In page 101, corollary 4.3.3., from Armitage and Gardiner's book on potential theory, the authors prove that any subharmonic function, can be identified with a positive measure (Riesz measure).

In doing so, they consider the functional $L_s(\phi)=\int_\Omega s(x)\Delta \phi(x)$, $\Omega$ a open set, $\phi\in C_0^\infty(\Omega)$ and $s$ the subharmonic function.

For any $\phi\in C_0(\Omega)$, he consider the mollified sequence $\phi_n$, which converges to $\phi$ uniformly and claim that $L_s(\phi_n)$ is Cauchy. Maybe, as he says, this is evident, however I fail to see it.

If, for instance the regularized sequence $s_n$, is such that, $\Delta s_n$ is bounded uniformly in $L^1_{loc}(\Omega)$ then, the above is true, however, I also fail to see it. Any help is appreciated.

Remark: I know how to extend $L_s$ to $C_0(\Omega)$, but with another approach, although the extension is the same, because of unicity.

Update: For $x\in \Omega$, fix some $r>0$ such that $\overline{B(x,r)}\subset \Omega$.The regularization $s_n$ of $s$, is a sequence of subharmonic functions in $B(x,r)$ which converges decreasingly to $u$ in $B(x,r)$. Note that (Green's theorem)

\begin{eqnarray} \int_{B(x,r)}|\Delta s_n| &=& \int_{B(x,r)} \Delta s_n \nonumber \\ &=& \int_{\partial B(x,r)} \frac{\partial s_n}{\partial\eta} \nonumber \\ &=& r^{N-1}\frac{\partial}{\partial r}\left(\frac{1}{r^{N-1}}\int_{\partial B(x,r)}s_n\right) \end{eqnarray}

Does anyone knows how to bound the right hand side uniformly in $n$? Also, note that the right hand side, can be written in terms of the average integral of $s_n$.


This is the "continuity from positivity" trick. First, recall that the authors use $C_0$ to denote the space of compactly supported functions, not those vanishing on the boundary. Since $\phi$ is compactly supported, so are $\phi_n$ for large $n$, and we can take a compact set $K\subset \Omega$ such that all large $n$ the functions $\phi_n$ and $\phi$ vanish outside of $K$.

Let $\chi\in C_0^\infty(\Omega)$ be a nonnegative function such that $\chi\equiv 1$ on $K$. Let $\delta_{nm}=\sup|\phi_n-\phi_m|$. Then $$- \delta_{nm} \chi \le \phi_n-\phi_m\le \delta_{nm} \chi$$ everywhere in $\Omega$. Since $L_s$ is a positive functional, it respects pointwise inequalities; thus $$-\delta_{nm} L_s(\chi) \le L_s(\phi_n)-L_s(\phi_m)\le \delta_{nm} L_s(\chi)$$ Here $L_s(\chi)$ is a fixed number, independent of $m,n$. Since $\delta_{nm}\to 0$ as $m,n\to\infty$, the conclusion follows.

  • $\begingroup$ I was going so far to prove a simple thing... $\endgroup$ – Tomás Jul 28 '14 at 22:09
  • $\begingroup$ Tomás. I can answer you old question: how to bound the integral. Since the sequence $(s_{n})$ is decreasing, $s_{n}\leq s_{0}$. But how did you go from the 2end equality to the third one? $\endgroup$ – M. Rahmat Jul 1 at 23:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.