[Paper Review] On Dirichlet problem for Beltrami equations with two characteristics

Bogdan Bojarski, Vladimir Gutlyanskiî+1 more|arXiv (Cornell University)|Nov 2, 2012

Analytic and geometric function theory20 references3 citations

TL;DR

This paper establishes sufficient conditions for the existence of regular solutions to the Dirichlet problem for degenerate Beltrami equations with two characteristics in arbitrary Jordan domains. Using an approximate procedure based on bounded dilatation theory and convergence theorems, it proves that if the dilatation coefficient $ K_{\mu,\nu} $ satisfies certain integral exponential integrability conditions, then a regular solution exists for any nonconstant continuous boundary data.

ABSTRACT

We establish a series of criteria on the existence of regular solutions for the Dirichlet problem to general degenerate Beltrami equations ${\bar{\partial}}f = μ{\partial f}+ν{\bar{\partial f}}$ in arbitrary Jordan domains in $\C$.

Motivation & Objective

• To establish existence criteria for regular solutions to the Dirichlet problem for general degenerate Beltrami equations with two measurable coefficients $ \mu $ and $ \nu $ in Jordan domains.
• To extend the solvability theory beyond the uniformly elliptic case, where $ K_{\mu,\nu} \in L^\infty $, to the case where $ K_{\mu,\nu} \in L^1_{\text{loc}} $ but possibly unbounded.
• To provide sufficient conditions on the growth of $ K_{\mu,\nu} $ ensuring existence of continuous, discrete, open, and non-vanishing Jacobian solutions.
• To generalize previous results on reduced Beltrami equations and homeomorphic solutions to the case with two characteristics.
• To unify and extend existing criteria using convex, non-decreasing functions $ \Phi $ satisfying specific integral conditions on $ K_{\mu,\nu} $.

Proposed method

• Adapts an approximate solution procedure based on known existence theorems for the uniformly elliptic case ($ K_{\mu,\nu} \in L^\infty $) via conformal mapping and Schwarz formula.
• Applies convergence theorems for Beltrami equations with $ K_{\mu,\nu} \in L^1_{\text{loc}} $ to pass to the limit in approximating sequences.
• Uses the Schwarz formula to recover analytic functions from their boundary real parts, enabling construction of solutions via composition with quasiconformal mappings.
• Employs the Stoilow factorization theorem to represent solutions as $ f = \mathcal{A} \circ h $, where $ h $ is a homeomorphic solution to a Beltrami equation and $ \mathcal{A} $ is analytic.
• Imposes integral conditions on $ K_{\mu,\nu} $ involving convex, non-decreasing functions $ \Phi $, including exponential-type growth control via $ \int_D \Phi(K_{\mu,\nu}) \, dxdy < \infty $.
• Considers the case where $ \Phi $ needs only be convex and non-decreasing on $[T, \infty) $, allowing truncation and replacement by a majorizing convex function without loss of generality.

Experimental results

Research questions

• RQ1 Under what conditions on the coefficients $ \mu $ and $ \nu $ does the Dirichlet problem for the degenerate Beltrami equation $ \overline{\partial}f = \mu \partial f + \nu \overline{\partial f} $ admit a regular solution in a Jordan domain?
• RQ2 Can the existence of regular solutions be guaranteed when the dilatation $ K_{\mu,\nu} $ is not essentially bounded but satisfies weaker integrability conditions?
• RQ3 What is the role of the boundary data $ \varphi \not\equiv \text{const} $ in ensuring the existence of non-constant solutions?
• RQ4 How do exponential integrability conditions on $ K_{\mu,\nu} $, such as $ \int_{D \cap U_{z_0}} e^{\alpha(z_0) K_{\mu,\nu}(z)} \, dxdy < \infty $, relate to the solvability of the Dirichlet problem?
• RQ5 To what extent can the results be extended to reduced Beltrami equations of the form $ f_{\overline{z}} = \lambda(z) \, \text{Re}(f_z) $, where $ \lambda $ is measurable and $ |\lambda| < 1 $ a.e.?

Key findings

• The Dirichlet problem for the degenerate Beltrami equation $ f_{\overline{z}} = \mu f_z + \nu \overline{f_z} $ has a regular solution in any Jordan domain $ D $ if the dilatation $ K_{\mu,\nu} $ satisfies $ \int_D \Phi(K_{\mu,\nu}) \, dxdy < \infty $ for a non-decreasing convex function $ \Phi $ satisfying at least one of the conditions (5.21)–(5.26).
• A sufficient condition for solvability is the local exponential integrability of $ K_{\mu,\nu} $, i.e., $ \int_{D \cap U_{z_0}} e^{\alpha(z_0) K_{\mu,\nu}(z)} \, dxdy < \infty $ for some $ \alpha(z_0) > 0 $ and neighborhood $ U_{z_0} $ of each $ z_0 \in \overline{D} $.
• The result extends to reduced Beltrami equations $ f_{\overline{z}} = \lambda(z) \, \text{Re}(f_z) $, where $ \lambda $ is measurable and $ |\lambda| < 1 $ a.e., with the same exponential integrability condition on $ K_\lambda = \frac{1+|\lambda|}{1-|\lambda|} $.
• The solution is constructed as $ f = \mathcal{A} \circ h $, where $ h $ is a homeomorphic solution to a Beltrami equation and $ \mathcal{A} $ is analytic, obtained via conformal mapping and Schwarz formula.
• The conditions on $ \Phi $ are sharp in the sense that they are necessary and sufficient for the existence of homeomorphic solutions to the corresponding Beltrami equations under the same integral constraints.
• The results remain valid even if $ \Phi $ is only convex and non-decreasing on $[T, \infty) $, as long as it is majorized by a convex function defined on $[0, \infty) $, allowing for truncation and convex extension.