Theory of variational methods for the Beltrami equations (Q2880010)

In the present paper, some problems connected with the Beltrami equation \(f_{\overline z}=\mu(z)\cdot f_z\) are studied. Let \(M(z)\) be a family of measurable compact sets in the unit disk \({\mathbb D}\). Denote by \({\mathfrak M}_M\) the class of all measurable functions satisfying the condition \(\mu(z)\in M(z)\) a.e., and by \(H_M^*\) the class of all regular sense-preserving homeomorphisms \(f:\overline{\mathbb{C}}\rightarrow \overline{\mathbb{C}}\) having the complex characteristics \(\mu(z)\) from the class \({\mathfrak M}_M\) normalized by \(f(0)=0\), \(f(1)=1\) and \(f(\infty)=\infty\). One of the main results of the work is the following. Let \(M(z)\subset {\mathbb D}\), \(z\in {\mathbb C}\), be a family of compact convex sets such that \(Q_M:=\frac{1+q_M(z)}{1-q_M(z)}\in L_{\text{loc}}^1\) and \(q_M(z):=\max\limits_{\nu\in M(z)}| \nu|\). Suppose that a functional \(\Omega:H_M^*\rightarrow {\mathbb R}\) is Gato differentiable without degeneration. If the maximum of \(\Omega\) is achieved by \(f\in H_M^*\), then its complex characteristics \(\mu(z)\) satisfies the condition \(\mu(z)\in \partial M(z)\) for almost every \(z\in {\mathbb C}\). Moreover, under the above conditions, the extremal \(f\) satisfies the requirement \(\mathrm{Re}\;\omega{\mathcal B}(z)\geq 0\) at a.e. \(z\), where \(\omega\neq 0\) is such that \(\mu(z)+\varepsilon\omega\in M(z)\) at a.e. \(z\) and \(\varepsilon\in [0, \varepsilon_0]\) at some \(\varepsilon_0>0\), \({\mathcal B}(z)={\mathcal A}(f(z))f_z^2\), \(A(\omega)=\frac{1}{\pi}\int\limits_{\mathbb C}\varphi(\omega, f(\zeta))d\chi(z)\) and \(\varphi(\omega, \omega^{\,\prime})=\frac{1}{\omega-\omega^{\,\prime}}\cdot\frac{\omega^{\,\prime}}{\omega} \cdot\frac{\omega^{\,\prime}-1}{\omega-1}\).