scientific article; zbMATH DE number 1462098

Other partial differential equations of complex analysis in several variables (32W50) Research exposition (monographs, survey articles) pertaining to partial differential equations (35-02) Second-order elliptic equations (35J15) Riemann-Hilbert problems in context of PDEs (35Q15) Electromagnetic theory (general) (78A25)

Review text

The book consists of 7 chapters united with the idea of a systematic usage of certain methods and results of theory of holomorphic functions of one complex variable for partial differential equations in the plane which are irregular in some sense. Chapter 1 deals with the equation \[ \frac{\partial^nu}{\partial t^n}+\sum_{k=0}^n\frac{\partial^ku}{\partial t^k}P_k\left(i\frac{\partial}{\partial x}\right)u=0\tag{1} \] for \(t>0\), \(x\in\mathbb{R}\), where \(\{P_k(\sigma)\}\) are polynomials of a real variable \(\sigma\) with constant coefficients. Let \(\rho(\sigma)\) be the number of the roots of the characteristic equation \[ \lambda^n+\sum_{k=0}^{n-1} P_k(\sigma)\lambda^k=0\tag{2} \] with \(\text{Re }\lambda\leq 0\) for a given \(\sigma\in\mathbb{R}\). Equation (1) is said to be regular if the function \(\rho\) is constant on \(\mathbb{R}\) except, possibly, a finite number of points. In Chapter 1, it is assumed that for some integers \(r,q\) there holds: \(\rho(\sigma)=r\) for \(\sigma>0\), \(\rho(\sigma)=q\) for \(\sigma<0\). The equation (1) is then considered coupled with the conditions: \[ \frac{\partial^ku}{\partial t^k}\left(x,0\right)=f_k(x)\qquad\text{ on }\mathbb{R}\text{, }k\in\{0,1,\dots,r-1\},\tag{3} \] \[ \text{Re}[b_k(x)\frac{\partial^ku}{\partial t^k}(x,0)]=f_k(x)\text{ \qquad on }\mathbb{R} \text{, }k\in\{0,1,\dots ,r-1\},\tag{4} \] where for any \(k\), \(b_{k}(x)\neq 0\) on \(\mathbb{R}\). Solutions of the problem (1), (3), (4) are looked for in a subclass of infinitely differentiable in \(\mathbb{R}_{+}^{2}\) functions, denoted by \(M_{0}(\mathbb{R})_{+}^{2})\), whose partial derivatives have at most a polynomial growth at infinity. Certain restrictions are imposed on \(b_{k},f_{k}\) as well. The main result says that in this case the problem is uniquely solvable. The proof is based on the properties of the Fourier transform and the Sokhotski-Plemelj formulas for the Cauchy-type integral. In Chapter 2, the equation \[ \frac{\partial ^{n}u}{\partial \overline{z}^{n}}+a_{1}\frac{\partial ^{n-1}u }{\partial \overline{z}^{n-1}}+\dots +a_{n} u=0\qquad\text{on }\mathbb{R}^2_+, \tag{5} \] where \(z=x+iy,\) \(\frac{\partial }{\partial \overline{z} }=\frac{1}{2}\left( \frac{\partial }{\partial x}+i\frac{\partial }{\partial y }\right) , \left\{ a_{1},\dots ,a_{n}\right\} \subset \mathbb{R}\), a particular case of (1), is considered. Its solutions are a kind of a generalization of the notion of the polyanalytic (or \(n\)-analytic) function. (5) is taken together with (4) where now \(k\in \left\{ 0,1,\dots ,n-1\right\} .\) The main result is given by Theorem 2.1 which states that if the Cauchy indices of all \(b_{k}\) are equal to zero and if all the roots of the equation \(\lambda ^{n}+a_{1}\lambda ^{n-1}+\dots +a_{n}=0\) are real, then the problem is uniquely solvable in \(M_{0}\left( \mathbb{R}_{+}^{2}\right).\) In order to prove this, first, the general solution is presented which makes use of the well-known general solution of the equation \(\frac{\partial u}{\partial \overline{z}} -\lambda u=0.\) Then the Cauchy problem is considered and it is proved that the problem (5), (4) is equivalent to the above mentioned Cauchy problem. Some cases not covered by the main result are also considered. Finally, in (5) complex coefficients are allowed which requires a different boundary value condition. Holomorphic function theory is, again, strongly exploited, in particular, the Riemann-Hilbert problem for the upper half-plane. Another particular case of (1) which is more general than (5) is studied in Chapter 3. Set \(L_{j}u:=\mu _{j}\frac{\partial u}{\partial x}+i\frac{\partial u}{ \partial t}+a_{j}u\) for \(j\in \left\{ 1,\dots ,n\right\} \) with \( \left\{ a_{j}\right\} \subset \mathbb{R}\) and \(\text{Re}\mu _{j}\neq 0,\) and consider \[ L_{1}L_{2}\dots L_{n}u=0\qquad\text{on }\mathbb{R}^2_+. \tag{6} \] fIt is now assumed that \(\text{Re}\mu _{j}>0\) for \(j\in \left\{ 1,\dots ,r\right\} ,\) \(\text{Re}\mu _{j}<0\) for \(j\in \left\{r+1,\dots ,n\right\} \) with \(r\geq q=n-r.\) (6) is taken together with (3) where \(k\in \left\{ 0,1,\dots ,q-1\right\} \) and (4) where \(k\in \left\{ q,\dots ,r-1\right\} .\) Theorem 3.1 states that the problem (6), (3), (4) in the class \(M_{0}\left( \mathbb{R}_{+}^{2}\right) ,\) is uniquely solvable. The methods of proving are analogous to those from Chapters 1-2. Then it is allowed for (6) to have complex coefficients \(a_{j}\). Theorem 3.5. The Dirichlet problem for (6) with \(n=2m\) is uniquely solvable. As an application, the Dirichlet problem for generalized analytic functions (in the sense of L. Bers - I. Vekua) is considered. Generalization to the situation when some \(\text{Re}\mu _{j}=0\) are given. In Chapter 4 there are considered electromagnetic fields depending on 1 space variable. Although formally there are no references here to the previous chapters, it follows their spirit. Subsequently, the following questions are discussed: a boundary value condition for which the solution in the hyper half-plane is unique, a more general condition which includes both components of the field and their partial derivatives with respect to \( x;\) the case of 2 homogeneous layers, Fredholmean property of the general boundary value problem for 2 layers; medium with many layers; harmonic oscillations of electromagnetic waves in non-homogeneous medium. Chapter 5 has the title ``Calculation of capacitances of cylindrical and spherical capacitors. In the author's opinion the main differences between the Chapter 5 and numerous papers on the same topic (there are references to seven of them) can be described as follows: 1. There are obtained exact formulae of capacitances for a wide class of cylindrical capacitors with cross-sections bounded by analytic curves. 2. For such capacitors with arbitrary cross-sections approximate formulae are given with estimates for the absolute and relative errors. 3. Some problems of the choice of cylindrical capacitors with given parameters are solved. 4. There are described some classes of mappings of cross-sections preserving the invariance of the capacitance which allow to choose the capacitors with prescribed parameters. 5. A new method of determining of capacitances for cylindrical capacitors is given. There are no comments concerning any relation between this chapter, on one hand, and both the preceding and the following chapters, on the other and I found not too much reasons for inserting all of them into the same book. Frankly speaking, almost the same should have been said about Chapter 4 where somewhat more relations are felt but no author's explanation are given with respect to that. In Chapter 6 we find another particular case of (1), that of the form \[ Lu:=\sum_{k=0}^nA_{k}\frac{\partial ^{n}u}{\partial x^{k}\partial y^{n-k}}=0 \tag{7} \] where \(\left( x,y\right) \in D,\) a simply connected, bounded domain in \(\mathbb{ R}^{2}\), \(\left\{ A_{k}\right\} \subset \mathbb{C}.\) The characteristic equation now is of the form \[ \sum_{k=0}^n A_{k}\lambda ^{n-k}=0 .\tag{8} \] If (8) has no real roots and \(A_{k}\neq 0\) then (7) is said to be elliptic. The proper ellipticity and strong ellipticity for (1) are defined. For an elliptic equation (7) the Dirichlet boundary condition is \[ \frac{\partial ^{k}u}{\partial N^{k}}\left( x,y\right) =f_{k}\left( x,y\right) \text{\;\;on \;\;}\Gamma =\partial \mathcal{D} \tag{9} \] for \(k\in \left\{0,1,\dots,\frac{n}{2}-1\right\},\) with \(N\) being the outer normal direction at \(\left( x,y\right) \in \Gamma.\) It's appeared that for a properly elliptic case, the Dirichlet problem is Fredholmean, and for a strongly elliptic one it is uniquely solvable. Taking \(\lambda =i\) as the root of (8) with the multiplicity \(n,\) one can rewrite (7) as \[ \frac{\partial ^{n}u}{\partial \overline{z}^{n}}=0 \tag{10} \] in \(\mathcal{D}\) which defines polyanalytic functions. The Riemann-Hilbert boundary condition for (10) is taken in the form \[ \text{Re}\frac{\partial ^{k}u}{\partial N^{k}}=f_{k}\qquad\text{ on }\Gamma, \quad k\in\{0,1,\dots,n-1\}.\tag{11} \] The general solution of the problem (10)-(11), i.e., the Riemann-Hilbert problem for polyanalytic functions is obtained, some additional conditions are indicated under which it has a unique solution. It is regretful that no bibliographical reference on polyanalytic functions including boundary value problems for them. The same questions are considered for the equation \[ \frac{\partial ^{k}u}{\partial \overline{z}^{p}\cdot \partial z^{n-p}}=0 \] which is an obvious generalization of (10) and, in fact, is equivalent to (10). A special treatment is given to the cases \(n=2\) and \(n=3.\) A small Chapter 7 is a supplement to Chapter 6 where certain details from the proof there, are justified. The bibliography contains 30 titles of which 18 sources cited, are in Russian, 1 in Japanese and 1 in Polish. For sure, some of them do not have English translations, but the majority do have it, for example, ``Izvestia Vuz'ov, Matematika is ``Soviet Mathematics, Izvestia Vuz'ov; many books cited have also English translations. The bibliography is not organized in alphabetic order; the authors' names are given sometimes in full form, and sometimes as initials only. Of those 30 titles only 4 are from 1990s, 3 from 1980s, and 7 from 1970s, the rest being older than 30 years. Certain terms do not correspond to what is canonically accepted in the English language literature, for instance, I guess that when it is written ``correct or incorrect problems it is meant to say ``well- and ill-posed problems. All this does not help to enjoy reading the book.

Cited in

(16)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4486017)