Section 4.1

Properties of compact sets

[38/16%]
Show
LaTeX
Problem 4.1

Prove Theorem 4.1.

Theorem 4.1: in any metric space XX, sequential compactness of a set MM is equivalent to compactness of MM.

?
Problem 4.2

Let UU be an open cover of a metric compact (X,ρ)(X, \rho ). Prove that there exists r>0r > 0 such that every ball B(x,r)B(x, r) is contained in some element of the cover UU.

?
Problem 4.3

Prove that in a metric space a compact set is closed and bounded.

?
Problem 4.4

Prove that any subset of a precompact set is precompact.

?
Problem 4.5

Prove that in any finite-dimensional normed space precompactness is equivalent to boundedness.

?
Problem 4.6

Prove that in a discrete metric space, in the space N\mathbb {N}, and in the space Z\mathbb {Z}, a set is precompact if and only if it consists of a finite number of points.

?
Problem 4.7

Prove that if a mapping f:X→Yf: X \rightarrow Y of metric spaces is continuous and XX is compact, then f(X)f(X) is compact in YY.

?
Problem 4.8

Prove that if a set MM is precompact in an incomplete metric space XX, then it is precompact also in the completion of this space1. Give a counterexample to the converse statement.

Definition 1.16: A metric space YY is called a completion of a metric space XX if: (1) YY is complete; (2) there is an isometric embedding f:X→Yf: X \rightarrow Y such that the set f(X)f(X) is everywhere dense in YY.

Footnotes

  1. More precisely, if (Y,f)(Y, f) is a completion of the metric space XX (see Definition 1.16), where f:X→Yf: X \rightarrow Y is an isometry, then f(M)f(M) is precompact in YY. ↩

?
Problem 4.9

Give an example of an incomplete metric space XX and a totally bounded, but not precompact, set MM in it (compare this problem with Theorem 4.2).

Theorem 4.2 (F. Hausdorff, 1914): In any metric space, precompactness of a set MM implies total boundedness of MM. In a complete metric space the converse also holds: every totally bounded set is precompact.

?
Problem 4.10

Prove that if a set MM in a metric space XX is totally bounded, then any sequence of points of the set MM contains a fundamental subsequence.

?
Problem 4.11

Prove that if for some ε>0\varepsilon > 0 a finite ε\varepsilon-net exists for a set MM of a metric space XX, then there exists a finite 2ε2 \varepsilon-net consisting of points of the set MM.

?
Problem 4.12

In the closed unit ball of the space lp,1⩽p⩽∞l_{p}, 1 \leqslant p \leqslant \infty, exhibit a sequence containing no fundamental subsequence.

?
Problem 4.13

Prove that in any infinite-dimensional normed space XX any ball (closed or open) is not precompact.

?
Problem 4.14

Prove that any precompact set in an infinite-dimensional Banach space is nowhere dense.

?
Problem 4.15

Prove that if in a normed space XX a set MM is precompact, then it is bounded and can be approximated with any precision by a finite-dimensional subspace, i.e., for any ε>0\varepsilon > 0 there exists a finite-dimensional subspace X0X_{0} such that dist⁡0(M,X0)<ε\operatorname {dist}_{0}\left(M, X_{0}\right) < \varepsilon (see Definition 1.17). Prove the converse statement for a Banach space XX.

Definition 1.17: Let XX be a metric space, A⊂XA \subset X a nonempty set and x∈Xx \in X. The distance from the point to the set is the number dist⁡(x,A):=inf⁡{ρ(x,y):y∈A}\operatorname {dist}(x, A) := \inf \left\{ \rho (x, y): y \in A\right\}. Here, as in the book's text before this problem, the distance between two sets is denoted dist⁡0(A,B)=inf⁡{ρ(x,y):x∈A,y∈B}\operatorname {dist}_{0}(A, B) = \inf \left\{ \rho (x, y): x \in A, y \in B\right\}.

?
Problem 4.16

Let (X,ρ)(X, \rho ) be a metric space, and let MM be a precompact set in it. Prove that (M,ρ)(M, \rho ) is a separable metric space.

?
Problem 4.17

Let MM be a compact set in a Banach space XX. Prove that there exists a sequence xn→0x_{n} \rightarrow 0 such that M⊂conv⁡{xn}1∞‾M \subset \overline{\operatorname {conv}\left\{ x_{n}\right\}_{1}^{\infty }}.

?
Problem 4.18

Let XX be a metric space, and let MM be a boundedly compact set in it. Prove that for any point x∈Xx \in X the distance from xx to MM is attained, i.e., there exists y∈My \in M such that dist⁡(x,M)=ρ(x,y)\operatorname {dist}(x, M) = \rho (x, y).

?
Problem 4.19

Let XX be a metric space, and let M1M_{1} and M2M_{2} be compact sets in it. Prove that dist⁡0(M1,M2)=ρ(x0,y0)\operatorname {dist}_{0}\left(M_{1}, M_{2}\right) = \rho \left(x_{0}, y_{0}\right) (see Definition 1.17) for some x0∈M1x_{0} \in M_{1} and y0∈M2y_{0} \in M_{2}.

Definition 1.17: Let XX be a metric space, A⊂XA \subset X a nonempty set and x∈Xx \in X. The distance from the point to the set is the number dist⁡(x,A):=inf⁡{ρ(x,y):y∈A}\operatorname {dist}(x, A) := \inf \left\{ \rho (x, y): y \in A\right\}. Here, as in the book's text before problem 4.15, the distance between two sets is denoted dist⁡0(A,B)=inf⁡{ρ(x,y):x∈A,y∈B}\operatorname {dist}_{0}(A, B) = \inf \left\{ \rho (x, y): x \in A, y \in B\right\}.

?
Problem 4.20

Let XX be a metric space, and let MM be a compact set in it. Prove that diam⁡(M)=ρ(x,y)\operatorname {diam}(M) = \rho (x, y) for some x,y∈Mx, y \in M.

?
Problem 4.21

Let XX be a metric compact. Prove that every subset of XX has a Chebyshev center.

?
Problem 4.22

Prove that in a complete metric space any system of closed nonempty nested sets has nonempty intersection, if all these sets are contained in some compact set. 1

Footnotes

  1. Or, in other words, all these sets are compact (see problem 4.4). ↩

?
Problem 4.23

Prove that the function dist⁡H\operatorname {dist}_{\mathrm{H}} satisfies the axioms of a metric.

?
Problem 4.24

Denote by N(X)\mathcal{N}(X) the family of all subsets of a metric space XX consisting of a finite number of points. Prove that if XX is complete, then the space (exp⁡(X),dist⁡H)(\exp (X), \operatorname {dist}_{\mathrm{H}}) is a completion of the space (N(X),dist⁡H)(\mathcal{N}(X), \operatorname {dist}_{\mathrm{H}}).

?
Problem 4.25

Let XX be a metric compact. Prove that the space (exp⁡(X),dist⁡H)(\exp (X), \operatorname {dist}_{\mathrm{H}}) is also a compact.

?
Problem 4.26

Let XX be a complete metric space, and let {fj}1n\left\{ f_{j}\right\}_{1}^{n} be contraction mappings from XX to XX. Define the mapping

F:exp⁡(X)→exp⁡(X),F(A)=⋃j=1nfj(A) F: \exp (X) \rightarrow \exp (X), \quad F(A) = \bigcup _{j = 1}^{n} f_{j}(A)

Prove that FF is a contraction mapping in (exp⁡(X),dist⁡H)\left(\exp (X), \operatorname {dist}_{\mathrm{H}}\right).

?
Problem 4.27

Prove that the following sets are fixed points for certain contraction mappings. Find these mappings.

?
(a)

The Cantor set on the segment [0,1][0,1] is a fixed point in the space (exp⁡([0,1]),dist⁡H)\left(\exp ([0,1]), \operatorname {dist}_{\mathrm{H}}\right).

(b)

The Sierpiński triangle is a fixed point in the space (exp⁡([0,1]2),dist⁡H)\left(\exp \left([0,1]^{2}\right), \operatorname {dist}_{\mathrm{H}}\right).

(c)

The Cantor staircase is a fixed point in the space (exp⁡([0,1]2),dist⁡H)\left(\exp \left([0,1]^{2}\right), \operatorname {dist}_{\mathrm{H}}\right).

(d)

The Cantor function is a fixed point in the set X={x∈C[0,1]:x(0)=0,x(1)=1}X = \left\{ x \in C[0,1]: x(0) = 0, x(1) = 1\right\}.

Problem 4.28

Let MM be a set in a complete metric space XX such that every real-valued function continuous on MM is bounded. Prove that MM is a compact.

?
Problem 4.29

Let MM be a set in a complete metric space XX such that every real-valued function continuous and bounded on MM attains its supremum and infimum. Prove that MM is a compact.

?
Problem 4.30

Let MM be a compact set in a complete metric space X1,f:M→X2X_{1}, f: M \rightarrow X_{2}, where X2X_{2} is also a complete metric space. Prove that the continuity of the mapping ff implies its uniform continuity.

?
Problem 4.31

Let MM be a set in a complete metric space XX such that every real-valued function continuous on MM is uniformly continuous. Prove that MM is a compact.

?
Problem 4.32

Let MM be a compact in a metric space XX, and let YY be a normed space. Denote by C(M,Y)C(M, Y) the set of continuous functions from MM to YY. Prove that the norm ∥x∥C(M,Y)=max⁡t∈M∥x(t)∥Y\left\| x\right\|_{C(M, Y)} = \max_{t \in M}\left\| x(t)\right\|_{Y} turns C(M,Y)C(M, Y) into a normed space (the linear operations are introduced pointwise), and if YY is a Banach space, then the space C(M,Y)C(M, Y) is also a Banach space. 1

Footnotes

  1. An important special case: Y=RY = \mathbb {R} or Y=CY = \mathbb {C} (compare with problem 2.61). ↩

?
Problem 4.33

(Dini, 1872). Let MM be a metric compact, and let the sequence of functions {fn:M→Y}1∞\left\{ f_{n}: M \rightarrow Y\right\}_{1}^{\infty }, continuously mapping MM into a normed space YY, converge to a continuous function ff pointwise, and for every x∈Mx \in M

∥f1(x)−f(x)∥Y⩾∥f2(x)−f(x)∥Y⩾… \left\| f_{1}(x)-f(x)\right\| _{Y} \geqslant \left\| f_{2}(x)-f(x)\right\| _{Y} \geqslant \ldots

Prove that fn→ff_{n} \rightarrow f in the space C(M,Y)C(M, Y). 1

Footnotes

  1. A special case of this theorem: Y=RY = \mathbb {R}, and the sequence {fn(x)}1∞\left\{ f_{n}(x)\right\}_{1}^{\infty } for every x∈Xx \in X increases monotonically (or for every x∈Xx \in X decreases monotonically) to f(x)f(x). ↩

?
Problem 4.34

Prove that if a set MM is compact in a metric space XX, and a mapping f:M→Mf: M \rightarrow M satisfies the inequality ρ(f(x),f(y))<ρ(x,y)\rho (f(x), f(y)) < \rho (x, y) for any x≠yx \neq y, then there exists a unique fixed point x∈Mx \in M (compare with problem 1.74 and Theorem 1.4).

Theorem 1.4 (contraction mapping principle): let f:X→Xf: X \rightarrow X be a contraction mapping (i.e., there exists q∈[0,1)q \in [0, 1) such that ρ(f(x1),f(x2))⩽q⋅ρ(x1,x2)\rho (f(x_{1}), f(x_{2})) \leqslant q \cdot \rho (x_{1}, x_{2}) for all x1,x2∈Xx_{1}, x_{2} \in X) on a complete metric space XX. Then there exists a unique point x0∈Xx_{0} \in X such that f(x0)=x0f(x_{0}) = x_{0}; this point is the limit of the recursively defined sequence xn=f(xn−1)x_{n} = f(x_{n-1}) with an arbitrary starting point x1∈Xx_{1} \in X.

?
Problem 4.35

Prove that in a metric space any isometric embedding of a compact into itself is an isometry of this compact (and show that this is false for an arbitrary closed bounded set).

?
Problem 4.36

Let XX be a complete metric space, MM a compact set in it, φ:M→M\varphi : M \rightarrow M a mapping satisfying the inequality ρ(φ(x),φ(y))⩾ρ(x,y)\rho (\varphi (x), \varphi (y)) \geqslant \rho (x, y). Prove that φ\varphi is an isometry of the set MM.

?
Problem 4.37

Find the approximative dimension:

?
(a)

of the segment [0,1][0,1];

(b)

of the square [0,1]×[0,1][0,1] \times [0,1].

Problem 4.38

Find the approximative dimension:

?
(a)

of the Cantor set in the space X=[0,1]X = [0,1];

(b)

of the Sierpiński triangle in R2\mathbb {R}^{2}.