On the solution of the Dirichlet problem with rational holomorphic boundary data (Q2492091): Difference between revisions

The following answer on the conjecture by D. Khavinson and H. Shapiro (1992) is given: \textbf{Theorem}. Let \(\Omega\) be a simply connected domain in \({\mathbb C}\) with a smooth real-analytic boundary, and \(a, a_1, a_2\) be any distinct fixed points of \(\Omega\). Let \(K(z, w)\) be the Bergman kernel associated with \(\Omega\). Then the following are equivalent: \noindent (i) the solution to the Dirichlet problem \[ \left\{ \begin{matrix} \triangle u = 0 & {\text{in}} & \Omega, \\ u(z, \overline{z}) = h(z, \overline{z}) & {\text{on}} & \partial\, \Omega, \end{matrix} \right.\tag{1} \] is rational for \(h(z, \overline{z}) = \frac{1}{z - a}\); \noindent (ii) the solution to (ref {1}) is rational for \(h(z, \overline{z}) = R(z)\) with \(R(z)\) being any rational function without poles on \(\partial\, \Omega\); \noindent (iii) the Riemann mapping \(f : \Omega \rightarrow {\mathbb D}\) is rational (\({\mathbb D}\) is the unit disc); \noindent (iv) \(K(z, a_1)\) and \(K(z, a_2)\) are rational functions of \(z\); \noindent (v) \(K(z, w)\) is a rational function of \(z\), and \(\overline{w}\).