2

In the book "Uniform Algebras and Jensen Measures" by T.W. Gamelin, p.39, says:

$---------------------$

By the F.Riesz Theorem, any subharmonic function $u$ in a neighborhood of a compact set $K$ in $C$ can be expressed in the form

$$u(z)= v(z) + \int log|z- \zeta|d\tau(\zeta),\quad z \in K$$

where $\tau$ is a positive measure supported on a compact neighborhood of $K$, and $v$ is harmonic in a neighborhood of $K$.

$---------------------$

Where can I find a reference for this result, especially for a proof? Can anyone provide a proof? Thank you.

LDS
  • 81

2 Answers2

1

A proof for the fundamental theorem on subharmonic functions may be found in Masatsugu Tsuji's "On F. Riesz' fundamental theorem on subharmonic functions" (Tohoku Math. J. (2) 4(2): 131-140 (1952)).

Alex M.
  • 35,207
Fred
  • 77,394
  • Fantastic. Thank you. Meanwhile I should read the proof...do you think the analogous result holds in $R^d$, $d \ge 3$, where $log|z-\zeta|$ is replaced by the potential function on $R^d$ in the statement? – LDS Nov 28 '16 at 13:33
1

This question was posed a long time ago so I will answer it from another point of view.

If you're willing to accept some things about distributions you can see this by looking at the potential as a convolution: if $\tau$ is seen as a distribution with compact support and then we have

\begin{equation} p_\tau(z) = \int \log |z - w|\ d\tau(w) = \log|z|*\tau \end{equation}

Now using the fact that $\Delta\log|z| = 2\pi\delta$, we have that $\Delta p_\tau = \Delta(\log|z|*\tau) = 2\pi(\delta*\tau) = 2\pi\tau$.

We need to accept some facts:

If we accept this, then the result follows since $\tau=\frac{1}{2\pi}\Delta f$ is a radon measure and $h = f - p_\tau$ is a harmonic function such that

\begin{equation} f = h + p_\tau. \end{equation}

PS: We also have to check that $\operatorname{supp} \tau$ is compact, so you actually need to change $f$ a little, but it is not too problematic.

PSS: I think you wanted a more classical answer but since 2 years had passed, i figured why not.

Thiago
  • 768
  • Also this result should work in $\mathbb{R}$ by convolutioning with $F$ such that $\Delta F = \delta$. All you would need to check is that $\Delta f$ is positive for any sub-harmonic function on dimension $n$. – Thiago Jul 13 '19 at 19:24