Some remarks on diametral dimension and approximate diametral dimension of certain nuclear Fréchet spaces (Q1999124)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Some remarks on diametral dimension and approximate diametral dimension of certain nuclear Fréchet spaces |
scientific article |
Statements
Some remarks on diametral dimension and approximate diametral dimension of certain nuclear Fréchet spaces (English)
0 references
18 March 2021
0 references
The Kolmogorov \(n\)-width, \(n\in\mathbb{N}\), of absolutely convex sets \(U\) and \(V\) is defined as the number \[ \delta_n(V,U) = \inf\{\delta > 0: V \subset \delta U + L, \ L \text{ a subspace of dimension at most } n\}. \] For a locally convex space \(E\) with a neighbourhood basis of zero \(\mathcal{U}_{0}(E)\) consisting of absolutely convex sets, the diametral dimension of \(E\) is the sequence space \[ \Delta(E) = \{(t_n)_{n\in\mathbb{N}}: \forall U\in\mathcal{U}_0(E)\, \exists V\in \mathcal{U}_0(E) : \lim_{n\to\infty} t_n \delta_n(V,U) = 0 \}. \] The approximate diametral dimension of \(E\) is defined as \[ \delta(E) = \left\{(t_n)_{n\in\mathbb{N}}: \exists U\in\mathcal{U}_0(E)\, \exists B\in \mathcal{B}(E) : \lim_{n\to\infty} \frac{t_n}{\delta_n(B,U)} = 0 \right\}, \] where \(\mathcal{B}(E)\) is the collection of all bounded subsets of \(E\). For a non-negative increasing sequence \((\alpha_n)_{n\in\mathbb{N}}\) with \(\lim_{n\to\infty} \alpha_n = \infty\), the power-series spaces \[ \Lambda_1(\alpha) = \left\{(x_n)_{n\in\mathbb{N}}: \|(x_n)\|_k = \sup_{n\in\mathbb{N}} |x_n| \,\mathrm{e}^{-\alpha_n/k} < \infty \;\text{for all}\; k\in\mathbb{N}\right\} \] and \[ \Lambda_\infty (\alpha) = \left\{(x_n)_{n\in\mathbb{N}}: \|(x_n)\|_k = \sup_{n\in\mathbb{N}} |x_n| \,\mathrm{e}^{k \alpha_n} < \infty \;\text{for all}\; k\in\mathbb{N}\right\} \] play an important role in the theory of Fréchet spaces. The space \(\Lambda_\infty(\alpha)\) is called a power series space of infinite type and \(\Lambda_1(\alpha)\) a power series space of finite type. This article addresses the following question: Let \(E\) be a nuclear Fréchet space with the properties \textit{\underline{DN}} and \(\Omega\). Is the condition \(\Delta(E)=\Delta(\Lambda_r(\alpha))\) equivalent to \(\delta(E)=\delta(\Lambda_r(\alpha))\)? In the case of power series spaces of infinite type, the author gives a positive answer to the above question. More precisely, she shows that, if \(\varepsilon\) is the associated exponent sequence of a Fréchet space \(E\) with the properties \textit{\underline{DN}} and \(\Omega\), then \(\Delta(E)=\Delta(\Lambda_\infty(\varepsilon))\) if and only if \(\delta(E)=\delta(\Lambda_\infty(\varepsilon))\). In the case of power series spaces of finite type, it is shown that, in the above situation, \(\delta(E)=\delta(\Lambda_1(\varepsilon))\) implies that \(\Delta(E)=\Delta(\Lambda_1(\varepsilon))\). In addition, a~number of additional assumptions on \(E\) are presented under which for power series spaces of finite type these conditions are equivalent.
0 references
nuclear Fréchet spaces
0 references
diametral dimension
0 references
topological invariants
0 references
prominent bounded subsets
0 references