that
lim
h0
f(x + h) + f(x h) 2f(x)
h
2
= f

(x).
Show by an example that the limit may exist even if f(x) dose not.
Hint: Use Theorem 5.13.
Proof: By using LHospitals rule: (respect to h.)
lim
h0
f(x + h) + f(x h) 2f(x)
h
2
= lim
h0
f

(x + h) f

(x h)
2h
Note that
f

(x) =
1
2
(f

(x) + f

(x))
=
1
2
(lim
h0
f

(x + h) f

(x)
h
+ lim
h0
f

(x h) f

(x)
h
)
=
1
2
lim
h0
f

(x + h) f

(x h)
h
= lim
h0
f

(x + h) f

(x h)
2h
Thus,
f(x + h) + f(x h) 2f(x)
h
2
f

(x)
as h 0. Counter-example: f(x) = x|x| for all real x.
12. If f(x) = |x|
3
, compute f

(x), f

(x) for all real x, and show that f

(3)
(0)
does not exist.
Proof: f

(x) = 3|x|
2
if x = 0. Consider
f(h) f(0)
h
=
|h|
3
h
49
Note that |h|/h is bounded and |h|
2
0 as h 0. Thus,
f

(0) = lim
h0
f(h) f(0)
h
= 0.
Hence, f

(x) = 3|x|
2
for all x. Similarly,
f

(x) = 6|x|.
Thus,
f

(h) f(0)
h
= 6
|h|
h
Since
|h|
h
= 1 if h > 0 and = 1 if h < 0, f

(0) does not exist.

13. Suppose a and c are real numbers, c > 0, and f is dened on [1, 1] by
f(x) =
_
x
a
sin (x
c
) (if x = 0),
0 (if x = 0).
Prove the following statements:
(a) f is continuous if and only if a > 0.
(b) f

(0) exists if and only if a > 1.

(c) f

is bounded if and only if a 1 + c.

(d) f

is continuous if and only if a > 1 + c.

(e) f

(0) exists if and only if a > 2 + c.

(f) f

is bounded if and only if a 2 + 2c.

(g) f

is continuous if and only if a > 2 + 2c.

Proof: For (a): () f is continuous i for any sequence {x
n
} 0
with x
n
= 0, x
a
n
sin x
c
n
0 as n . In particular, take
x
n
= (
1
2n + /2
)
1
c
> 0
and thus x
a
n
0 as n . Hence a > 0. (If not, then a = 0 or a < 0.
When a = 0, x
a
n
= 1. When a < 0, x
a
n
= 1/x
a
n
as n . It
contradicts.)
() f is continuous on [1, 1] {0} clearly. Note that
|x
a
| x
a
sin (x
c
) |x
a
|,
50 CHAPTER 5. DIFFERENTIATION
and |x
a
| 0 as x 0 since a > 0. Thus f is continuous at x = 0.
Hence f is continuous.
For (b): f

(0) exists i x
a1
sin (x
c
) 0 as x 0. In the previous
proof we know that f

(0) exists if and only if a1 > 0. Also, f

(0) = 0.
14. Let f be a dierentiable real function dened in (a, b). Prove that f is
convex if and only if f

is monotonically increasing. Assume next that

f

(x) exist for every x (a, b), and prove that f is convex if and only
if f

(x) 0 for all x (a, b).

15. Suppose a R
1
, f is a twice-dierentiable real function on (a, ),
and M
0
, M
1
, M
2
are the least upper bounds of |f(x)|, |f

(x)|, |f

(x)|,
respectively, on (a, ). Prove that
M
2
1
4M
0
M
2
.
Hint: If h > 0, Taylors theorem shows that
f

(x) =
1
2h
[f(x + 2h) f(x)] hf

()
for some (x, x + 2h). Hence
|f

(x)| hM
2
+
M
0
h
.
To show that M
2
1
= 4M
0
M
2
can actually happen, take a = 1, dene
f(x) =
_
2x
2
1 (1 < x < 0),
x
2
1
x
2
+1
(0 x < ),
and show that M
0
= 1, M
1
= 4, M
2
= 4. Does M
2
1
4M
0
M
2
hold for
vector-valued functions too?
Proof: Suppose h > 0. By using Taylors theorem:
f(x + h) = f(x) + hf

(x) +
h
2
2
f

()
for some x < < x + 2h. Thus
h|f

(x)| |f(x + h)| +|f(x)| +

h
2
2
|f

()|
51
h|f

(x)| 2M
0
+
h
2
2
M
2
.
h
2
M
2
2h|f

(x)| + 4M
0
0 ()
Since equation (*) holds for all h > 0, its determinant must be non-
positive.
4|f

(x)|
2
4M
2
(4M
0
) 0
|f

(x)|
2
4M
0
M
2
(M
1
)
2
2M
0
M
2
Note: There is a similar exercise:
Suppose f(x)( < x < +) is a twice-dierentiable real function,
and
M
k
= sup
<x<+
|f
(k)
(x)| < + (k = 0, 1, 2).
Prove that M
2
1
2M
0
M
2
.
Proof of Note:
f(x + h) = f(x) + f

(x)h +
f

(
1
)
2
h
2
(x <
1
< x + h or x >
1
> x + h) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (*)
f(x h) = f(x) f

(x)h +
f

(
2
)
2
h
2
(x h <
2
< x or x h >
2
> x) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (**)
(*) minus (**):
f(x + h) f(x h) = 2f

(x)h +
h
2
2
(f

(
1
) f

(
2
)).
2h|f

(x)| |2hf

(x)|
2h|f

(x)| |f(x + h)| +|f(x h)| +

h
2
2
(|f

(
1
)| +|f

(
2
)|)
2h|f

(x)| 2M
0
+ h
2
M
2
M
2
h
2
2|f

(x)|h + 2M
0
0
52 CHAPTER 5. DIFFERENTIATION
Since this equation holds for all h, its determinant must be non-positive:
4|f

(x)|
2
4M
2
(2M
0
) 0,
|f

(x)|
2
2M
0
M
2
Thus
M
2
1
2M
0
M
2
16. Suppose f is twice-dierentiable on (0, ), f

is bounded on (0, ),
and f(x) 0 as x . Prove that f

(x) 0 as x . Hint: Let

a in Exercise 15.
Proof: Suppose a (0, ), and M
0
, M
1
, M
2
are the least upper
bounds of |f(x)|, |f

(x)|, |f

(x)| on (a, ). Hence, M

2
1
4M
0
M
2
. Let
a , M
0
= sup |f(x)| 0. Since M
2
is bounded, therefore M
2
1
0
as a . It implies that sup |f

(x)| 0 as x .
17. Suppose f is a real, three times dierentiable function on [1, 1], such
that
f(1) = 0, f(0) = 0, f(1) = 1, f

(0) = 0.
Prove that f
(3)
(x) 3 for some x (1, 1).
Note that equality holds for
1
2
(x
3
+ x
2
).
Hint: Use Theorem 5.15, with = 0 and = 1, 1, to show that
there exist s (0, 1) and t (1, 0) such that
f
(3)
(s) + f
(3)
(t) = 6.
Proof: By Theorem 5.15, we take = 0, = 1,
f(1) = f(0) + f

(0) +
f

(0)
2
+
f
(3)
(s)
6
where s (0, 1). Take = 0, and = 1,
f(1) = f(0) f

(0) +
f

(0)
2

f
(3)
(t)
6
53
where t (1, 0). Thus
1 =
f

(0)
2
+
f
(3)
(s)
6
, s (0, 1) ()
0 =
f

(0)
2

f
(3)
(s)
6
, s (1, 0) ()
Equation (*) - equation (**):
f
(3)
(s)
6
+
f
(3)
(t)
6
, s (0, 1), t (1, 0).
f
(3)
(s) + f
(3)
(t) = 6, s, t (1, 1).
f
(3)
(x) 3 for some x (1, 1).
Theorem 5.15: Suppose f is a real function on [a, b], n is a positive
integer, f
(n1)
is continuous on [a, b], f
(n)
(t) exists for every t (a, b).
Let , be distinct points of [a, b], and dene
P(t) =
n1

k=0
f
(k)
()
k!
(t )
k
.
Then there exists a point x between and such that
f() = P() +
f
(n)
(x)
n!
( )
n
.
18. Suppose f is a real function on [a, b], n is a positive integer, and f
(n1)
exists for every t [a, b]. Let , , and P be as in Taylors theorem
(5.15). Dene
Q(t) =
f(t) f()
t
for t [a, b], t = , dierentiate
f(t) f() = (t )Q(t)
54 CHAPTER 5. DIFFERENTIATION
n 1 times at t = , and derive the following version of Taylors
theorem:
f() = P() +
Q
(n1)
()
(n 1)!
( )
n
.
19. Suppose f is dened in (1, 1) and f

(0) exists. Suppose 1 <

n
<

n
< 1,
n
0, and
n
0 as n . Dene the dierence quotients
D
n
=
f(
n
) f(
n
)

n
Prove the following statements:
(a) If
n
< 0 <
n
, then limD
n
= f

(0).
(b) If 0 <
n
<
n
and {
n
/(
n

n
)} is bounded, then limD
n
= f

(0).
(c) If f

is continuous in (1, 1), then limD

n
= f

(0).
Give an example in which f is dierentiable in (1, 1) (but f

is not
continuous at 0) and in which
n
,
n
tend to 0 in such a way that
limD
n
exists but is dierent from f

(0).
Proof: For (a):
D
n
=
f(
n
) f(0)

n
+
f(
n
) f(0)

n
Note that
f

(0) = lim
n
f(
n
) f(0)

n
= lim
n
f(
n
) f(0)

n
Thus for any > 0, there exists N such that
L <
f(
n
) f(0)

n
< L + ,
L <
f(
n
) f(0)

n
< L + ,
55
whenever n > N where L = f

(0) respectively. Note that

n
/(
n

n
)
and
n
/(
n

n
) are positive. Hence,

n
(L ) <
f(
n
) f(0)

n
<

n

n
(L + )

n
(L ) <
f(
n
) f(0)

n
<

n

n
(L + )
Combine two inequations,
L < D
n
< L +
Hence, limD
n
= L = f

(0).
For (b): We process as above prove, but note that
n
/(
n

n
) < 0.
Thus we only have the following inequations:

n
(L ) <
f(
n
) f(0)

n
<

n

n
(L + )

n
(L + ) <
f(
n
) f(0)

n
<

n

n
(L )
Combine them:
L

n
+
n

n
< D
n
< L +

n
+
n

Note that {
n
/(
n

n
)} is bounded, ie,
|

n

n
| M
for some constant M. Thus
|

n
+
n

n
| = |
2
n

n
1| 2M + 1
Hence,
L (2M + 1) < D
n
< L + (2M + 1)
Hence, limD
n
= L = f

(0).
56 CHAPTER 5. DIFFERENTIATION
For (c): By using Mean-Value Theorem,
D
n
= f

(t
n
)
where t
n
is between
n
and
n
. Note that
min{
n
,
n
} < t
n
< max{
n
,
n
}
and
max{
n
,
n
} =
1
2
(
n
+
n
+|
n

n
|)
min{
n
,
n
} =
1
2
(
n
+
n
|
n

n
|)
Thus, max{
n
,
n
} 0 and min{
n
,
n
} 0 as
n
0 and
n
0.
By squeezing principle for limits, t
n
0. With the continuity of f

,
we have
limD
n
= limf

(t
n
) = f

(limt
n
) = f

(0).
Example: Let f be dened by
f(x) =
_
x
2
sin(1/x) (x = 0),
0 (x = 0).
Thus f

(x) is not continuous at x = 0, and f

(0) = 0. Take
n
=
1
/2+2n
and
n
=
1
2n
. It is clear that
n
0, and
n
0 as n .
Also,
D
n
=
4n
(/2 + 2n)

2

as n . Thus, limD
n
= 2/ exists and is dierent from 0 = f

(0).
20.
21.
22. Suppose f is a real function on (, ). Call x a xed point of f if
f(x) = x.
(a) If f is dierentiable and f

(t) = 1 for every real t, prove that f has

at most one xed point.
57
(b) Show that the function f dened by
f(t) = t + (1 + e
t
)
1
has no xed point, although 0 < f

(t) < 1 for all real t.

(c) However, if there is a constant A < 1 such that |f

(t)| A for all

real t, prove that a xed point x of f exists, and that x = limx
n
, where
x
1
is an arbitrary real number and
x
n+1
= f(x
n
)
for n = 1, 2, 3, ...
(d) Show that the process describe in (c) can be visualized by the zig-
zag path
(x
1
, x
2
) (x
2
, x
2
) (x
2
, x
3
) (x
3
, x
3
) (x
3
, x
4
) ....
Proof: For (a): If not, then there exists two distinct xed points, say
x and y, of f. Thus f(x) = x and f(y) = y. Since f is diereniable,
by applying Mean-Value Theorem we know that
f(x) f(y) = f

(t)(x y)
where t is between x and y. Since x = y, f

(t) = 1. It contradicts.
For (b): We show that 0 < f

(t) < 1 for all real t rst:

f

(t) = 1 + (1)(1 + e
t
)
2
e
t
= 1
e
t
(1 + e
t
)
2
Since
e
t
> 0
(1 + e
t
)
2
= (1 + e
t
)(1 + e
t
) > 1(1 + e
t
) = 1 + e
t
> e
t
> 0
for all real t, thus
(1 + e
t
)
2
e
t
> 0
(1 + e
t
)
2
e
t
< 1
for all real t. Hence 0 < f

(t) < 1 for all real t.

58 CHAPTER 5. DIFFERENTIATION
Next, since f(t) t = (1 e
t
)
1
> 0 for all real t, f(t) has no xed
point.
For (c): Suppose x
n+1
= x
n
for all n. (If x
n+1
= x
n
, then x
n
= x
n+1
=
... and x
n
is a xed point of f).
By Mean-Value Theorem,
f(x
n+1
) f(x
n
) = f

(t
n
)(x
n+1
x
n
)
where t
n
is betweem x
n
and x
n+1
. Thus,
|f(x
n+1
) f(x
n
)| = |f

(t
n
)||(x
n+1
x
n
)|
Note that |f

(t
n
)| is bounded by A < 1, f(x
n
) = x
n+1
, and f(x
n+1
) =
x
n+2
. Thus
|x
n+2
x
n+1
| A|x
n+1
x
n
|
|x
n+1
x
n
| CA
n1
where C = |x
2
x
1
|. For two positive integers p > q,
|x
p
x
q
| |x
p
x
p1
| + ... +|x
q+1
x
q
|
= C(A
q1
+ A
q
+ ... + A
p2
)

CA
q1
1 A
.
Hence
|x
p
x
q
|
CA
q1
1 A
.
Hence, for any > 0, there exists N = [log
A
(1A)
C
] + 2 such that
|x
p
x
q
| < whenever p > q N. By Cauchy criterion we know that
{x
n
} converges to x. Thus,
lim
n
x
n+1
= f( lim
n
x
n
)
since f is continuous. Thus,
x = f(x).
x is a xed point of f.
For (d): Since x
n+1
= f(x
n
), it is trivial.
59
23.
24.
25. Suppose f is twice dierentiable on [a, b], f(a) < 0, f(b) > 0, f

(x)
> 0, and 0 f

(x) M for all x [a, b]. Let be the unique point

in (a, b) at which f() = 0.
Complete the details in the following outline of Newtons method for
computing .
(a) Choose x
1
(, b), and dene {x
n
} by
x
n+1
= x
n

f(x
n
)
f

(x
n
)
.
Interpret this geometrically, in terms of a tangent to the graph of f.
(b) Prove that x
n+1
< x
n
and that
lim
n>oo
x
n
= .
(c) Use Taylors theorem to show that
x
n+1
=
f

(t
n
)
2f

(x
n
)
(x
n
)
2
for some t
n
(, x
n
).
(d) If A =
M
2
, deduce that
0 x
n+1

1
A
[A(x
1
)]
2
n
.
(Compare with Exercise 16 and 18, Chap. 3)
(e) Show that Newtons method amounts to nding a xed point of the
function g dened by
g(x) = x
f(x)
f

(x)
.
How does g

(x) behave for x near ?

(f) Put f(x) = x
1/3
on (, ) and try Newtons method. What
happens?
60 CHAPTER 5. DIFFERENTIATION
Proof: For (a): You can see the picture in the following URL:
http://archives.math.utk.edu/visual.calculus/3/newton.5/1.html.
For (b): We show that x
n
x
n+1
. (induction). By Mean-Value
Theorem, f(x
n
) f() = f

(c
n
)(x
n
) where c
n
(, x
n
). Since
f

0, f

is increasing and thus

f(x
n
)
x
n

= f

(c
n
) f

(x
n
) =
f(x
n
)
x
n
x
n+1
f(x
n
)(x
n
) f(x
n
)(x
n
x
n+1
)
Note that f(x
n
) > f() = 0 since f

> 0 and f is strictly increasing.

Thus,
x
n
x
n
x
n+1
x
n+1
Note that f(x
n
) > 0 and f

(x
n
) > 0. Thus x
n+1
< x
n
. Hence,
x
n
> x
n+1
.
Thus, {x
n
} converges to a real number . Suppose = , then
x
n+1
= x
n

f(x
n
)
f

(x
n
)
Note that
f(x
n
)
f

(x
n
)
>
f()

. Let =
f()

> 0, be a constant. Thus,

x
n+1
< x
n

for all n. Thus, x

n
< x
1
(n 1), that is, x
n
as n . It
contradicts. Thus, {x
n
} converges to .
For (c): By using Taylors theorem,
f() = f(x
n
) + f

(x
n
)( x
n
) +
f

(t
n
)
2(x
n
)
2
0 = f(x
n
) + f

(x
n
)( x
n
) +
f

(t
n
)
2(x
n
)
2
0 =
f(x
n
)
f

(x
n
)
x
n
+ +
f

(t
n
)
2f

(x
n
)
(x
n
)
2
61
x
n+1
=
f

(t
n
)
2f

(x
n
)
(x
n
)
2
where t
n
(, x
n
).
For (d): By (b) we know that 0 x
n+1
for all n. Next by (c) we
know that
x
n+1
=
f

(t
n
)
2f

(x
n
)
(x
n
)
2
Note that f

M and f

> 0. Thus
x
n+1
A(x
n
)
2

1
A
(A(x
1
))
2
n
by the induction. Thus,
0 x
n+1

1
A
[A(x
1
)]
2
n
.
For (e): If x
0
is a xed point of g(x), then g(x
0
) = x
0
, that is,
x
0

f(x
0
)
f

(x
0
)
= x
0
f(x
0
) = 0.
It implies that x
0
= and x
0
is unique since f is strictly increasing.
Thus, we choose x
1
(, b) and apply Newtons method, we can nd
out . Hence we can nd out x
0
.
Next, by calculating
g

(x) =
f(x)f

(x)
f

(x)
2
0 g

(x) f(x)
M

2
.
As x near from right hand side, g

(x) near f() = 0.

For (f ): x
n+1
= x
n

f(x
n
)
f

(x
n
)
= 2x
n
by calculating. Thus,
x
n
= (2)
n1
x
1
for all n, thus {x
n
} does not converges for any choice of x
1
, and we
cannot nd such that f() = 0 in this case.
62 CHAPTER 5. DIFFERENTIATION
26. Suppose f is dierentiable on [a, b], f(a) = 0, and there is a real number
A such that |f

(x)| A|f(x)| on [a, b]. Prove that f(x) = 0 for all

x [a, b]. Hint: Fix x
0
[a, b], let
M
0
= sup |f(x)|, M
1
= sup |f

(x)|
for a x x
0
. For any such x,
|f(x)| M
1
(x
0
a) A(x
0
a)M
0
.
Hence M
0
= 0 if A(x
0
a) < 1. That is, f = 0 on [a, x
0
]. Proceed.
Proof: Suppose A > 0. (If not, then f = 0 on [a, b] clearly.) Fix
x
0
[a, b], let
M
0
= sup |f(x)|, M
1
= sup |f

(x)|
for a x x
0
. For any such x,
f(x) f(a) = f

(c)(x a)
where c is between x and a by using Mean-Value Theorem. Thus
|f(x)| M
1
(x a) M
1
(x
0
a) A(x
0
a)M
0
Hence M
0
= 0 if A(x
0
a) < 1. That is, f = 0 on [a, x
0
] by taking
x
0
= a +
1
2A
. Repeat the above argument by replacing a with x
0
, and
note that
1
2A
is a constant. Hence, f = 0 on [a, b].
27. Let be a real function dened on a rectangle R in the plane, given
by a x b, y . A solution of the initial-value problem
y

= (x, y), y(a) = c ( c )

is, by denition, a dierentiable function f on [a, b] such that f(a) = c,
f(x) , and
f

(x) = (x, f(x)) (a x b)

Prove that such a problem has at most one solution if there is a constant
A such that
|(x, y
2
) (x, y
1
)| A|y
2
y
1
|
63
whenever (x, y
1
) R and (x, y
2
) R.
Hint: Apply Exercise 26 to the dierence of two solutions. Note that
this uniqueness theorem does not hold for the initial-value problem
y

= y
1/2
, y(0) = 0,
which has two solutions: f(x) = 0 and f(x) = x
2
/4. Find all other
solutions.
Proof: Suppose y
1
and y
2
are solutions of that problem. Since
|(x, y
2
) (x, y
1
)| A|y
2
y
1
|,
y(a) = c, y

1
= (x, y
1
), and y

2
= (x, y
2
), by Exercise 26 we know that
y
1
y
2
= 0, y
1
= y
2
. Hence, such a problem has at most one solution.
Note: Suppose there is initial-value problem
y

= y
1/2
, y(0) = 0.
If y
1/2
= 0, then y
1/2
dy = dx. By integrating each side and noting that
y(0) = 0, we know that f(x) = x
2
/4. With y
1/2
= 0, or y = 0. All
solutions of that problem are
f(x) = 0 and f(x) = x
2
/4
Why the uniqueness theorem does not hold for this problem? One
reason is that there does not exist a constant A satisfying
|y

1
y

2
| A|y
1
y
2
|
if y
1
and y
2
are solutions of that problem. (since 2/x as x 0
and thus A does not exist).
28. Formulate and prove an analogous uniqueness theorem for systems of
dierential equations of the form
y

j
=
j
(x, y
1
, ..., y
k
), y
j
(a) = c
j
(j = 1, ..., k)
Note that this can be rewritten in the form
y

= (x, y), y(a) = c

64 CHAPTER 5. DIFFERENTIATION
where y = (y
1
, ..., y
k
) ranges over a k-cell, is the mapping of a (k+1)-
cell into the Euclidean k-space whose components are the function

1
, ...,
k
, and c is the vector (c
1
, ..., c
k
). Use Exercise 26, for vector-
valued functions.
Theorem: Let
j
(j = 1, ..., k) be real functions dened on a rectangle
R
j
in the plane given by a x b,
j
y
j

j
.
A solution of the initial-value problem
y

j
= (x, y
j
), y
j
(a) = c
j
(
j
c
j

j
)
is, by denition, a dierentiable function f
j
on [a, b] such that f
j
(a) =
c
j
,
j
f
j
(x)
j
, and
f

j
(x) =
j
(x, f
j
(x)) (a x b)
Then this problem has at most one solution if there is a constant A
such that
|
j
(x, y
j
2
)
j
(x, y
j
1
)| A|y
j
2
y
j
1
|
whenever (x, y
j
1
) R
j
and (x, y
j
2
) R
j
.
Proof: Suppose y
1
and y
2
are solutions of that problem. For each
components of y
1
and y
2
, say y
1j
and y
2j
respectively, y
1j
= y
2j
by
using Exercise 26. Thus, y
1
= y
2
29. Specialize Exercise 28 by considering the system
y

j
= y
j+1
(j = 1, ..., k 1),
y

k
= f(x)
k

j=1
g
j
(x)y
j
where f, g
1
, ..., g
k
are continuous real functions on [a, b], and derive a
uniqueness theorem for solutions of the equation
y
(k)
+ g
k
(x)y
(k1)
+ ... + g
2
(x)y

+ g
1
(x)y = f(x),
65
subject to initial conditions
y(a) = c
1
, y

(a) = c
1
, ..., y
(k1)
(a) = c
k
.
Theorem: Let R
j
be a rectangle in the plain, given by a x b,
min y
j
y
j
max y
j
. (since y
j
is continuous on the compact set, say
[a, b], we know that y
j
attains minimal and maximal.) If there is a
constant A such that
_
|y
j+1,1
y
j+1,2
| A|y
j,1
y
j,2
| (j < k)
|

k
j=1
g
j
(x)(y
j,1
y
j,2
)| A|y
k,1
y
k,2
|
whenever (x, y
j,1
) R
j
and (x, y
j,2
) R
j
.
Proof: Since the system y

1
, ..., y

k
with initial conditions satises a fact
that there is a constant A such that |y

1
y

2
| A|y
1
y
2
|, that system
has at most one solution. Hence,
y
(k)
+ g
k
(x)y
(k1)
+ ... + g
2
(x)y

+ g
1
(x)y = f(x),
with initial conditions has at most one solution.
66 CHAPTER 5. DIFFERENTIATION
Chapter 6
Sequences and Series of
Functions
67
68 CHAPTER 6. SEQUENCES AND SERIES OF FUNCTIONS