A new algorithmic method for finding the Carathéodory extremal maps (Q953512)

From MaRDI portal





scientific article; zbMATH DE number 5362173
Language Label Description Also known as
default for all languages
No label defined
    English
    A new algorithmic method for finding the Carathéodory extremal maps
    scientific article; zbMATH DE number 5362173

      Statements

      A new algorithmic method for finding the Carathéodory extremal maps (English)
      0 references
      0 references
      6 November 2008
      0 references
      The paper is devoted to the extremal problem: find the analytic map \(f\) from a bounded domain \(D\subset\mathbb C^n\), \(0\in D\), into the unit ball \(B_n\subset\mathbb C^n\) with \(f(0)=0\) such that the value of \(| \det Jf(0)| \) is as large as possible. This value is called the \(C\)-extremal value \(J(D)\) and this map is called \(C\)-extremal map for \(D\). The author introduces a novel algorithmic method for finding a \(C\)-extremal map for a complex ellipsoid. Let \(D\) be a balanced bounded domain, for which the minimal circumscribed Hermitian ellipsoid is of the form \(\{(z_1,\dots,z_n)\in\mathbb C^n: a_1| z_1| ^2+\dots+a_n| z_n| ^2\}\), \(a_j>0\) for each \(1\leq j\leq n\). Denote \[ x_j=| z_j| ^2,\quad S:=\{x\in\mathbb R^n: f(x_1,\dots,x_n)=0\}, \quad H:=\biggl\{x\in\mathbb R^n: \sum_{i=1}^na_ix_i=1\biggr\}. \] \[ f_j:=\frac{\partial f}{\partial x_j},\qquad f_{ij}:=\frac{\partial^2f}{\partial x_i\partial x_j}. \] Split into the following two cases: (C1) \(f_{jj}\neq0\) for any \(1\leq j\leq n\); (C2) If \(f_{jj}=0\) for some \(j\), after rearranging, assume that there exists a \(\nu\) such that \(f_{jj}\neq0\) for \(1\leq j\leq\nu\) and \(f_{jj}=0\) for \(\nu+1\leq j\leq n\). The extremal problem is reduced to find the condition for \(F:=\prod_{j=1}^na_j^{g_j}\), \(g_j>0\), having the local extremal values when \(S\) and \(H\) meet tangentially. The main result is given in Theorem 2.6: Assume that \(f(x_1,\dots,x_n)\) is decoupled, so that \(f_{ij}=0\) for \(i\neq j\), and \(f_j\neq0\) for each \(j\). (i) If \(f\) satisfies (C1), then \(F\) has a local extremal point at \((x_1,\dots,x_n)\) only when \[ \frac{x_1f_1}{g_1}=\dots=\frac{x_nf_n}{g_n}\quad \text{or} \quad\sum_{j=1}^n\frac{(f_j)^2}{f_{jj}}=0. \] (ii) If \(f\) satisfies (C2), then \(F\) has a local extremal point at \((x_1,\dots,x_n)\) only when \[ \frac{x_1f_1}{g_1}=\dots=\frac{x_{\nu}f_{\nu}}{g_{\nu}}. \] Theorem 2.6 is applied to solve the extremal problem for complex ellipsoids and Cartan-Hartogs domains.
      0 references
      extremal map
      0 references
      Hermitian ellipsoid
      0 references
      balanced domain
      0 references

      Identifiers