Mazur's lemma (weak limits and convex hulls)

[8/25%]
Show
LaTeX

Let XX be an NLS. A set M⊂XM \subset X is called weakly closed if for any sequence {xn:n≥1}⊂M\left\{ x_{n}: n \geq 1\right\} \subset M such that xn→wx∈Xx_{n} \xrightarrow {w} x \in X, it follows that x∈Mx \in M. Prove that

?
(1)

every weakly closed set in XX is strongly closed;

(2)

the converse statement to statement (1) is false;

(3)

every subspace in XX is a weakly closed set;

(4)

every closed ball in XX is a weakly closed set.

Let XX be a Banach space, and let the system {en}1∞\left\{ e_{n}\right\}_{1}^{\infty } be weakly complete in it, i.e., for every vector x∈Xx \in X there exists a sequence {xk}1∞⊂Lin⁡{en}1∞\left\{ x_{k}\right\}_{1}^{\infty } \subset \operatorname {Lin}\left\{ e_{n}\right\}_{1}^{\infty }, weakly convergent to xx. Prove that then the system {en}1∞\left\{ e_{n}\right\}_{1}^{\infty } is complete in XX in the usual sense.

?

Let XX be a Banach space, and UU a convex set in it. Prove that UU is closed with respect to the norm if and only if UU is closed in the weak topology.

?

(S. Mazur, 1933). Let a sequence xnx_{n} converge weakly to xx in a normed space XX. Prove that x∈conv⁡{xn}1∞‾x \in \overline{\operatorname {conv}\left\{ x_{n}\right\}_{1}^{\infty }}.

?

Let X\mathcal{X} be a normed space and let {xn}\left\{ x_n\right\} be a sequence in X\mathcal{X} such that xn→xx_n \to x weakly. Show that there is a sequence {yn}\left\{ y_n\right\} such that yn∈co⁡{x1,x2,…,xn}y_n \in \operatorname {co}\left\{ x_1, x_2, \dots , x_n\right\} and ∥yn−x∥→0\left\| y_n - x\right\| \to 0.

?

Every weakly closed set is strongly closed, but the converse is not true. Nevertheless every subspace of a Hilbert space (i.e., every strongly closed linear manifold) is weakly closed.

?
Remark.
?

The metric topology (or norm topology) of a Hilbert space is often called the strong topology. A base for the strong topology is the collection of open balls, i.e., sets of the form {f:∥f−f0∥<ε}\left\{ f : \left\| f - f_0\right\| < \varepsilon \right\}, where f0f_0 (the center) is a vector and ε\varepsilon (the radius) is a positive number.

A subbase (not a base) for the weak topology is the collection of all sets of the form {f:∣⟨f−f0,g0⟩∣<ε}\left\{ f : \left| \left\langle f - f_0, g_0 \right\rangle \right| < \varepsilon \right\}. It follows that a base for the weak topology is the collection of all sets of the form {f:∣⟨f−f0,gi⟩∣<ε,i=1,…,k}\left\{ f : \left| \left\langle f - f_0, g_i \right\rangle \right| < \varepsilon , i = 1, \ldots , k\right\}, where kk is a positive integer, f0,g1,…,gkf_0, g_1, \ldots , g_k are vectors, and ε\varepsilon is a positive number.

A sequence, or a net, {fn}\left\{ f_n\right\} is weakly convergent to ff if and only if ⟨fn,g⟩→⟨f,g⟩\left\langle f_n, g \right\rangle \to \left\langle f, g \right\rangle for each gg.

Prove that a subspace of a normed linear space is closed if and only if it is weakly closed.

?

Put fN(t)=N−1∑n=1N2eintf_{N}(t) = N^{-1} \sum_{n = 1}^{N 2} e^{i n t}. Prove that fN→0f_{N} \rightarrow 0 weakly in L2(−π,π)L^{2}(-\pi , \pi ).

By Theorem 3.13, some sequence of convex combinations of the fNf_{N} converges to 0 in the L2L^{2}-norm. Find such a sequence. Show that gN=N−1(f1+⋯+fN)g_{N} = N^{-1}\left(f_{1}+\cdots +f_{N}\right) will not do.

?

References

[BCK+24] Brayman V.B., Chaikovskyi A.V., Konstantinov O.Y., Kukush A.G., Mishura Y.S., Nesterenko O.N. Functional Analysis and Operator Theory. Springer, 2024.
[БСШ17] Бородин П.А., Савчук А.М., Шейпак И.А. Задачи по функциональному анализу. МЦНМО, 2017.
[Con07] Conway J.B. A Course in Functional Analysis. 2nd ed., Springer, 2007.
[Hal82] Halmos P.R. A Hilbert Space Problem Book. 2nd ed., Springer-Verlag, 1982.
[Pet25] Petrovic J.S. Measure and Integral: Theory and Practice. CRC Press, 2025.
[Rud91] Rudin W. Functional Analysis. 2nd ed., McGraw-Hill, 1991.