Chapter 2

Hilbert Spaces

[64/83%]
Show
LaTeX
Problem 2.1

Prove that a scalar product in a Hilbert space is a continuous function in each of the arguments.

?
Problem 2.2

Let LL and MM be linear sets in a Hilbert space HH, and L⊥ML \perp M.

?
(1)

Prove that for all x∈L⨁Mx \in L \bigoplus M the representation x=x′+x′′,x′∈L,x′′∈Mx = x^{\prime }+x^{\prime \prime }, x^{\prime } \in L, x^{\prime \prime } \in M, is unique.

(2)

Let H=L⨁MH = L \bigoplus M. Prove that M⊥=LM^{\perp } = L.

Problem 2.3

Find the orthogonal complement in L2([−1,1])L_{2}(\left[-1,1\right]) of the set of all even functions.

?
Remark.
?

A function x∈L2([−1,1])x \in L_{2}(\left[-1,1\right]) is said to be even if x(t)=x(−t)x(t) = x(-t) for almost all t∈[−1,1]t \in [-1,1]. The notion of an odd function is introduced in a similar way.

Problem 2.4

Let LL be a total set in a Hilbert space HH. Prove that L⊥={0}L^{\perp } = \left\{ 0\right\}.

?
Problem 2.5

Find the orthogonal complement of the set M={ekt∣k≥10}M = \left\{ e^{k t} \mid k \geq 10\right\} in L2([0,1])L_{2}(\left[0,1\right]).

?
Problem 2.6

Let α=(α1,α2,…)\alpha = \left(\alpha_{1}, \alpha_{2}, \ldots \right), where αk>0,k≥1\alpha_{k} > 0, k \geq 1. Consider the set l2,αl_{2, \alpha } of all numerical sequences x=(x1,x2,…)x = \left(x_{1}, x_{2}, \ldots \right) for which ∑k=1∞αk∣xk∣2<∞\sum_{k = 1}^{\infty } \alpha_{k}\left|x_{k}\right|^{2} < \infty. Verify that l2,αl_{2, \alpha } is a Hilbert space with the scalar product (x,y)=∑k=1∞αkxkyk‾,x,y∈l2,α(x, y) = \sum_{k = 1}^{\infty } \alpha_{k} x_{k} \overline{y_{k}}, x, y \in l_{2, \alpha }.

Construct an orthonormal basis in the space l2,αl_{2, \alpha } in the following cases:

?
(1)

αk=k,k≥1\alpha_{k} = k, k \geq 1;

(2)

αk=k2,k≥1\alpha_{k} = k^{2}, k \geq 1;

(3)

αk=e−k,k≥1\alpha_{k} = e^{-k}, k \geq 1.

Problem 2.7

Determine for which α=(α1,α2,…)\alpha = \left(\alpha_{1}, \alpha_{2}, \ldots \right), where αk>0,k≥1\alpha_{k} > 0, k \geq 1, the inclusion l2,α⊂l2l_{2, \alpha } \subset l_{2} holds.

?
Problem 2.8
?
(1)

Let HH be a Hilbert space, ∥⋅∥\left\| \cdot \right\| be a norm in HH, generated by a scalar product. Check that

∀x,y∈H:∥x+y∥2+∥x−y∥2=2(∥x∥2+∥y∥2) \forall x, y \in H:\left\| x+y\right\| ^{2}+\left\| x-y\right\| ^{2} = 2\left(\left\| x\right\| ^{2}+\left\| y\right\| ^{2}\right)

(parallelogram law).

(2)

Prove that in the Banach space C([0,1])C(\left[0,1\right]) the norm is not generated by a scalar product.

Problem 2.9

Prove that in the Banach spaces c0,lp,Lp([a,b]),p≠2c_{0}, l_{p}, L_{p}(\left[a, b\right]), p \neq 2, the norm is not generated by a scalar product.

?
Problem 2.10

Prove that for arbitrary x,y,zx, y, z in a Hilbert space Apollonius identity holds:

∥x−z∥2+∥y−z∥2=12∥x−y∥2+2∥z−x+y2∥2. \left\| x-z\right\| ^{2}+\left\| y-z\right\| ^{2} = \frac{1}{2}\left\| x-y\right\| ^{2}+2\left\| z-\frac{x+y}{2}\right\| ^{2}.
?
Problem 2.11

Prove that for arbitrary x,y,z,ux, y, z, u in a Hilbert space Ptolemy's inequality holds:

∥x−z∥⋅∥y−u∥≤∥x−y∥⋅∥z−u∥+∥y−z∥⋅∥x−u∥. \left\| x-z\right\| \cdot \left\| y-u\right\| \leq \left\| x-y\right\| \cdot \left\| z-u\right\| +\left\| y-z\right\| \cdot \left\| x-u\right\| .
?
Problem 2.12

Prove that in a Hilbert space over a field K\mathbb {K} the elements xx and yy are orthogonal if and only if

?
(1)

∥x+y∥2=∥x∥2+∥y∥2\left\| x+y\right\|^{2} = \left\| x\right\|^{2}+\left\| y\right\|^{2} in the case of K=R\mathbb {K} = \mathbb {R};

(2)

∥λx+μy∥2=∥λx∥2+∥μy∥2\left\| \lambda x+\mu y\right\|^{2} = \left\| \lambda x\right\|^{2}+\left\| \mu y\right\|^{2} for any λ,μ∈C\lambda , \mu \in \mathbb {C} in the case of K=C\mathbb {K} = \mathbb {C}.

Problem 2.13

Prove that the polarization identity holds in a complex Hilbert space:

(x,y)=14(∥x+y∥2−∥x−y∥2+i∥x+iy∥2−i∥x−iy∥2). (x, y) = \frac{1}{4}\left(\left\| x+y\right\| ^{2}-\left\| x-y\right\| ^{2}+i\left\| x+i y\right\| ^{2}-i\left\| x-i y\right\| ^{2}\right).

What does it look like in a real Hilbert space?

?
Problem 2.14

Let XX be a complex NLS, and let the function S:X×X→CS: X \times X \rightarrow \mathbb {C} has the following properties:

?
(1)

∀x∈X:S(x,x)≥0\forall x \in X: S(x, x) \geq 0, and S(x,x)=0⇔x=0S(x, x) = 0 \Leftrightarrow x = 0;

(2)

∀x,y∈X:S(y,x)=S(x,y)‾\forall x, y \in X: S(y, x) = \overline{S(x, y)};

(3)

∀x1,x2,y∈X:S(x1+x2,y)=S(x1,y)+S(x2,y)\forall x_{1}, x_{2}, y \in X: S\left(x_{1}+x_{2}, y\right) = S\left(x_{1}, y\right)+S\left(x_{2}, y\right);

(4)

∀x,y∈X:S(ix,y)=iS(x,y)\forall x, y \in X: S(i x, y) = i S(x, y);

(5)

∀{xn:n≥1}⊂X,y∈X,xn→x,n→∞:S(xn,y)→S(x,y),n→∞\forall \left\{ x_{n}: n \geq 1\right\} \subset X, y \in X, x_{n} \rightarrow x, n \rightarrow \infty : S\left(x_{n}, y\right) \rightarrow S(x, y), n \rightarrow \infty.

Prove that S(⋅,⋅)S(\cdot , \cdot ) is a scalar product in the space XX. Is XX necessarily a Hilbert space with this scalar product if XX is a Banach space?

Problem 2.15

Let XX be a complex NLS in which the norm satisfies the parallelogram law ∥x+y∥2+∥x−y∥2=2(∥x∥2+∥y∥2)\left\| x+y\right\|^{2}+\left\| x-y\right\|^{2} = 2\left(\left\| x\right\|^{2}+\left\| y\right\|^{2}\right) for all x,y∈Xx, y \in X. Prove that the formula

S(x,y)=14(∥x+y∥2−∥x−y∥2+i∥x+iy∥2−i∥x−iy∥2), S(x, y) = \frac{1}{4}\left(\left\| x+y\right\| ^{2}-\left\| x-y\right\| ^{2}+i\left\| x+i y\right\| ^{2}-i\left\| x-i y\right\| ^{2}\right),

x,y∈Xx, y \in X, defines a scalar product in the space XX, and ∥⋅∥\left\| \cdot \right\| is the norm generated by this scalar product.

?
Problem 2.16

Let pp be a Lebesgue-measurable function on (a,b)(a, b) such that p(t)>0p(t) > 0 for almost all t∈(a,b)t \in (a, b), and L2,p([a,b])={x∣ p12x∈L2([a,b])}L_{2, p}(\left[a, b\right]) = \left\{ x \left\lvert \, p^{\frac{1}{2}} x \in L_{2}(\left[a, b\right])\right.\right\}. Prove that L2,p([a,b])L_{2, p}(\left[a, b\right]) is a Hilbert space with scalar product (x,y)=∫abx(t)y(t)‾p(t)dt(x, y) = \int_{a}^{b} x(t) \overline{y(t)} p(t) d t. For which pp the inclusion L2,p([a,b])⊂L2([a,b])L_{2, p}(\left[a, b\right]) \subset L_{2}(\left[a, b\right]) holds?

?
Problem 2.17

Let {xk:k≥1}\left\{ x_{k}: k \geq 1\right\} be an orthogonal system in a Hilbert space HH. Prove that the series ∑k=1∞xk\sum_{k = 1}^{\infty } x_{k} converges in HH if and only if ∑k=1∞∥xk∥2<∞\sum_{k = 1}^{\infty }\left\| x_{k}\right\|^{2} < \infty.

?
Problem 2.18

Prove that a finite system {x1,…,xn}\left\{ x_{1}, \ldots , x_{n}\right\} of elements of a Hilbert space is linearly independent if and only if its Gram determinant

[(x1,x1)(x1,x2)…(x1,xn)(x2,x1)(x2,x2)…(x2,xn)…………(xn,x1)(xn,x2)…(xn,xn)] \left[\begin{smallmatrix} \left(x_{1}, x_{1}\right) & \left(x_{1}, x_{2}\right) & \ldots & \left(x_{1}, x_{n}\right) \\ \left(x_{2}, x_{1}\right) & \left(x_{2}, x_{2}\right) & \ldots & \left(x_{2}, x_{n}\right) \\ \ldots & \ldots & \ldots & \ldots \\ \left(x_{n}, x_{1}\right) & \left(x_{n}, x_{2}\right) & \ldots & \left(x_{n}, x_{n}\right) \end{smallmatrix}\right]

is nonzero.

?
Problem 2.19

Let HH be a Hilbert space, x,y∈H,∥x∥=∥y∥=1,a,b∈Rx, y \in H,\left\| x\right\| = \left\| y\right\| = 1, a, b \in \mathbb {R}. Prove that ∥ax+by∥=∥bx+ay∥\left\| a x+b y\right\| = \left\| b x+a y\right\|.

?
Problem 2.20

Let HH be a Hilbert space, {xn}⊂H,{yn}⊂H\left\{ x_{n}\right\} \subset H,\left\{ y_{n}\right\} \subset H, and ∥xn∥=∥yn∥=1,n≥\left\| x_{n}\right\| = \left\| y_{n}\right\| = 1, n \geq 1. Prove that

?
(1)

(xn,yn)→1,n→∞⇒∥xn−yn∥→0,n→∞\left(x_{n}, y_{n}\right) \rightarrow 1, n \rightarrow \infty \Rightarrow \left\| x_{n}-y_{n}\right\| \rightarrow 0, n \rightarrow \infty;

(2)

∥xn+yn∥→2,n→∞⇒∥xn−yn∥→0,n→∞\left\| x_{n}+y_{n}\right\| \rightarrow 2, n \rightarrow \infty \Rightarrow \left\| x_{n}-y_{n}\right\| \rightarrow 0, n \rightarrow \infty.

Problem 2.21

Prove that an element xx of a Hilbert space HH is orthogonal to a subspace LL if and only if ∥x∥≤∥x−y∥\left\| x\right\| \leq \left\| x-y\right\| for all y∈Ly \in L.

?
Problem 2.22

Prove that for an arbitrary set M⊂HM \subset H the orthogonal complement M⊥M^{\perp } is a subspace.

?
Problem 2.23

Let M⊂N⊂HM \subset N \subset H. Prove that M⊥⊃N⊥M^{\perp } \supset N^{\perp }.

?
Problem 2.24

Prove that

?
(1)

M⊂(M⊥)⊥M \subset \left(M^{\perp }\right)^{\perp } for any M⊂HM \subset H;

(2)

the equality M=(M⊥)⊥M = \left(M^{\perp }\right)^{\perp } holds if and only if MM is a subspace.

Problem 2.25

Let HH be a Hilbert space. Prove that

?
(1)

M⊥=(span⁡(M))⊥=(span⁡(M)‾)⊥M^{\perp } = (\operatorname {span}(M))^{\perp } = (\overline{\operatorname {span}(M)})^{\perp } for all subsets M⊂HM \subset H;

(2)

(⋃α∈AMα)⊥=⋂α∈AMα⊥\left(\bigcup_{\alpha \in A} M_{\alpha }\right)^{\perp } = \bigcap_{\alpha \in A} M_{\alpha }^{\perp }, where {Mα∣α∈A}\left\{ M_{\alpha } \mid \alpha \in A\right\} are subsets of HH;

(3)

(⋂α∈ALα)⊥=span⁡(⋃α∈ALα⊥)‾\left(\bigcap_{\alpha \in A} L_{\alpha }\right)^{\perp } = \overline{\operatorname {span}\left(\bigcup_{\alpha \in A} L_{\alpha }^{\perp }\right)}, where {Lα∣α∈A}\left\{ L_{\alpha } \mid \alpha \in A\right\} are subspaces of HH.

Problem 2.26

Let L={x∈L2(R)∣x(t)=0,t≥0( mod m)}L = \left\{ x \in L_{2}(\mathbb {R}) \mid x(t) = 0, t \geq 0(\bmod m)\right\}. Prove that LL is a subspace of L2(R)L_{2}(\mathbb {R}) and find the orthogonal complement of LL.

?
Problem 2.27

Let L⊂HL \subset H. Prove that span⁡(L)‾=H\overline{\operatorname {span}(L)} = H if and only if L⊥={0}L^{\perp } = \left\{ 0\right\}.

?
Problem 2.28

Find in L2([0,1])L_{2}(\left[0,1\right]) the orthogonal complement of the set

?
(1)

C([0,1])C(\left[0,1\right]);

(2)

all polynomials;

(3)

all polynomials of t2t^{2};

(4)

all polynomials x(t)x(t), for which x(0)=0x(0) = 0;

(5)

all polynomials of ete^{t};

(6)

{tk,t∈[0,1]∣k≥10}\left\{ t^{k}, t \in [0,1] \mid k \geq 10\right\};

(7)

{t3k,t∈[0,1]∣k≥1}\left\{ t^{3 k}, t \in [0,1] \mid k \geq 1\right\}.

Problem 2.29

Find in L2([−1,1])L_{2}(\left[-1,1\right]) the orthogonal complement of the set

?
(1)

{x(t)=tk,t∈[−1,1]∣k≥13}\left\{ x(t) = t^{k}, t \in [-1,1] \mid k \geq 13\right\};

(2)

{t2k,t∈[−1,1]∣k≥0}\left\{ t^{2 k}, t \in [-1,1] \mid k \geq 0\right\};

(3)

{t2k+1,t∈[−1,1]∣k≥0}\left\{ t^{2 k+1}, t \in [-1,1] \mid k \geq 0\right\};

(4)

{t2k,t∈[−1,1]∣k≥7}\left\{ t^{2 k}, t \in [-1,1] \mid k \geq 7\right\}.

Problem 2.30

Find in L2([−π,π])L_{2}(\left[-\pi , \pi \right]) the orthogonal complement of the set

?
(1)

{sin⁡kt∣k≥1}\left\{ \sin k t \mid k \geq 1\right\};

(2)

{cos⁡kt∣k≥1}\left\{ \cos k t \mid k \geq 1\right\};

(3)

{eikt∣k≥5}\left\{ e^{i k t} \mid k \geq 5\right\};

(4)

{e−ikt∣k≥3}\left\{ e^{-i k t} \mid k \geq 3\right\}.

Problem 2.31

Find in l2l_{2} the orthogonal complement of the set

?
(1)

{(1,1,0,…)}\left\{ (1,1,0, \ldots )\right\};

(2)

{ek,∣1≤k≤n},n∈N\left\{ e_{k}, \mid 1 \leq k \leq n\right\} , n \in \mathbb {N};

(3)

{e2k∣k∈N}\left\{ e_{2 k} \mid k \in \mathbb {N}\right\}.

Problem 2.32

Prove that the function

(x,y)=∫abx(t)y(t)‾dt+∫abx′(t)y′(t)‾dt,x,y∈C1([a,b]) (x, y) = \int _{a}^{b} x(t) \overline{y(t)} d t+\int _{a}^{b} x^{\prime }(t) \overline{y^{\prime }(t)} d t, x, y \in C^{1}(\left[a, b\right])

defines a scalar product in the linear space C1([a,b])C^{1}(\left[a, b\right]). The completion of the space C1([a,b])C^{1}(\left[a, b\right]) in the norm generated by this scalar product is a Hilbert space called Sobolev space W21([a,b])W_{2}^{1}(\left[a, b\right]).

?
Problem 2.33

Find the orthogonal complement of the set {eikt∣k∈Z}\left\{ e^{i k t} \mid k \in \mathbb {Z}\right\} in the space W21([−π,π])W_{2}^{1}(\left[-\pi , \pi \right]).

?
Problem 2.34

Consider the space HH that consists of the functions x:R→Rx: \mathbb {R} \rightarrow \mathbb {R}, for which {x∈R∣x(t)≠0}\left\{ x \in \mathbb {R} \mid x(t) \neq 0\right\} is a countable set and ∑t∈R∣x(t)∣2<+∞\sum_{t \in \mathbb {R}}\left|x(t)\right|^{2} < +\infty. Prove that HH is a nonseparable Hilbert space with the scalar product (x,y)=∑t∈Rx(t)y(t),x,y∈H(x, y) = \sum_{t \in \mathbb {R}} x(t) y(t), x, y \in H.

?
Problem 2.35

Prove that the system of functions {e−int∣n∈Z}\left\{ e^{-i n t} \mid n \in \mathbb {Z}\right\} is complete in L2([a,b])L_{2}(\left[a, b\right]) if and only if b−a≤2πb-a \leq 2 \pi.

?
Problem 2.36

Let LL be a linear span of the set {eiλt∣λ∈R}\left\{ e^{i \lambda t} \mid \lambda \in \mathbb {R}\right\} equipped with the scalar product (x,y)=lim⁡T→∞12T∫−TTx(t)y(t)‾dt,x,y∈L(x, y) = \lim_{T \rightarrow \infty } \frac{1}{2 T} \int_{-T}^{T} x(t) \overline{y(t)} d t, x, y \in L. Prove that

?
(1)

(eiλt,eiμt)=δλμ={1,λ=μ,0,λ≠μ,λ,μ∈R\left(e^{i \lambda t}, e^{i \mu t}\right) = \delta_{\lambda \mu } = \begin{cases} 1, & \lambda = \mu , \\ 0, & \lambda \neq \mu , \end{cases} \quad \lambda , \mu \in \mathbb {R};

(2)

the completion of the set LL is nonseparable.

Problem 2.37

Prove that the system of Haar functions

{xkn∣1≤k≤2n,n≥0}∪{x0} \left\{ x_{k n} \mid 1 \leq k \leq 2^{n}, n \geq 0\right\} \cup \left\{ x_{0}\right\}

where xkn(t)={2n2,t∈[k−12n,k−1/22n),−2n2,t∈[k−1/22n,k2n),0,t∈[0,1]\[k−12n,k2n),x0(t)=1,t∈[0,1]x_{k n}(t) = \begin{cases} 2^{\frac{n}{2}}, & t \in \left[\frac{k-1}{2^{n}}, \frac{k-1 / 2}{2^{n}}\right), \\ -2^{\frac{n}{2}}, & t \in \left[\frac{k-1 / 2}{2^{n}}, \frac{k}{2^{n}}\right), \\ 0, & t \in [0,1] \backslash \left[\frac{k-1}{2^{n}}, \frac{k}{2^{n}}\right), \end{cases} \quad x_{0}(t) = 1, t \in [0,1], is an orthonormal basis in the space L2([0,1])L_{2}(\left[0,1\right]).

?
Problem 2.38

Verify that the following systems are orthogonal in HH :

?
(1)

Rademacher's system of functions

xn(t)={(−1)m,t∈[m2n,m+12n),m=0,1,…,2n−1−1,t=1 x_{n}(t) = \begin{cases} (-1)^{m}, & t \in \left[\frac{m}{2^{n}}, \frac{m+1}{2^{n}}\right), m = 0,1, \ldots , 2^{n}-1 \\ -1, & t = 1 \end{cases}

n≥1,H=L2([0,1]);n \geq 1, H = L_{2}(\left[0,1\right]) ;

(2)

Legendre polynomials Pn(t)=12nn!dndtn(t2−1)n,n≥0,H=L2([−1,1])P_{n}(t) = \frac{1}{2^{n} n!} \frac{d^{n}}{d t^{n}}\left(t^{2}-1\right)^{n}, n \geq 0, H = L_{2}(\left[-1,1\right]);

(3)

Hermite functions Hn(t)=et2/2dndtne−t2,n≥0,H=L2(R)H_{n}(t) = e^{t^{2} / 2} \frac{d^{n}}{d t^{n}} e^{-t^{2}}, n \geq 0, H = L_{2}(\mathbb {R});

(4)

xn(t)=dndtn[(t−a)(t−b)]n,n≥0,H=L2([a,b])x_{n}(t) = \frac{d^{n}}{d t^{n}}\left[(t-a)(t-b)\right]^{n}, n \geq 0, H = L_{2}(\left[a, b\right]);

(5)

Laguerre functions xn(t)=et/2dndtn(tne−t),n≥0,H=L2([0,+∞))x_{n}(t) = e^{t / 2} \frac{d^{n}}{d t^{n}}\left(t^{n} e^{-t}\right), n \geq 0, H = L_{2}([0,+\infty )).

Problem 2.39

Prove that in problem 2.38

?
(1)

Rademacher's system of function {xn(t):n≥1}\left\{ x_{n}(t): n \geq 1\right\} is not an orthonormal basis in L2([0,1])L_{2}(\left[0,1\right]);

(2)

the functions {2n+12Pn(t):n≥0}\left\{ \sqrt{\frac{2 n+1}{2}} P_{n}(t): n \geq 0\right\} form an orthonormal basis in L2([−1,1])L_{2}(\left[-1,1\right]).

Problem 2.40

In the space L2(R)L_{2}(\mathbb {R}), consider the set

M={x∈C0(R)∣∫−∞∞x(t)dt=0} M = \left\{ x \in C_{0}(\mathbb {R}) \mid \int _{-\infty }^{\infty } x(t) d t = 0\right\}

where C0(R)C_{0}(\mathbb {R}) is the set of continuous functions on R\mathbb {R} which vanish outside some finite interval, see problem 1.54∘1.54^{\circ }. Prove that MM is dense in L2(R)L_{2}(\mathbb {R}).

?
Problem 2.41

In the space l2l_{2}, find the orthogonal complement of the set

?
(1)

{x=(1,α,α2,…,αk,…)∣α∈(0,1)}\left\{ x = \left(1, \alpha , \alpha^{2}, \ldots , \alpha^{k}, \ldots \right) \mid \alpha \in (0,1)\right\};

(2)

{x=(1,1n,1n2,…,1nk,…)∣ n≥2}\left\{ \left.x = \left(1, \frac{1}{n}, \frac{1}{n^{2}}, \ldots , \frac{1}{n^{k}}, \ldots \right) \right\rvert \, n \geq 2\right\}.

Problem 2.42

Let M⊂HM \subset H, where HH is a Hilbert space. Is the formula H=M⊕M⊥H = M \oplus M^{\perp } always true?

?
Problem 2.43

Let {en:n≥1}\left\{ e_{n}: n \geq 1\right\} be a standard basis in l2,x1=∑k=1∞ekkl_{2}, x_{1} = \sum_{k = 1}^{\infty } \frac{e_{k}}{k} and H=span⁡({x1,ek:k≥2})⊂l2H = \operatorname {span}\left(\left\{ x_{1}, e_{k} : k \geq 2\right\} \right) \subset l_{2}. Prove that in the pre-Hilbert space HH the set M={en:n≥2}M = \left\{ e_{n}: n \geq 2\right\} is a complete orthonormal system, but span⁡(M)‾≠H\overline{\operatorname {span}(M)} \neq H.

?
Problem 2.44

Prove that a separable space equipped with a scalar product is complete if and only if every complete orthonormal system in this space is a basis.

?
Problem 2.45

Prove that in a Hilbert space there exists a dense linear subset, which does not coincide with the entire space, if and only if the space is infinite-dimensional.

?
Problem 2.46

Let {en:n≥1}\left\{ e_{n}: n \geq 1\right\} be linearly independent system of elements of a Hilbert space HH such that ∑n=1∞cnen=0\sum_{n = 1}^{\infty } c_{n} e_{n} = 0 for some sequence {cn:n≥1}⊂K\left\{ c_{n}: n \geq 1\right\} \subset \mathbb {K}. Is it true that cn=0,n≥1c_{n} = 0, n \geq 1?

?
Problem 2.47

Let HH be a Hilbert space over a field K,{en:n≥1}⊂H\mathbb {K},\left\{ e_{n}: n \geq 1\right\} \subset H be an orthonormal system and x=∑n=1∞xnenx = \sum_{n = 1}^{\infty } x_{n} e_{n}, where xn∈K,n≥1x_{n} \in \mathbb {K}, n \geq 1. Prove that xn=(x,en),n≥1x_{n} = \left(x, e_{n}\right), n \geq 1.

?
Problem 2.48

Prove that the system {tn,t∈[0,1]:n≥0}\left\{ t^{n}, t \in [0,1]: n \geq 0\right\} is linearly independent, and its closed linear span is equal to L2([0,1])L_{2}(\left[0,1\right]), but not for all x∈L2([0,1])x \in L_{2}(\left[0,1\right]) there exist coefficients {cn:n≥0}\left\{ c_{n}: n \geq 0\right\} such that x(t)=∑n=0∞cntnx(t) = \sum_{n = 0}^{\infty } c_{n} t^{n} (the convergence here is in the sense of convergence in L2([0,1])L_{2}(\left[0,1\right]).) Does this contradict to the statement of Theorem 2.6? Why?

Theorem 2.6: Let {en:n≥1}\left\{ e_{n}: n \geq 1\right\} be an orthonormal sequence in a Hilbert space HH. Then the following conditions are equivalent:

?
(1)

{en:n≥1}\left\{ e_{n}: n \geq 1\right\} is an orthonormal basis;

(2)

∀x∈H:x=∑n=1∞(x,en)en\forall x \in H: x = \sum_{n = 1}^{\infty }\left(x, e_{n}\right) e_{n};

(3)

∀x∈H:∥x∥2=∑n=1∞∣(x,en)∣2\forall x \in H:\left\| x\right\|^{2} = \sum_{n = 1}^{\infty }\left|\left(x, e_{n}\right)\right|^{2} (Parseval's identity);

(4)

{en:n≥1}\left\{ e_{n}: n \geq 1\right\} is a complete system.

Problem 2.49

Let D=Bˉ(0,1)⊂CD = \bar{B}(0,1) \subset \mathbb {C} be the closed unit circle in the complex plane, A2(D)A_{2}(D) be the set of functions x∈L2(D)x \in L_{2}(D), for which x(z)=∑n=0∞anznx(z) = \sum_{n = 0}^{\infty } a_{n} z^{n} for ∣z∣<1\left|z\right| < 1. Prove that A2(D)A_{2}(D) is the completion of the set P(D)P(D) of all polynomials defined on DD, in the norm of L2(D)L_{2}(D).

?
Problem 2.50

Apply the orthogonalization process to the sequence {zn:n≥0},z=t+is\left\{ z^{n}: n \geq 0\right\} , z = t+i s, with respect to the following scalar product on the set of all polynomials P(C)P(\mathbb {C}) :

(x,y)=∫Cx(z)y(z)‾e−∣z∣2dtds,x,y∈P(C) (x, y) = \int _{\mathbb {C}} x(z) \overline{y(z)} e^{-\left|z\right|^{2}} d t d s, x, y \in P(\mathbb {C})

Describe the completion of the set P(C)P(\mathbb {C}).

?
Problem 2.51

Let {φn:n≥1}\left\{ \varphi_{n}: n \geq 1\right\} and {ψn:n≥1}\left\{ \psi_{n}: n \geq 1\right\} be complete orthonormal systems in L2(X1,F1,μ1)L_{2}\left(X_{1}, \mathcal{F}_{1}, \mu_{1}\right) and L2(X2,F2,μ2)L_{2}\left(X_{2}, \mathcal{F}_{2}, \mu_{2}\right), respectively, and μi\mu_{i} be σ\sigma-finite measures on Fi,i=1,2\mathcal{F}_{i}, i = 1,2. Prove that en,m(x,y)=φn(x)ψm(y),(x,y)∈X=X1×X2,n,m≥e_{n, m}(x, y) = \varphi_{n}(x) \psi_{m}(y),(x, y) \in X = X_{1} \times X_{2}, n, m \geq 1, is a complete orthonormal system in L2(X,F,μ)L_{2}(X, \mathcal{F}, \mu ), where μ=μ1×μ2,F=σ(F1×F2)\mu = \mu_{1} \times \mu_{2}, \mathcal{F} = \sigma \left(\mathcal{F}_{1} \times \mathcal{F}_{2}\right).

?
Problem 2.52

Prove that a Hilbert space is uniformly convex (see problem 1.98).

?
Problem 2.53

Let MM be a nonempty closed convex set in a Hilbert space HH. Prove that in MM there exists a unique element with the smallest norm. Show that the convexity condition is essential.

?
Problem 2.54

Let HH be a Hilbert space, x0∈H,r>0x_{0} \in H, r > 0. For x∈H\S(x0,r)x \in H \backslash S\left(x_{0}, r\right) find y∈S(x0,r)y \in S\left(x_{0}, r\right) such that ρ(x,S(x0,r))=∥x−y∥\rho \left(x, S\left(x_{0}, r\right)\right) = \left\| x-y\right\|.

?
Problem 2.55

Let x∈Hx \in H and M⊂HM \subset H be a closed convex set in a real Hilbert space HH. Prove that y∈My \in M satisfies the condition ρ(x,M)=∥x−y∥\rho (x, M) = \left\| x-y\right\| if and only if (x−y,y−z)≥0(x-y, y-z) \geq 0 for all z∈Mz \in M.

?
Problem 2.56

Let LL be a one-dimensional subspace of a Hilbert space H,a∈L,a≠0H, a \in L, a \neq 0. Prove that ρ(x,L⊥)=∣(x,a)∣∥a∥\rho \left(x, L^{\perp }\right) = \frac{\left|(x, a)\right|}{\left\| a\right\| } for all x∈Hx \in H.

?
Problem 2.57

For the set Mn={x∈l2∣∑k=1nxk=0},n∈NM_{n} = \left\{ x \in l_{2} \mid \sum_{k = 1}^{n} x_{k} = 0\right\} , n \in \mathbb {N}, find Mn⊥,ρ(x0,Mn)M_{n}^{\perp }, \rho \left(x_{0}, M_{n}\right) and lim⁡n→∞ρ(x0,Mn)\lim_{n \rightarrow \infty } \rho \left(x_{0}, M_{n}\right), where x0=(1,0,…)x_{0} = (1,0, \ldots ).

?
Problem 2.58

Let M,NM, N be subspaces of a Hilbert space HH and for each x∈Hx \in H there exist unique x1∈Mx_{1} \in M and x2∈Nx_{2} \in N such that x=x1+x2x = x_{1}+x_{2}. Is it true that N=M⊥N = M^{\perp }?

?
Problem 2.59

Let

M={(x1,0,x3,0,…)∣x=(x1,x2,x3,…)∈l2}N={(x1,x1,x3,x33,x5,x55,…)∣ x=(x1,x2,x3,…)∈l2} \begin{aligned} M & = \left\{ \left(x_{1}, 0, x_{3}, 0, \ldots \right) \mid x = \left(x_{1}, x_{2}, x_{3}, \ldots \right) \in l_{2}\right\} \\ N & = \left\{ \left.\left(x_{1}, x_{1}, x_{3}, \frac{x_{3}}{3}, x_{5}, \frac{x_{5}}{5}, \ldots \right) \right\rvert \, x = \left(x_{1}, x_{2}, x_{3}, \ldots \right) \in l_{2}\right\} \end{aligned}

Prove that MM and NN are subspaces of l2l_{2} and the set M+NM+N is dense in l2l_{2}, but M+N≠l2M+N \neq l_{2}, i.e. M+NM+N is not a subspace of l2l_{2}.

?
Problem 2.60

Let {en:n≥1}\left\{ e_{n}: n \geq 1\right\} be an orthonormal basis in H,un=cos⁡1n⋅e2n+sin⁡1n⋅e2n+1H, u_{n} = \cos \frac{1}{n} \cdot e_{2 n}+\sin \frac{1}{n} \cdot e_{2 n+1}. Prove that {un:n≥1}\left\{ u_{n}: n \geq 1\right\} is an orthonormal system. Put M=span⁡({e2n:n≥1})‾M = \overline{\operatorname {span}\left(\left\{ e_{2 n}: n \geq 1\right\} \right)} and N=span⁡({un:n≥1})N = \operatorname {span}\left(\left\{ u_{n}: n \geq 1\right\} \right). Prove that the linear set M+NM+N is not closed.

?
Problem 2.61

Let MM and NN be subspaces of a Hilbert space HH. Prove that M+NM+N is a subspace of HH, if the following condition holds:

∃ε∈(0,1]:sup⁡{∣(x,y)∣∣∥x∥=∥y∥=1,x∈M,y∈N}≤1−ε, \exists \varepsilon \in (0,1]: \sup \left\{ \left|(x, y)\right| \mid \left\| x\right\| = \left\| y\right\| = 1, x \in M, y \in N\right\} \leq 1-\varepsilon ,

in particular if M⊥NM \perp N.

?
Problem 2.62

Let {xn:n≥1},{yn:n≥1}\left\{ x_{n}: n \geq 1\right\} ,\left\{ y_{n}: n \geq 1\right\} be orthonormal systems in L2([a,b])L_{2}(\left[a, b\right]) and L2([b,c])L_{2}(\left[b, c\right]) respectively, λ>0,μ>0,λ2+μ2=1\lambda > 0, \mu > 0, \lambda^{2}+\mu^{2} = 1. Put

zn(t)={λxn(t),a≤t≤b,μyn(t),b≤t≤c. z_{n}(t) = \begin{cases} \lambda x_{n}(t), & a \leq t \leq b, \\ \mu y_{n}(t), & b \leq t \leq c. \end{cases}

Prove that {zn:n≥1}\left\{ z_{n}: n \geq 1\right\} is an orthonormal system in L2([a,c])L_{2}(\left[a, c\right]). Is {zn:n≥1}\left\{ z_{n}: n \geq 1\right\} necessarily an orthonormal basis in L2([a,c])L_{2}(\left[a, c\right]), if {xn:n≥1}\left\{ x_{n}: n \geq 1\right\} and {yn:n≥1}\left\{ y_{n}: n \geq 1\right\} are orthonormal bases in L2([a,b])L_{2}(\left[a, b\right]) and L2([b,c])L_{2}(\left[b, c\right]), respectively?

?
Problem 2.63

Let a<b<ca < b < c and {xn:n≥1}\left\{ x_{n}: n \geq 1\right\} be an arbitrary countable system of functions from L2([a,b])L_{2}(\left[a, b\right]). Prove that the functions xnx_{n} can be extended to the segment [b,c]\left[b, c\right] to obtain an orthogonal system in L2([a,c])L_{2}(\left[a, c\right]).

?
Problem 2.64

Show that the functions xn,k(t)=2χ[k−1,k](t)sin⁡πntx_{n, k}(t) = \sqrt{2} \chi_{\left[k-1, k\right]}(t) \sin \pi n t for n∈N,k∈Nn \in \mathbb {N}, k \in \mathbb {N} form an orthonormal basis in L2([0,+∞))L_{2}([0,+\infty )), and for n∈N,k∈Zn \in \mathbb {N}, k \in \mathbb {Z} they form an orthonormal basis in L2(R)L_{2}(\mathbb {R}).

?