Reinhardt domains and toric models. (Q1396344): Difference between revisions

Let \(T\) be the compact \(n\)-torus, \(T={\mathbb R}^n/{\mathbb Z}^n\) and \(\Omega \in {\mathbb R}^n\) a strictly convex domain. The set \(X=\Omega +\sqrt{-1}T=\{x+\sqrt{-1}y; \, x\in \Omega, y\in T\}\) is called \textit{Reinhardt domain} and is a \(T\)-manifold with the \(T\)-action on \(X\) given by addition on \(y\in T\). Let \(\omega \) be a \(T\)-invariant symplectic form on \(X\) and \(H_{\omega }\) an unitary \(T\)-representation. If \(H_{\omega }\) contains every irreducible \(T\)-representation exactly once, it is called a \textit{toric model} after \textit{I. M. Gelfand} and \textit{A. Zelevinski} [Funct. Anal. Appl. 18, 183--198 (1984; Zbl 0573.22008)]. In this paper, the Kähler \(T\)-invariant structures \(\omega \) on \(X\) are obtained with the expression \(\omega =\sqrt{-1}\partial {\bar \partial}F\) where \(\partial, {\bar \partial}\) are the Dolbeault operators on \(X\) and \(F\) is a strictly convex smooth function on \(\Omega \). The main result of paper show that the conditions for \(H_{\omega }\) to be a toric model are closely related to the convergence of the integrals \(\int_{x\in \Omega}e^{-F(x)+\lambda x}dV\) where \(\lambda \in {\mathbb Z}^n\) and \(dV\) is the Lebesgue measure on \({\mathbb R}^n\).