Dirichlet space

In mathematics, the Dirichlet space on the domain \Omega \subseteq \mathbb{C}, \, \mathcal{D}(\Omega) (named after Peter Gustav Lejeune Dirichlet), is the reproducing kernel Hilbert space of holomorphic functions, contained within the Hardy space H^2(\Omega), for which the Dirichlet integral, defined by

 \mathcal{D}(f) := {1\over \pi} \iint_\Omega |f^\prime(z)|^2 \, dA = {1\over 4\pi}\iint_\Omega |\partial_x f|^2 + |\partial_y f|^2 \, dx \, dy

is finite (here dA denotes the area Lebesgue measure on the complex plane \mathbb{C}). The latter is the integral occurring in Dirichlet's principle for harmonic functions. The Dirichlet integral defines a seminorm on \mathcal{D}(\Omega). It is not a norm in general, since \mathcal{D}(f) = 0 whenever f is a constant function.

For f,\, g \in \mathcal{D}(\Omega), we define

\mathcal{D}(f, \, g) : = {1\over \pi} \iint_\Omega f'(z) \overline{g'(z)} \, dA(z).

This is a semi-inner product, and clearly \mathcal{D}(f, \, f) = \mathcal{D}(f). We may equip \mathcal{D}(\Omega) with an inner product given by

 \langle f, g \rangle_{\mathcal{D}(\Omega)} := \langle f, \, g \rangle_{H^2 (\Omega)} + \mathcal{D}(f, \, g) \; \; \; \; \; (f, \, g \in \mathcal{D}(\Omega)),

where  \langle \cdot, \, \cdot \rangle_{H^2 (\Omega)} is the usual inner product on H^2 (\Omega). The corresponding norm  \| \cdot \|_{\mathcal{D}(\Omega)} is given by

 \|f\|^2_{\mathcal{D}(\Omega)} := \|f\|^2_{H^2 (\Omega)} + \mathcal{D}(f) \; \; \; \; \; (f \in \mathcal{D} (\Omega)).

Note that this definition is not unique, another common choice is to take  \|f\|^2 = |f(c)|^2 + \mathcal{D}(f), for some fixed  c \in \Omega .

The Dirichlet space is not an algebra, but the space \mathcal{D}(\Omega) \cap H^\infty(\Omega) is a Banach algebra, with respect to the norm

 \|f\|_{\mathcal{D}(\Omega) \cap H^\infty(\Omega)} := \|f\|_{H^\infty(\Omega)} + \mathcal{D}(f)^{1/2} \; \; \; \; \; (f \in \mathcal{D}(\Omega) \cap H^\infty(\Omega)).


We usually have \Omega = \mathbb{D} (the unit disk of the complex plane \mathbb{C}), in that case \mathcal{D}(\mathbb{D}):=\mathcal{D}, and if

 f(z) = \sum_{n \ge 0} a_n z^n \; \; \; \; \; (f \in \mathcal{D}),

then

 D(f) =\sum_{n\ge 1} n |a_n|^2,

and

 \|f \|^2_\mathcal{D} = \sum_{n \ge 0} (n+1) |a_n|^2.

Clearly, \mathcal{D} contains all the polynomials and, more generally, all functions f, holomorphic on \mathbb{D} such that f' is bounded on \mathbb{D}.

The reproducing kernel of \mathcal{D} at w \in \mathbb{C} \setminus \{ 0 \} is given by

 k_w(z) = \frac{1}{z\overline{w}} \log \left( \frac{1}{1-z\overline{w}} \right) \; \; \; \; \; (z \in \mathbb{C} \setminus \{ 0 \}).

See also

References

This article is issued from Wikipedia - version of the Thursday, July 30, 2015. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.