It's easy to see that the Phargmen-Lindelöf theorem from complex analysis can be generalized to non-simply-connected regions. Namely to regions $G$ with the property that for each $z \in \partial_\infty G $ there is a sphere $V$ in $\mathbb{C}_\infty $ centered at $z$ such that $V \cap G$ is simply connected.

The problem is that I can't think about any example of non-simply connected regions that have this property and of simply-connected regions that don't have this property...

Does the exterior of a unit ball can be considered as an example of non-simply connected region that don't satisfy the property above? What about regions that do satisfy this property?

BTW - $ \partial _ \infty $ is the boundary of $G$ with the boundary at infinity (and $\mathbb{C}_\infty$ is the Riemann-Sphere ) .

Thanks in advance

  • 2
    $\begingroup$ A sphere in $C^\infty$? What's that? $\endgroup$ – Marc Palm Jul 18 '12 at 17:04
  • $\begingroup$ It's $C$ with infinity... $\endgroup$ – Jason Mraz Jul 18 '12 at 17:07
  • 1
    $\begingroup$ Firstly, there are a lot of results in complex analysis labeled "Phragmen-Lindelöf Theorems". It might be useful here to indicate which Theorem you are refering to.... Secondly, I'm not sure I understand the question... Doesn't an annulus, for example, satisfy the property? $\endgroup$ – Malik Younsi Jul 18 '12 at 18:09
  • 1
    $\begingroup$ By "a sphere V in $\mathbb{C}_{\infty}$" did you mean a circle, i.e. the boundary of a disc? $\endgroup$ – Kevin Jul 18 '12 at 20:59
  • 1
    $\begingroup$ For an example of unbounded non simply-connected region which satisfies your property, just take the complement in $\mathbb{C}_{\infty}$ of two disjoint closed disks... $\endgroup$ – Malik Younsi Jul 18 '12 at 22:53

Phragmen-Lindelof theorem has nothing to do with simple or multiple connectedness. The general formulation is the following. Let D be an region on the Riemann sphere. Let $w_0$ be a boundary point of D. Suppose that u is a subharmonic function in D which is bounded from above. Suppose that $$\limsup_{z\to w} u(z)\leq 0$$ for all boundary points $w$ of $D$, except $w_0$. Then $u\leq 0$.

One point $w_0$ can be replaced by any closed set of zero logarithmic capacity. But the original argument of Phragmen and Lindelof was for one point, and $u=\log|f|$ where $f$ is analytic.

For the proof, assume first that $D$ is contained in the unit disc, and $w_0=0$. Apply the Maximum principle to $v=u+\epsilon\log|z|$. You obtain that $v\leq 0$. Now fix $z$ and let $\epsilon\to 0$. Generalization to arbitrary $D$ is trivial.

| cite | improve this answer | |

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.