Nevanlinna classes on the intersection of strictly pseudoconvex domains (Q1903621): Difference between revisions

Let \(\Omega\) be a transverse finite intersection of strictly pseudo-convex domains \(\Omega_i\) in \(\mathbb{C}^n\), \(i=1, \dots, N\); \(\Omega_i\) are starlike with respect to 0, bounded, with \(C^2\) regular border and defined by \(\Omega_i= \{z\in\mathbb{C}^n: \rho_i (z)<0\}\), \(\rho_i\)-strictly plurisubharmonic \((s psh)\), \(C^2\) or \(C^3\) in a neighbourhood of \(\overline \Omega_i\), with \(d\rho_i \neq 0\) in a neighbourhood of \(\partial \Omega_i\). Author proves that the function \(\rho\) defined by \(\rho^{-1}= \Sigma_i \rho_i^{-1}\) has the same regularity as \(\rho_i\), the behaviour of the boundary-distance function and is \(s psh\) in \(\Omega\). Two Nevanlinna classes of \(psh\) functions \(W\) in \(\Omega\) are considered: \[ N_1 (\Omega)= \Bigl\{W: \sup_{\varepsilon>0} \int_{\partial \Omega_\varepsilon} W^+i \overline \partial \rho \wedge \beta^{n-1} <+ \infty\Bigr\},\;\beta= {i\over 2} \partial\overline \partial|z |^2,\;\Omega_\varepsilon= \bigl\{z\in \Omega: \rho(z)< -\varepsilon\bigr\} \] and \[ N_2 (\Omega)= \Bigl\{W: \sup_{\varepsilon>0} \int_{\partial \Omega_\varepsilon} W^+ i\overline \partial\rho \wedge (i\partial \overline \partial\rho)^{n-1} <+\infty\Bigr \}. \] An example proves that Blaschke condition for \(\Theta=i\partial\overline\partial W\) is not necessary in order that \(W\in N_1 (\Omega)\). The main theorems \(A\) and \(B\) provide sufficient Blaschke type conditions for a positive closed \((1,1)\)-current \(\Theta\), such that the equation \(\Theta=i \partial\overline \partial W\) has a solution \(W\) in \(N_j (\Omega)\), \(j=1,2\) respectively. Thus Theorem A shows that \(W\in N_1 (\Omega)\) if \(\Theta\) satisfies \(\int_\Omega- \rho i\partial \overline \partial \rho \wedge \Theta \wedge \beta^{n-2} <+\infty\). The solution \(W\) is obtained by solving the equations \(idw= \Theta\) and \(\overline \partial U=w_{0,1}=\) the \((0,1)\) component of \(w\). Then \(W=U+ \overline U\), and since \(w_{0,1}\) satisfies ``good estimations'' \(W\in N_1 (\Omega)\). Many techniques are ingeniously combined, e.g. Fornaess's support functions and weighted kernels of Berndtsson and Andersson to solve \(\overline \partial u=w\), change of variables of type Range-Siu to estimate \(u\), homotopy to obtain \(w_{0,1}\), regularization. Similarly, Theorem B gives conditions for \(N_2 (\Omega)\), however only in the case \(N=2\).