Section 5.4

The second adjoint space. Reflexivity

[15/20%]
Show
LaTeX
Problem 5.67

Prove that if XX is a reflexive space, then for any functional f∈X∗f \in X^{*} there exists an element x∈Xx \in X such that ∥x∥=1\left\| x\right\| = 1 and f(x)=∥f∥f(x) = \left\| f\right\| (necessity in James's theorem).

?
Problem 5.68

Prove that the canonical embedding π:X→X∗∗\pi : X \rightarrow X^{* *} is an isometry.

?
Problem 5.69

Prove the non-reflexivity of the spaces:

?
(a)

C[a,b]C[a, b];

(b)

Cn[0,1],n=1,2,…C^{n}[0,1], n = 1,2, \ldots;

(c)

l1l_{1};

(d)

c0c_{0};

(e)

cc;

(f)

L1[a,b]L_{1}[a, b].

Problem 5.70

Prove that if the space XX is reflexive, then X∗X^{*} is also reflexive.

?
Problem 5.71

Prove the non-reflexivity of the spaces:

?
(a)

l∞l_{\infty };

(b)

L∞[0,1]L_{\infty }[0,1].

Problem 5.72

Prove that any finite-dimensional normed space and any Hilbert space is reflexive.

?
Problem 5.73

Prove the reflexivity of the spaces:

?
(a)

lp,p∈(1,∞)l_{p}, p \in (1, \infty );

(b)

Lp[0,1],p∈(1,∞)L_{p}[0,1], p \in (1, \infty ).

Problem 5.74

Prove that if XX is a Banach space and X∗X^{*} is reflexive, then XX is also reflexive.

?
Problem 5.75

Prove that there does not exist a Banach space XX such that X∗=c0X^{*} = c_{0}.

?
Problem 5.76

Prove that a Banach space XX is reflexive if and only if every closed subspace Y⊂XY \subset X of codimension 1 is a subspace of existence (compare with problem 14.127).

?
Problem 5.77

Give an example of a Banach space XX and a closed subspace Y⊂XY \subset X that is not a subspace of existence (compare with problem 5.22).

?
Problem 5.78

Let MM be an arbitrary set in a normed space XX.

?
(a)

Prove that M⊥M^{\perp } is a closed subspace of X∗X^{*}.

(b)

Prove that the set {x∈X:f(x)=0 for every f∈M⊥}\left\{ x \in X: f(x) = 0 \text{ for every } f \in M^{\perp }\right\} coincides with Lin⁡(M)‾\overline{\operatorname {Lin}(M)}.

(c)

Prove that if MM is a closed linear subspace in a reflexive space XX, then (M⊥)⊥=M\left(M^{\perp }\right)^{\perp } = M.

(d)

Prove that if MM is a linear subspace in XX, then dim⁡(M⊥)=codim⁡M\operatorname {dim}\left(M^{\perp }\right) = \operatorname {codim} M.

Problem 5.79

Find the set M⊥M^{\perp } if:

?
(a)

M=B(x,r)M = B(x, r) in the normed space XX;

(b)

M={x∈C[−1,1]:x(t)=0 for every t<0}M = \left\{ x \in C[-1,1]: x(t) = 0\text{ for every }t < 0\right\} in the space С [−1,1][-1,1];

(c)

M={x∈C[0,1]:x(0)=0}M = \left\{ x \in C[0,1]: x(0) = 0\right\} in the space C[0,1]C[0,1];

(d)

M={x∈L1[0,1]:x is continuous at the point a and x(a)=0}M = \left\{ x \in L_{1}[0,1]: x \text{ is continuous at the point } a \text{ and } x(a) = 0\right\} in the space L1[0,1]L_{1}[0,1];

(e)

M={x∈L1[−1,1]:x(t)=x(−t) for every t∈[−1,1]}M = \left\{ x \in L_{1}[-1,1]: x(t) = x(-t) \text{ for every } t \in [-1,1]\right\} in the space L1[−1,1]L_{1}[-1,1];

(f)

the set MM is the set of all polynomials in the space C[0,1]C[0,1];

(g)

the set MM in the space C[0,1]C[0,1] is the set of all polynomials pp such that p(0)=0p(0) = 0;

(h)

M={x∈C[0,1]:x(t)⩾0 for every t∈[0,1]}M = \left\{ x \in C[0,1]: x(t) \geqslant 0\text{ for every }t \in [0,1]\right\} in the space C[0,1]C[0,1] over the field R\mathbb {R};

(i)

M={x∈C[0,1]:x(a)⩾x(b) for every a>b}M = \left\{ x \in C[0,1]: x(a) \geqslant x(b)\text{ for every }a > b\right\} in the space C[0,1]C[0,1] over the field R\mathbb {R};

(j)

M={x∈L3[−1,1]:∫01x(t)dt=0}M = \left\{ x \in L_{3}[-1,1]: \int_{0}^{1} x(t) d t = 0\right\} in the space L3[0,1]L_{3}[0,1];

(k)

M=c00M = c_{00} in the space c0c_{0};

(l)

M=c0M = c_{0} in the space cc;

(m)

M={x∈l1:x1>x2>x3>…}M = \left\{ x \in l_{1}: x_{1} > x_{2} > x_{3} > \ldots \right\} in the space l1l_{1} over the field R\mathbb {R};

(n)
  • M={x=(λ,λ2,λ3,…):λ∈{1,12,13,…}}M = \left\{ x = \left(\lambda , \lambda^{2}, \lambda^{3}, \ldots \right): \lambda \in \left\{ 1, \frac{1}{2}, \frac{1}{3}, \ldots \right\} \right\} in the space с.
Problem 5.80

Let YY be a linear subspace of a normed space XX. Prove that for any element x∈Xx \in X the equality dist⁡(x,Y)=max⁡{f(x):f∈Y⊥,∥f∥=1}\operatorname {dist}(x, Y) = \max \left\{ f(x): f \in Y^{\perp },\left\| f\right\| = 1\right\} holds.

?
Problem 5.81

Give an example of a normed space XX, a closed subspace YY of X∗X^{*}, and a bounded linear functional f0∉Yf_{0} \notin Y such that for every vector x∈Xx \in X, from the equalities f(x)=0f(x) = 0 for every f∈Yf \in Y it follows that f0(x)=0f_{0}(x) = 0 (the closed subspace YY and the functional f0f_{0} cannot be separated by means of a vector). Prove that in a reflexive space there necessarily exists a vector x0∈Xx_{0} \in X such that f(x0)=0f\left(x_{0}\right) = 0 for every f∈Yf \in Y, but f0(x0)≠0f_{0}\left(x_{0}\right) \neq 0 (compare with problem 5.37).

?