• 沒有找到結果。

Advanced Calculus (II)

N/A
N/A
Protected

Academic year: 2022

Share "Advanced Calculus (II)"

Copied!
45
0
0

加載中.... (立即查看全文)

全文

(1)

Advanced Calculus (II)

WEN-CHINGLIEN

Department of Mathematics National Cheng Kung University

2009

(2)

Ch10: Metric Spaces

10.4: Compact Sets

Definition (10.41)

Let V = {Vα}α∈A be a collection of subsets of a metric space X and suppose that E is a subset of X .

(i) V is said to cover E (or be a covering of E ) if and only if E ⊆ [

α∈A

Vα.

(ii) V is said to be an open covering of E if and only if V covers E and each Vα is open.

(iii) Let V be a covering of E . V is said to have a finite (respectively, countable) subcovering if and only if there is

(3)

Definition (10.42)

A subset H of a metric space X is said to be compact if and only if every open covering of H has a finite subcover.

(4)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(5)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed.Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(6)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 .Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(7)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0;hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(8)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 .Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(9)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0;hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(10)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact,we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(11)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(12)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact,we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(13)

Remark (10.44)

A compact set is always closed.

Proof.

Suppose that H is compact but not closed. Then H is nonempty and (by Theorem 10.16) there is a convergent sequence xk ∈ H whose limit x does not belong to H. For each y ∈ H, set r (y ) := ρ(x ,y )2 . Since x does not belong to H, r (y ) > 0; hence, each Br (y )r (y ) is open and contains y ; i.e., {Br (y )r (y ) : y ∈ H} is an open covering of H. Since H is compact, we can choose points yj and radii is rj :=r (yj) such that {Brj(yj) :j = 1, 2, . . . , N} covers H.

(14)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.)Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N.Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(15)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k .But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r ,and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(16)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N.Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj)≥ ρ(x, yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(17)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r ,and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(18)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj)≥ ρ(x, yj) − ρ(xk,x )

=2rj− ρ(xk,x )>2rj− r ≥ 2rj − rj =rj

a contradiction.

(19)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(20)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x )>2rj− r ≥ 2rj − rj =rj

a contradiction.

(21)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(22)

Proof.

Set r := min{r1, . . . ,rN}. (This is a finite set of positive numbers, so r is also positive.) Since xk → x as

k → ∞, xk ∈ Br(x ) for large k . But xk ∈ Br(x ) ∩ H implies xk ∈ Brj(yj)for some j ∈N. Therefore, it follows from the choices of rj and r , and from the triangle inequality, that

rj ≥ ρ(xk,yj) ≥ ρ(x , yj) − ρ(xk,x )

=2rj− ρ(xk,x ) > 2rj− r ≥ 2rj − rj =rj

a contradiction.

(23)

Remark (10.45)

A closed subset of H of a compact set is compact.

(24)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded.Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(25)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact.By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X .Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(26)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded.Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact,it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(27)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X .Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N.Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(28)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact,it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(29)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N.Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(30)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(31)

Theorem (10.46)

Let H be a subset of metric space X . If H is compact, then H is closed and bounded.

Proof.

Suppose that H is compact. By Remark 10.44, H is closed. It is also bounded. Indeed, fix b ∈ X and observe that {Bn(b) : n ∈N} covers X . Since H is compact, it follows that

H ⊂

N

[

n=1

Bn(b)

for some N ∈N. Since these balls are nested, we conclude that H ⊂ BN(b); i.e., H is bounded.

(32)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R,every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(33)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3.Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded.Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(34)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R,every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed.Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(35)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded.Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X .Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(36)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed.Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover,we conclude that [0, 1] is closed and bounded, but not compact.

(37)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X .Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(38)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover,we conclude that [0, 1] is closed and bounded, but not compact.

(39)

Remark (10.47)

The converse of Theorem 10.46 is false for arbitrary metric spaces.

Proof.

Let X =R be the discrete metric space introduced in Example 10.3. Since σ(0, x ) ≤ 1 for all x ∈R, every subset of X is bounded. Since xk → x in X implies xk → x for large k , every subset of X is closed. Thus [0, 1] is a closed bounded subset of X . Since {x }x ∈[0,1] is an uncountable open covering of [0, 1] that has no proper finite subcover, we conclude that [0, 1] is closed and bounded, but not compact.

(40)

Definition (10.48)

A metric space X is said to be separable if and only if it contains a countable dense subset; i.e., there is a

countable set Z of X such that for every point a ∈ X there is a sequence xk ∈ Z such that xk → a as x → ∞.

(41)

Theorem (10.49 Lindel ¨of)

Let E be a subset of separable metric space X . If {Vα}α is a collection of open sets and E ⊆ ∪α∈AVα, then there is a countable subset A0of A such that

E ⊆ [

α∈A0

Vα.

(42)

Theorem (10.50 Heine-Borel)

Let X be a separable metric space that satisfies the Bolzano-Weierstrass Property and H be a subset of X . Then H is compact if and only if it is closed and bounded.

(43)

Definition (10.51)

Let X be a metric space, E be a nonempty subset of X , and f : E → Y . Then f is said to be uniformly continuous on E (notation: f : E → Y is uniformly continuous) if and only if given ε > 0 there is a δ > 0 such that

ρ(x , a) < δ and x , a ∈ E imply τ (f (x ), f (a)) < ε.

(44)

Theorem (10.52)

Suppose that E is a compact subset of X and f : X → Y . Then F is uniformly continuous on E if and only if f is continuous on E .

(45)

Thank you.

參考文獻

相關文件

Department of Mathematics National Cheng Kung

Department of Mathematics National Cheng Kung

On the other hand, by (22), the triangle inequality, and the definition of the operator norm... On the other hand, by (22), the triangle inequality, and the definition of the

Department of Mathematics National Cheng Kung

Department of Mathematics National Cheng Kung

Department of Mathematics National Cheng Kung

Since ρ is obviously symmetric, it remains to verify the triangle

Department of Mathematics National Cheng Kung