Hilbert Space of Complex-Valued Harmonic Functions in the Unit Disc

Tseganesh Getachew Gebrehana and Hunduma Legesse Geleta

arXiv: 2410.22045v1 · PDF · 29 October 2024
Categories: math.CV
Licence: CC BY 4.0 The site shows the text.

Abstract

We investigate an extended version of Hilbert space of analytic functions called Hilbert space of complex-valued harmonic functions. It is found that functions in Hilbert space of complex-valued harmonic functions exhibit many properties analogous to its analytic counter part such as complex-valued harmonic function analogous of norm, equivalent norms, reproducing kernels, growth estimates and Littlewood-Paley Identity Theorem. In conclusion we prove that many results in Hilbert space of analytic functions also hold in larger Hilbert space of complex-valued harmonic functions.

Items 14

Theorem

Theorem 2.1

PQFA

Let f be a complex-valued harmonic function on the unit disc \mathbb{D} given by

f(z)=h(z)+\overline{g(z)}

where h(z)=\sum_{n=0}^{\infty} a_{n}z^{n} and g(z)=\sum_{n=0}^{\infty} {b_{n}z^{n}} are analytic. Suppose

\mathnormal{H}_{h}^{2}(\mathbb{D})= \lbrace f:\mathbb{D} \rightarrow \mathbb{C}: f(z)= \sum_{n=0}^{\infty} a_{n}z^{n} + \sum_{n=0}^{\infty} \overline{b_{n}z^{n}} with \sum_{n=0}^{\infty} |a_{n}|^{2}+|b_{n}|^{2} <\infty \rbrace.

Then ||.||_{\mathnormal{H}_{h}^{2}}: {\mathnormal{H}_{h}^{2}}(\mathbb{D}) \rightarrow \mathbb{R} defined by ||f||_{\mathnormal{H}_{h}^{2}}=({\sum_{n=0}^{\infty}|a_{n}|^{2} + |b_{n}|^{2}})^{\frac{1}{2}} is a norm.

Proof of Theorem 2.1

(i). By definition, ||f||_{{H}_{h}^{2}}^{2} \geq 0.
(ii). ||\alpha f||_{{H}_{h}^{2}}^{2}= \sum_{n=0}^{\infty}|\alpha a_{n}|^{2} + |\alpha b_{n}|^{2}=|\alpha|^{2} \sum_{n=0}^{\infty} |a_{n}|^{2} + |b_{n}|^{2}= |\alpha|^{2}||f||_{{H}_{h}^{2}}^{2}. Thus, ||\alpha f||_{{H}_{h}^{2}}=|\alpha|||f||_{{H}_{h}^{2}}.
(iii.) ||f+F||_{{H}_{h}^{2}}^{2} = \sum_{n=0}^{\infty} |a_{n}+A_{n}|^{2} + |b_{n}+B_{n}|^{2}
= \sum_{n=0}^{\infty} |a_{n}|^{2}+ |A_{n}|^{2} + 2\Re (a_{n}\overline{A_{n}}) +|b_{n}|^{2}+|B_{n}|^{2}+ 2\Re (b_{n}\overline{B}_{n})
\leq ||f||_{{H}_{h}^{2}}^{2} + ||F||_{{H}_{h}^{2}}^{2}+2||f||_{{H}_{h}^{2}}||F||_{{H}_{h}^{2}}=(||f||_{{H}_{h}^{2}}+||F||_{{H}_{h}^{2}})^{2}.
This implies that
||f+F||_{{H}_{h}^{2}} \leq ||f||_{{H}_{h}^{2}}+||F||_{{H}_{h}^{2}}.

Therefore,

||f||_{{H}_{h}^{2}}(\mathbb{D}) is a norm on \mathnormal{H}_{h}^{2}(\mathbb{D}).

Theorem

Theorem 2.2

0JGW

Let \mathnormal{H}_{h}^{2} be the space of complex-valued harmonic functions on the unit disc \mathbb{D}:=\lbrace z \in \mathbb{C} : |z|<1 \rbrace whose Taylor series representation has square-summable coefficients. Then \mathnormal{H}_{h}^{2} is a Hilbert space.

Proof of Theorem 2.2

Since the sequences \lbrace a_{n} \rbrace_{n=0}^{\infty} and \lbrace b_{n} \rbrace_{n=0}^{\infty} of power series coefficients belong to the Hilbert space \mathnormal{l^2}(\mathbb{Z^+}), then the sequence \lbrace c_{n} \rbrace_{n=0}^{\infty} defined by c_{2n} = a_n for each n= 0, 1, 2, ... and c_{2n-1}= b_{n-1} for each n =1, 2, 3, ... belongs to the Hilbert space \mathnormal{l^2}(\mathbb{Z^+}). Conversely, every sequence in \mathnormal{l^2}(\mathbb{Z^+}) defines a complex-valued harmonic function on the open unit disk belongs to \mathnormal{H_h^2}(\mathbb{D}) by means of the map \lbrace c_{n} \rbrace _{n=0}^{\infty} \rightarrow \sum_{n=0}^{\infty} a_{n}z^{n}+\sum_{n=0}^{\infty} \overline{b_{n}z^{n}}. Thus, \mathnormal{H_h^2}(\mathbb{D}) is isometrically isomorphic to \mathnormal{l^2}(\mathbb{Z^+}). Thus we conclude that \mathnormal{H_h^2}(\mathbb{D}) is a Hilbert space.

Theorem

Theorem 3.1

9Q3S

The norm defined on \mathnormal{H}_{h}^{2} (in Theorem 2.1) has another equivalent representation by integral means which is denoted by \mathnormal{M}_{2}^{2} (f,r) and defined by

\mathnormal{M}_{2}^{2} (f,r) = \frac{1}{2\pi} \int_{-\pi}^{\pi} |f(re^{i\theta})|^{2} d\theta,

where f(z)=\sum_{n=0}^{\infty} a_{n}{z}^{n}+\sum_{n=0}^{\infty} \overline{ b_{n}{z}^{n}} on \mathbb{D} and 0 \leq r < 1.

Proof of Theorem 3.1

Now using polar form of the series representation of f, we get

f(re^{i\theta} )=\sum_{n=0}^{\infty} a_{n}{r}^{n}{e^{in\theta}}+\sum_{n=0}^{\infty} \overline{ b_{n}{r}^{n}{e^{in\theta}}}.

After some algebraic manipulation, we obtain

|f(re^{i\theta})|^{2}=\sum_{n=0}^{\infty} \sum_{m=0}^{\infty} a_{n}\overline{a_{m}}{r}^{n+m}{e^{i(n-m)\theta}} +\sum_{n=0}^{\infty}\sum_{m=0}^{\infty} a_{m}b_{n}{r}^{m+n}{e^{i(n+m)\theta}} + \sum_{n=0}^{\infty}\sum_{m=0}^{\infty}\overline{ a_{m}b_{n}{r}^{m+n}{e^{i(n+m)\theta}}}+\sum_{n=0}^{\infty} \sum_{m=0}^{\infty} b_{m}\overline{b_{n}}{r}^{n+m}{e^{i(n-m)\theta}}.

It is clear that the integral of exponential function \lbrace e^{in\theta} \rbrace_{n=0}^{\infty} is 2\pi when n=m and 0 when n \neq m. Multiplying both sides of the above equation by \frac{1}{2\pi} and integrating with respect to \theta from -\pi to \pi, we get

\frac{1}{2\pi}\int_{-\pi}^{\pi}|f(re^{i\theta})|^{2} d\theta=\sum_{n=0}^{\infty} |a_{n}|^{2}{r}^{2n} +\sum_{n=0}^{\infty} |b_{n}|^{2}{r}^{2n} +\frac{1}{2\pi}\int_{-\pi}^{\pi} \sum_{n=0}^{\infty}\sum_{n=0}^{\infty} 2\Re ({a_{n}b_{n}{r}^{2n}{e^{2in\theta}}})d\theta

=\sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n} + \sum_{n=0}^{\infty}\sum_{n=0}^{\infty} \frac{2{r}^{2n}}{2\pi} \Re ({a_{n}b_{n}\int_{-\pi}^{\pi} {e^{2in\theta}}}d\theta)

=\sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n}

Therefore,

M_{2}^{2} (f,r)=\frac{1}{2\pi}\int_{-\pi}^{\pi}|f(re^{i\theta})|^{2}d\theta=\sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n}.

To complete the proof we need to show ||f||_{\mathnormal{H}_{h}^2}= \lim_{r \rightarrow 1^{-}} M_{2} (f,r). So, from the above equation we have

M_{2}^{2}(f,r)=\sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n}

\leq \sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2})=||f||_{\mathnormal{H}_{h}^{2}}^{2},

whenever f \in {\mathnormal{H}_{h}^{2}} and 0 \leq r<1. So M_{2} (f,r) is bounded by the \mathnormal{H}_{h}^{2}-norm. It remains to show that whenever \lim_{r \rightarrow 1^{-}}M_{2} (f,r)=\mathnormal{M}< \infty, then the partial sum of the series of the form M_{2}^{2} (f,r)=\sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n} are bounded by \mathnormal{M^2}:

\sum_{n=0}^{\mathnormal{N}} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n} \leq \sum_{n=0}^{\infty} (|a_{n}|^{2} + |b_{n}|^{2}){r}^{2n} \leq \mathnormal{M^2}.

As r \rightarrow 1^{-}, these partial sums converges to functions in \mathnormal{H}_{h}^{2}, which must therefore be bounded by \mathnormal{M^2} as well. If every partial sum of Taylor series representation of functions in \mathnormal{H}_{h}^{2} is bounded by \mathnormal{M^2} , then this is also true for the entire series. This completes the proof.

Corollary

Corollary 3.2

VIAU

The space of bounded complex-valued harmonic functions on the unit disc \mathnormal{H}_{h}^{\infty} is a subset of \mathnormal{H}_{h}^2.

Proof of Corollary 3.2

Let f \in \mathnormal{H}_{h}^\infty(\mathbb{D}). Then ||f||_{\mathnormal{H}_{h}^\infty} = \sup_{z \in \mathbb{D}}|f(z)|. Now, \frac{1}{2\pi}\int_{-\pi}^{\pi}|f(re^{i\theta})|^{2}d\theta \leq \frac{1}{2\pi}\int_{-\pi}^{\pi} \sup |f(re^{i\theta})|^{2}d\theta \newline =\frac{1}{2\pi}\int_{-\pi}^{\pi}||f||_{\mathnormal{H}_{h}^{\infty}}^{2}d\theta=||f||_{\mathnormal{H}_{h}^{\infty}}^{2}, which holds true for every 0<r<1. So for any f \in {\mathnormal{H}_{h}^{\infty}} we get

\lim_{r \rightarrow 1^{-}}M_{2} (f,r) \leq ||f||_{\mathnormal{H}_{h}^{\infty}}^{2}.

Hence, f \in \mathnormal{H}_{h}^2.

Theorem

Theorem 3.3

4V13

For every complex-valued harmonic functions f \in \mathnormal{H}_{h}^2 on the unit disc we have

||f||_{\mathnormal{H}_{h}^2}=|f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2} \log \frac{1}{|z|} dA(z),

where d\mathnormal{A} denotes the normalized Lebesgue measure on \mathbb{D} (d\mathnormal{A}=\frac{1}{\pi} dxdy=\frac{1}{\pi} r dr d\theta).

Proof of Theorem 3.3

We start by considering the right hand side of equation of Theorem 3.3., and converting f into polar form, we obtain

\begin{aligned} |f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2}\log \frac{1}{|z|} d\mathnormal{A} =|f(0)|^{2}+2\int_{-\pi}^{\pi} \frac{1}{\pi} \int_{0}^{1}|f'(re^{i\theta})|^{2} (\log \frac{1}{r})r dr d\theta \end{aligned}

Interchanging the two integrals which can be justified by Fubini’s theorem; we have the following

\begin{aligned} |f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2}\log \frac{1}{|z|} d\mathnormal{A} =|f(0)|^{2}+2\int_{0}^{1} (\frac{1}{\pi} \int_{-\pi}^{\pi}|f'(re^{i\theta})|^{2} d\theta) (\log \frac{1}{r})r dr \end{aligned}

Applying simple algebraic manipulations, we obtain

\begin{aligned} |f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2}\log \frac{1}{|z|} d\mathnormal{A}=|f(0)|^{2}+4\int_{0}^{1} \mathnormal{M}_{2}^{2}(f', r) (\log \frac{1}{r})r dr \end{aligned}

Replacing by the Taylor series representation, we get

\begin{aligned} |f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2}\log \frac{1}{|z|} d\mathnormal{A} =|f(0)|^{2}+4\int_{0}^{1} \sum_{n=1}^{\infty} (|n|^{2}|a_{n}|^{2}+ |n|^{2}|b_{n}|^{2}) r^{(2n-2)} (\log \frac{1}{r})r dr\\ =|f(0)|^{2}+4\sum_{n=1}^{\infty} |n|^{2}(|a_{n}|^{2}+ |b_{n}|^{2}) \int_{0}^{1} r^{(2n-1)} (\log \frac{1}{r})r dr \\ =|f(0)|^{2}+4\sum_{n=1}^{\infty} n^{2}(|a_{n}|^{2} + |b_{n}|^{2}) \frac{1}{4n^{2}}\\ = |f(0)|^{2}+\sum_{n=1}^{\infty} (|a_{n}|^{2}+ |b_{n}|^{2})=||f||_{\mathnormal{H}_{h}^{2}}. \end{aligned}

From which we obtain,

|f(0)|^{2}+2\int_{\mathbb{D}}|f'(z)|^{2}\log \frac{1}{|z|} d\mathnormal{A}=||f||_{\mathnormal{H}_{h}^{2}}.

This completes the proof.

Theorem

Theorem 4.1

LRHB

(Growth estimate). For each z \in \mathbb{D}, the growth estimate of f \in \mathnormal{H}_{h}^{2} is given by

|f(z)| \leq \frac{{2}||f||_{\mathnormal{H}_{h}^{2}}}{\sqrt{1-|z|^{2}}}.

Proof of Theorem 4.1

By applying the triangle inequality for the modulus and Cauchy-Schwartz inequality to the complex-valued harmonic function of f in Theorem 3.4 we obtain for each z \in \mathbb{D},

\begin{aligned} |f(z)|=|\sum_{n=0}^{\infty}a_{n}z^{n}+\sum_{n=0}^{\infty} \overline{b_{n}z^{n}}| \leq \sum_{n=0}^{\infty}|a_{n}||z^{n}| +\sum_{n=0}^{\infty} |{b_{n}||z^{n}}|\\ \leq (\sum_{n=0}^{\infty}|a_{n}|^{2})^{\frac{1}{2}} (\sum_{n=0}^{\infty}|z^{n}|^{2})^{\frac{1}{2}} +(\sum_{n=0}^{\infty} |b_{n}|^{2})^{\frac{1}{2}} (\sum_{n=0}^{\infty}|z^{n}|^{2})^{\frac{1}{2}} \\ \leq [(\sum_{n=0}^{\infty}|a_{n}|^{2})^{\frac{1}{2}} + (\sum_{n=0}^{\infty} |b_{n}|^{2})^{\frac{1}{2}}](\sum_{n=0}^{\infty}|z|^{2n})^{\frac{1}{2}}\\ \leq (||h||_{\mathnormal{H}_{h}^{2}} +||g||_{\mathnormal{H}_{h}^{2}})(\sum_{n=0}^{\infty}|z|^{2n})^{\frac{1}{2}} \leq \frac{||h||_{\mathnormal{H}_{h}^{2}} +||g||_{\mathnormal{H}_{h}^{2}}}{\sqrt{1-|z|^{2}}}. \end{aligned}

But then again, ||h||_{\mathnormal{H}_{h}^{2}} \leq ||f||_{\mathnormal{H}_{h}^{2}}; ||g||_{\mathnormal{H}_{h}^{2}} \leq ||f||_{\mathnormal{H}_{h}^{2}} implying that,

|f(z)| \leq \frac{2||f||_{\mathnormal{H}_{h}^{2}}}{\sqrt{1-|z|^{2}}}.

Theorem

Theorem 4.2

IH8H

(Kernel) The reproducing kernel function K_{\alpha} , for a point \alpha in \mathbb{D} is defined by

K_{\alpha}(z)=\frac{1}{1-\overline{\alpha}{z}}+ \frac{1}{1-\alpha \overline{z}} \hspace{2mm}for |\overline{\alpha} z|<1.

Proof of Theorem 4.2

For a point \alpha \in \mathbb{D}, K_{\alpha} is the functional in \mathnormal{H}_{h}^{2}(\mathbb{D}) such that for all f in \mathnormal{H}_{h}^{2}(\mathbb{D}), we have

\langle f, K_{\alpha} \rangle=f(\alpha)

where f and K_{\alpha} are in \mathnormal{H}_{h}^{2}. Let f(z)=\sum_{n=0}^{\infty}a_{n}z^{n}+\sum_{n=0}^{\infty} \overline{b_{n}z^{n}} and K_{\alpha}(z)=\sum_{n=0}^{\infty}c_{n}z^{n}+\sum_{n=0}^{\infty} \overline{d_{n}z^{n}} for some coefficients. Thus for each f \in \mathnormal{H}_{h}^{2},

\sum_{n=0}^{\infty}a_{n}\alpha^{n}+\sum_{n=0}^{\infty} \overline{b_{n}\alpha^{n}} = f(\alpha) = \langle f, K_{\alpha} \rangle = \sum_{n=0}^{\infty}a_{n}\overline{c_{n}}+\overline{b_{n}}d_{n}.

This implies that

\sum_{n=0}^{\infty}a_{n}\alpha^{n}+\sum_{n=0}^{\infty} \overline{b_{n}} \hspace{1.5mm}\overline{\alpha^{n}} =\sum_{n=0}^{\infty}a_{n}\overline{c_{n}}+ \sum_{n=0}^{\infty}\overline{b_{n}} \hspace{1.5mm}d_{n}.

So this can be true if the following holds true

\alpha^{n}=\overline{c_{n}} and \overline{\alpha^{n}}= d_{n}.

Thus,

K_{\alpha}(z)=\sum_{n=0}^{\infty}\overline{\alpha^{n}}{z^{n}}+\sum_{n=0}^{\infty}{\alpha^{n}} \overline{z^{n}}

=\sum_{n=0}^{\infty}(\overline{\alpha}z)^{n}+\sum_{n=0}^{\infty} (\alpha\overline{z})^{n}

=\frac{1}{1- \overline{\alpha} z}+ \frac{1}{1-\alpha \overline{z}} .

Therefore, the required result holds true for |\overline{\alpha} z|<1.

Source and licence

The text on this page is from arXiv:2410.22045v1 by Tseganesh Getachew Gebrehana and Hunduma Legesse Geleta, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2410/2410.22045.jsonl.