4
$\begingroup$

Let $$ \tilde{\mathbb C}={\mathbb C}\smallsetminus (-\infty,0] $$ the complex plane without the negative real axis. Let $V$ denote the set of all holomorphic functions $f:\tilde{\mathbb C}\to{ \mathbb C}$ such that $$ f^+(0)=\lim_{\lambda\searrow 0}f(\lambda) $$ exists. If $K\subset \{\mathrm{Re}(z)>0\}$ is a compact subset which is not finite, then a given $f\in V$ is uniquely determined by its restriction to $K$. Therefore it is reasonable to ask:

Does there exists a compact set $K$ and a complex valued Radon measure $\mu$ on $K$ such that $$ f^+(0)=\int_K f(\lambda)\,d\mu(\lambda) $$ holds for every $f\in V$?

Improve this question
$\endgroup$

2 Answers 2

Reset to default
4
$\begingroup$

No. The existence of such a representation would imply that the linear functional $\Phi:V\to \mathbb C$, $f\mapsto f^+(0)$ is a continuous with respect to the topology of uniform convergence on compact subsets of $K$. By Hahn-Banach one could thus extend $\Phi$ to $H(\tilde{\mathbb C})$. This is certainly not true: The functions $f_n(z)= \frac{1}{z+1/n}$ are in $V$ and converge in $H(\tilde{\mathbb C})$. But $\Phi(f_n)=n$ does not converge.

Improve this answer
$\endgroup$
2
  • 2
    $\begingroup$ Where do we use HB in this argument? All these functions belong to $V$ itself and they are uniformly bounded on $K$ while their limits in 0 are not, is not it already a contradiction? $\endgroup$
    – Fedor Petrov
    Commented Jul 13, 2016 at 8:43
  • 1
    $\begingroup$ Right. But Hahn-Banach was my first idea because it is ''obvious'' (even without looking at examples) that $\Phi$ cannot extend continuously to $H(\tilde{\mathbb C})$ ''because'' $f^+(0)$ need not exist for $f\in H(\tilde{\mathbb C})$. $\endgroup$
    – Jochen Wengenroth
    Commented Jul 13, 2016 at 8:53
1
$\begingroup$

Here is a simple proof without Hahn-Banach. I assume that the complex Radon measure is finite. Then your representation would imply $$|\lim_{\lambda\to 0}f(\lambda)|\leq C\max_{K}|f(z)|,$$ which is evidently not true.

Improve this answer
$\endgroup$

You must log in to answer this question.