0% found this document useful (0 votes)
43 views7 pages

Exercises 2.3

1. The document provides solutions to exercises related to limits, compact sets, and continuous functions. 2. It shows that the union of a bounded set and its limit points is compact, and that a compact set of real numbers contains its infimum and supremum. 3. It also proves that if a set S is compact and z0 is not in S, then the infimum of the distances from z0 to points in S is positive.

Uploaded by

이퓨처
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
43 views7 pages

Exercises 2.3

1. The document provides solutions to exercises related to limits, compact sets, and continuous functions. 2. It shows that the union of a bounded set and its limit points is compact, and that a compact set of real numbers contains its infimum and supremum. 3. It also proves that if a set S is compact and z0 is not in S, then the infimum of the distances from z0 to points in S is positive.

Uploaded by

이퓨처
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Exercises 2.3 (#5 may be skipped.

)
1. Show that the union of any bounded set and its limit points is a compact set.
Solution. A = A ∪ A0 is closed. If A is bounded then A is also bounded.
(∵ Since A is bounded, there exists a constant R > 0 such that A ⊂ N (0; R) = {z | |z| ≤ R}.
Since N (0; R) is closed and A is the smallest closed set containing A (#9 in Exercises 2.2),
it follows that A ⊂ {z | |z| ≤ R}. Thus A is bounded.)
A = A ∪ A0 is closed and bounded, and hence compact by Theorem 2.11. ¤

2. Show that a compact set of real numbers contains its infimum and its supremum. Can this
occur for a set of real numbers that is not compact?
Solution. Let A ⊂ R be a compact set (A 6= ∅). Since A is bounded, both sup A(∈ R) and
inf A(∈ R) exist by the completeness axiom.
We argue by contradiction and suppose that sup A ∈/ A. Given ² > 0, there exists an element
y ∈ A such that y > sup A − ². Since sup A ∈ / A, it follows that sup A − ² < y < sup A.
Therefore sup A is a limit point of A. Since sup A ∈/ A, A is not closed, which leads to a
contradiction. Thus sup A ∈ A.
Similarly we can prove that inf A ∈ A. The details are left to the readers. ¤

3. If S(6= ∅) is compact and z0 ∈


/ S, prove that inf |z − z0 | > 0.
z∈S
Solution. Since S is closed and bounded, S c is a nonempty open set. Then there exists a
number ² > 0 such that N (z0 ; ²) ⊂ S c . Consequently |z − z0 | ≥ ² for any z ∈ S, which
implies that ² is a lower bound of {|z − z0 | | z ∈ S} ⊂ R. Therefore inf |z − z0 | > 0. ¤
z∈S
Remark. The conclusion is still true when S is nonempty and closed.

4. If {Sn } is a sequence of nonempty compact sets with Sn+1 ⊂ Sn for every n, show that
T∞
n=1 Sn 6= ∅.
T∞
Solution. We argue by contradiction, and suppose that n=1 Sn = ∅, in other words,
¡ T∞ ¢ ¡ T∞ ¢c S∞
S1 ∩ n=2 Sn = ∅. Then it follows that S1 ⊂ n=2 Sn = n=2 Snc . Consequently,
{Snc }∞ c ∞
n=2 is an open cover of S1 . Since S1 is compact, {Sn }n=2 has a finite subcover, say,
Sm c
S1 ⊂ k=1 Snk (n1 < n2 < · · · < nm ). Since Snm ⊂ · · · ⊂ Sn1 ⊂ S1 ,
³\
m ´
Snm = S1 ∩ Snk = ∅,
k=1

which leads to a contradiction. ¤

5. In Theorem 2.12, prove as many different implications you can.

1
Solution. (i) ⇒ (iii): Let A be an infinite subset of S. Since S is bounded, Bolzano-
Weierstrass theorem implies that A has a limit point, say, z0 . Since A ⊂ S, it follows that
z0 is a limit point of S. Since S is closed, z0 ∈ S. ////
(i) ⇒ (iv): similar to the proof of (i) ⇒ (iii). ////
(ii) ⇒ (i): We claim that S is closed if it is compact. For this purpose, we show that S c is
open.
Choose a point z0 ∈ S c . For any q ∈ S, we let rq = 12 |q − z0 |. Then {N (q; rq )}q∈S is an
open cover of S, because q ∈ N (q; rq ) for all q ∈ S. Since S is compact, there exists a finite
subcover of {N (q; rq )}q∈S , namely, there exist q1 , . . . , qm ∈ S and rq1 , . . . , rqm > 0 such
that
[m
1
S⊂ N (qk ; rqk ), rqk = |qk − z0 |.
2
k=1
m
[
Let ² = min{rq1 , . . . , rqm }. Then it follows that ² > 0 and N (z0 ; ²) ∩ N (qk ; rqk ) = ∅.
k=1
Consequently N (z0 ; ²) ⊂ S c .
Since S is closed, S = S. In problem 1, we showed that S is bounded if S is. ////
(ii) ⇒ (iv): similar to the proof of Theorem 2.12. ((ii) ⇒ (iii))
(iii) ⇒ (i): If S is not bounded, we can choose a sequence {zn } such that |zn | ≥ n for all
n. Then every subsequence of {zn } diverges, which contradicts (iii). ////
Let q be a limit point of S (if S 0 6= ∅). Theorem 2.8 shows that there exists a sequence
{zn } ⊂ S such that zn → q and zn 6= q. Then {zn | n ∈ N} is infinite, and every subsequence
of {zn } converges to z0 . In other words, z0 is the only limit point of {zn | n ∈ N}. (iii)
implies that z0 ∈ S, and S is closed. ////
(iii) ⇒ (ii): similar to the proof of (iv) ⇒ (ii). Indeed, we can extract a sequence of distinct
points in an infinite set.
(iv) ⇒ (ii): Note that S is bounded (see the proof of (iii) ⇒ (i)).
We can modify the proof of Theorem 2.11: We argue by contradiction and suppose that S
is not compact. Then there exists an open cover {Oα } of S such that no finite subcover
contains S. Then we can choose a sequence {Rn } of rectangles such that
(a) Rn = [an , bn ] × [cn , dn ] (an < bn , cn < dn ),
(b) S ⊂ R1 and Rn+1 ⊂ Rn ,
(c) bn − an = (b1 − a1 )/2n−1 and dn − cn = (d1 − c1 )/2n−1 ,
(d) No finite subcollection of {Oα } covers S ∩ Rn for any n.
In particular S ∩ Rn is nonempty. Then we can choose a sequence {zn } of points such that
zn ∈ S ∩ Rn . By (iv), {zn } has a subsequence {znk } which converges to a point p ∈ S.
Then p ∈ Oα1 for some α1 . Note that S ∩ Rn shrinks to a point p. Then S ∩ Rn ⊂ Oα1 for
n sufficiently large, which yields a contradiction. ////
(iv) ⇒ (iii): Let A be an infinite subset of S. Choose a sequence {zn } of distinct points
in A. Then we can extract a subsequence which converges to z0 ∈ S. Clearly, z0 ∈ S is a
limit point of A. ////

2
Exercises 2.5 (#1, #2, #3, #10, #11, #12, #13: You may skip #4 ∼ #9)
1. Find the following limits.
z2 + 9
(a) Note that z 2 + 9 = (z + 3i)(z − 3i). Then lim = lim (z + 3i) = 6i.
z→3i z − 3i z→3i

z + z2 −2i + (2i)2 −8 + 6i
(b) lim = = .
z→2i 1 − z 1 − (−2i) 5
z+1 ³1 1´
(c) lim = lim + = 0.
z→∞ z 2 z→∞ z z2
z 2 + 10z + 2 1 + 10 2
z + z2 1
(d) lim = lim = .
z→∞ 2z 2 − 11z − 6 z→∞ 2 − 11 − 62 2
z z

2. Discuss continuity and uniform continuity for the following functions.

(a) f (z) = z 2 , |z| < 5


f is uniformly continuous, and hence continuous. (∵ Given ² > 0, we let δ = ²/10.
Then |z1 − z2 | < δ implies

|f (z1 ) − f (z2 )| = |z1 + z2 ||z1 − z2 | < δ|z1 + z2 | ≤ δ(|z1 | + |z2 |) < 10δ = ².

(b) f (z) = 1/z, |z| ≥ 1


Since g(z) = z is continuous and nonzero for |z| ≥ 1, f is continuous. Moreover f is
uniformly continuous. (∵ For any ² > 0, we let δ = ². Then |z1 − z2 | < δ implies
¯1 1 ¯¯ |z1 − z2 |
¯
|f (z1 ) − f (z2 )| = ¯ − ¯ = ≤ |z1 − z2 | < δ = ².
z1 z2 |z1 ||z2 |

1
(c) f (z) = , |z| < 1
1−z
Since g(z) = 1 − z is continuous and nonzero in |z| < 1, f is continuous. However f
is not uniformly continuous. (∵ Let xn = (n − 1)/n and yn = (2n − 1)/(2n). Then
|xn |, |yn | < 1 and |xn − yn | = 1/(2n) → 0 while |(f /g)(xn ) − (f /g)(yn )| = n → ∞.)
(
(Re z)/z, z 6= 0, |z| < 1,
(d) f (z) =
1, z = 0.
Since f (i/n) = 0 6= f (0), f is not continuous. f is not uniformly continuous.
(
|z|/z, z 6= 0, |z| ≤ 1,
(e) f (z) =
0, z = 0.
Since f (1/n) = 1 6= f (0), f is not continuous. f is not uniformly continuous.

3. If lim f (z) = a and f (z) is defined for every positive integer n, prove that lim f (n) = a.
z→∞ n→∞
Give an example to show that the converse is false.

3
Solution. (i) Given ² > 0 there exists an M > 0 such that |f (z) − a| < ² for |z| ≥ M . In
particular, if n ≥ M then |f (n) − a| < ². ////
(ii) Counterexample. Let f (z) = sin(π|z|). Then f (n) = 0 for all n, while f (z) diverges as
|z| → ∞. ¤

4. Show that a monotonic real-valued function of a real variable cannot have uncountably many
discontinuities.
Solution. Without loss of generality, we may assume that f is monotonically increasing on
an open set A ⊂ R. For p ∈ A, we let f (p−) = lim− f (x) and f (p+) = lim+ f (x).
x→p x→p

We claim that for every p ∈ A, there holds

f (p−) ≤ f (p) ≤ f (p+).

Indeed, we note that the set B = {f (x) | x < p} is nonempty, and f (p) is an upper bound
of B. Then it follows that sup B ≤ f (p). Moreover, for every ² > 0, there exists x0 < p such
that f (x0 ) > sup B − ². Let δ = p − x0 . Then it is easily checked that |f (x) − sup B| < ²
whenever −δ < x − p < 0. Therefore f (p−) = sup B ≤ f (p).
Similarly we can prove that f (p+) = inf{f (x) | x > p} ≥ f (p). Our claim is proved.
Therefore, f is discontinuous at p ∈ A if and only if f (p−) < f (p+). In this case, we can
choose a rational number rp ∈ (f (p−), f (p+)).
Let D be the set of discontinuities of f . Define a function ϕ : D → Q by ϕ(p) = rp . Then
ϕ is well defined and one-to-one, and consequently, it is a bijection from D onto a subset
of Q. Since Q is countable, D is also (at most) countable. ¤

Remark. Let O be a subset of E. We say that O is open relative to E if O = U ∩ E for some


open set U . It is obvious that O is open if and only if O is open relative to C.
Lemma. O is open relative to E ⇔ for every p ∈ O there exists an ² > 0 such that
N (p; ²) ∩ E ⊂ O.

5. Show that f : A → B is continuous if and only if for every open set, O relative to B (see
Remark 2, p.48), f −1 (O) is an open set relative to A.
Remark. Note that f −1 (A ∩ B) = f −1 (A) ∩ f −1 (B).
Solution. (⇒) Let O = U ∩ B. For any p ∈ f −1 (O), it follows that f (p) ∈ O ⊂ U . Then
there is an ² = ²(p) > 0 such that N (f (p); ²) ⊂ U . Since f is continuous, we can choose a
δ > 0 such that |z−p| < δ implies |f (z)−f (p)| < ². Therefore, f (N (p; δ)) ⊂ N (f (p); ²) ⊂ U ,
and hence
N (p; δ) ∩ A ⊂ f −1 (U ) ∩ f −1 (B) = f −1 (U ∩ B) = f −1 (O).

(⇐) Choose any p ∈ A. Let ² > 0 be given. Note that N (f (p); ²) ∩ B is open relative to
B. By assumption, f −1 (N (f (p); ²) ∩ B) = U ∩ A for some open set U . Then there exists a
δ > 0 such that N (p; δ) ⊂ U , in other words, N (p; δ) ∩ A ⊂ U ∩ A = f −1 (N (f (p); ²) ∩ B).

4
The last inclusion shows that |z − p| < δ (z ∈ A) implies |f (z) − f (p)| < ². ¤

6. Prove that the continuous image of a compact subset is compact.


Remark. f (f −1 (A)) ⊂ A.
Solution. Let K be a compact subset of C, and f : K → C be continuous. Let {Oα } be an
open cover of f (K). Since f is continuous, f −1 (Oα ) is open relative to K (see #5). For any
S S
α, f −1 (Oα ) = Uα ∩ K for some open set Uα . Therefore, K ⊂ α f −1 (Oα ) ⊂ α Uα . Since
{Uα } is an open cover of a compact set K, there exists finitely many indices α1 , . . . , αn
such that K ⊂ Uα1 ∪ · · · ∪ Uαn , namely,

K ⊂ (Uα1 ∩ K) ∪ · · · ∪ (Uαn ∩ K) = f −1 (Oα1 ) ∪ · · · ∪ f −1 (Oαn ) = f −1 (Oα1 ∪ · · · ∪ Oαn ).

Then
¡ ¢
f (K) ⊂ f f −1 (Oα1 ∪ · · · ∪ Oαn ) ⊂ Oα1 ∪ · · · ∪ Oαn .
Thus {Oα } has a finite subcover, and f (K) is compact. ¤

7. Show that f : A → B is continuous if and only if for every closed set F relative B, f −1 (F )
is a closed set relative to A.
Remark 1. We say F (⊂ A) is closed relative to A if F = K ∩ A for some closed set K.
Remark 2. f −1 (A ∩ B) = f −1 (A) ∩ f −1 (B) and f −1 (A ∪ B) = f −1 (A) ∪ f −1 (B). Moreover,
f −1 (Ac ) = [f −1 (A)]c .
Solution. We use #5 and the above remarks.
(⇒) For every set F closed relative to B, F = Oc ∩ B for some open set O. (Recall that A
is open ⇔ Ac is open.) Then it follows that

f −1 (F ) = f −1 (Oc ∩ B) = f −1 (Oc ) ∩ f −1 (B) = [f −1 (O)]c ∩ A = [U ∩ A]c ∩ A = U c ∩ A

for some open set U . Thus f −1 (F ) is closed relative to A.


(⇐) Let O be a set open relative to B. Then O = F c ∩ B for some closed set F . Then it
follows that

f −1 (O) = f −1 (F c ∩ B) = f −1 (F c ) ∩ f −1 (B) = [f −1 (F )]c ∩ A = [K ∩ A]c ∩ A = K c ∩ A

for some closed set K. Thus f −1 (O) is open relative to A. By #5, f is continuous. ¤

8. Prove that the continuous image of a connected set is connected.


Solution. Suppose that f (A) is not connected, i.e., there exist two disjoint open subsets U
and V such that
f (A) ⊂ U ∪ V, U ∩ f (A) 6= ∅, V ∩ f (A) 6= ∅.
Then we obtain that

A ⊂ f −1 (U ∪ V ) = f −1 (U ) ∪ f −1 (V ), A ∩ f −1 (U ) 6= ∅, A ∩ f −1 (V ) 6= ∅.

5
Since f is continuous, f −1 (U ) = O1 ∩ A and f −1 (V ) = O2 ∩ A for some open sets O1 and
O2 . Therefore,
A ⊂ O1 ∪ O2 , A ∩ O1 6= ∅, A ∩ O2 6= ∅. (1)
On the other hand, note that U ∩ V = ∅, and hence f −1 (U ) ∩ f −1 (V ) = f −1 (U ∩ V ) = ∅.
Then, by choosing new open sets satisfying (1) if necessary, we may assume that O1 ∩O2 = ∅
(why?). Therefore A is not connected. ¤

9. If a function, defined on a compact set, is continuous, one-to-one, and onto, show that the
inverse function also has these properties. Can compactness be omitted?
Solution. (i) If A is compact and f : A → B is bijective, the inverse function g : B → A
exists and it is bijective. Moreover, the domain B = f (A) of g is compact. It remains to
show that g is continuous.
Let K ⊂ A be any set closed relative to A. Since A is compact, K = K ∩ A is compact.
Then g −1 (K) = {z ∈ B | g(z) ∈ K} = f (K) is compact, and hence closed. Then Problem
# 7 implies that g is continuous. ¤

10. Let f and g be continuous on a set A. Show that f + g, f g, and f /g (g 6= 0) are also
continuous on A. What can we say if f and g are uniformly continuous on A?
Solution. (Continuity) Show that the following limits: For any p ∈ A,
(a) lim (f (z) + g(z)) = f (p) + g(p),
z→p

(b) lim f (z)g(z) = f (p)g(p),


z→p

(c) lim (f (z)/g(z)) = f (p)/g(p).


z→p

Proof of (b): Let ² > 0 be given.n There exist twoonumbers δ1 , δ2 > 0 such that |z − p| <
²
δ1 implies |f (z) − f (p)| < min , 1 , and moreover, |z − p| < δ2 implies
2(|g(p)| + 1)
²
|g(z) − g(p)| < .
2(|f (p)| + 1)
Let δ = min{δ1 , δ2 }. Then |z − p| < δ implies |f (z)| ≤ |f (p)| + 1, and

|(f g)(z) − (f g)(p)| = |f (z)g(z) − f (z)g(p) + f (z)g(p) − f (p)g(p)|


≤ |f (z)||g(z) − g(p)| + |g(p)||f (z) − f (p)| < ²/2 + ²/2 = ².

////
(Uniform continuity) f +g is uniformly continuous. Indeed, given ² > 0, there exist δ1 , δ2 > 0
such that |x−y| < δ1 implies |f (x)−f (y)| < ²/2, and |x−y| < δ2 implies |g(x)−g(y)| < ²/2.
Thus if |x − y| < δ = min{δ1 , δ2 } then

|(f + g)(x) − (f + g)(y)| ≤ |f (x) − f (y)| + |g(x) − g(y)| < ².

However, when f and g are uniformly continuous, f g and f /g may not be uniformly
continuous. For example, if we let f (z) = z = g(z) then both f and g are uniformly
continuous, but f g is not.

6
(∵ Let xn = n and yn = n + (1/n). Then |xn − yn | = 1/n → 0 while |(f g)(xn ) − (f g)(yn )| =
2 + (1/n2 ) ≥ 2.)
Moreover, let A = {z | |z| < 1} and f (z) = 1 and g(z) = z − 1. Then f /g is not uniformly
continuous (see #2-(c)).
Remark. If A is compact then the above three functions are uniformly continuous if both f and
g are.

11. Show that f (z) is continuous in a region R if and only if both Re f (z) and Im f (z) are
continuous in R.
Solution. For any p ∈ R, there holds

|Re f (z) − Re f (p)|, |Im f (z) − Im f (p)| ≤ |f (z) − f (p)|


≤ |Re f (z) − Re f (p)| + |Im f (z) − Im f (p)|.

(⇒) Given ² > 0 there exists a number δ > 0 such that |x − p| < δ implies |f (z) − f (p)| < ².
Therefore |Re f (z) − Re f (p)|, |Im f (z) − Im f (p)| < ² for all |x − p| < δ.
(⇐) Let ² > 0 be given. There exist δ1 , δ2 > 0 such that |Re f (z) − Re f (p)| < ²/2 for
|z − p| < δ1 , and |Im f (z) − Im f (p)| < ²/2 for |z − p| < δ2 . Let δ = min{δ1 , δ2 }. Then
|z − p| < δ implies |f (z) − f (p)| < ². ¤

12. Show that every polynomial is continuous in the complex plane.


Solution. For a constant a ∈ C, the function f and g defined by f (z) = z and g(z) = a
is continuous. Then g(z)[f (z)]k = az k is also continuous (k ∈ N). Every polynomial is a
finite sum of the continuous function of the form h(z) = ak z k (ak is a constant), and hence
continuous. ¤

13. Let f (z) be continuous in the complex plane. Let A = {z ∈ C | f (z) = 0}. Show that A is
a closed set.
Solution. Let p be a limit point of A (if A0 6= ∅). Then there exists a sequence {zn } ⊂ A
such that zn → p and zn 6= p. Since f is continuous, f (zn ) → f (p): For every ² > 0 there
exists an N such that n ≥ N implies

|f (p)| = |f (p) − f (zn )| < ².

Therefore f (p) = 0, namely, p ∈ A. ¤

You might also like