Bergman-Shilov Boundary and Peak Points

581 Views Asked by At

Let $\Omega$ be a domain in $\mathbb{C}^n.$ Consider the Banach algebra $A(\Omega):=\mathcal{C}({\overline{\Omega}})\cap\mathcal{O}(\Omega).$ Denote the Bergman-Shilov boundary of $A(\Omega)$ by $\partial_S(\Omega)$ . From the very definiton of peak points we know that every peak point belongs to $\partial_S(\Omega).$ Now the examples that I know the set of peak points coincides with the $\partial_S(\Omega).$ My question here is if there are domains for which the set of peak points are properly contained in $\partial_S(\Omega)$?

1

There are 1 best solutions below

2
On

You ask difficult questions, and it's been a while since I've thought about peak points. I'm not sure if this is open or not, at least for smoothly bounded domains. Replacing $A(\Omega)$ with $A^1(\Omega) = C^1(\bar\Omega) \cap \mathcal{O}(\Omega)$ however, there are known examples.

Note that the Shilov boundary is closed by definition. Also, for a smoothly bounded pseudoconvex domain $\Omega$, it's well known that every strictly pseudoconvex point is a peak point for $A(\Omega)$ (in fact for $A^\infty(\Omega)$).

On the other hand, Kohn and Nirenberg (1973) constructed an example of a bounded pseduconvex domain $\Omega$ in $\mathbb{C}^2$ such that $\partial\Omega$ is strictly pseudoconvex at every point expect a point $P$ that doesn't admit a peak function that is holomorphic on a neighbourhood of $P$. This was later sharpened by several people to produce a smoothly bounded pseudoconvex domain with strictly pseudoconvex boundary except at one single point, admitting no peak function in $A^1(\Omega)$ at the bad point. Hence the Shilov boundary is the entire boundary, but the set of peak points for $A^1$ is a proper subset.

You should read Noell, Alan: "Peak points for pseudoconvex domains: a survey." J. Geom. Anal. 18 (2008), no. 4, 1058–1087 if you haven't already.

I'll think a bit about the non-smooth case and get back to you if I come up with something.


By the way, you may want to accept some of the answers you have received to your previous questions. This shows your appreciation for the people who have helped you.