卷 221, 编号 5 (2017)

In the geometric theory of functions of complex variables, we consider V.N. Dubinin’s extremal problem associated with estimates of a functional defined on a system of nonoverlapping domains and obtain a particular solution of this problem.

The survey is devoted to recent advances in nonclassical solutions of the main boundary-value problems such as the well-known Dirichlet, Hilbert, Neumann, Poincaré, and Riemann problems in the plane. Such solutions are essentially different from the variational solutions of the classical mathematical physics and based on the nonstandard point of view of the geometric function theory with a clear visual sense. The traditional approach of the latter is the meaning of the boundary values of functions in the sense of the so-called angular limits or limits along certain classes of curves terminated at the boundary. This become necessary if we start to consider boundary data that are only measurable, and it is turned out to be useful under the study of problems in the field of mathematical physics as well. Thus, we essentially widen the notion of solutions and, furthermore, obtain spaces of solutions of the infinite dimension for all the given boundary-value problems. The latter concerns the Laplace equation, as well as its counterparts in the potential theory for inhomogeneous and anisotropic media.

The main object of the paper is a symmetric system J_y^′ − B(t)y = ⋋∆(t)y defined on an interval Ι = [a, b) with the regular endpoint a. Let φ(⋅, λ) be a matrix solution φ(⋅, λ) of this system of an arbitrary dimension, and let \( \left( V\kern0.5em f\right)(s)={\displaystyle \underset{I}{\int }{\varphi}^{\ast}\left( t, s\right)\varDelta (t) f(t) d t} \) be the Fourier transform of the function f(⋅) ∈ L_Δ²(I). We define a pseudospectral function of the system as a matrix-valued distribution function σ(·) of the dimension n_σ such that V is a partial isometry from \( {L}_{\varDelta}^2(I)\kern0.5em \mathrm{t}\mathrm{o}\kern0.5em {L}^2\left(\sigma; \kern0.5em {\mathbb{C}}^{n_{\sigma}}\right) \) with minimally possible kernel. Moreover, we find the minimally possible value of n_σ and parametrize all spectral and pseudospectral functions of every possible dimensions n_σ by means of a Nevanlinna boundary parameter. The obtained results develop the results by Arov and Dym; Sakhnovich, Sakhnovich and Roitberg; Langer and Textorius.