The abstract linear boundary value problem (a nongroup approach) (Q910683)

\(\{\) This is a posthumous paper by the well-known functional analyst, long an associate of M. G. Kreĭn.\(\}\) Let \(D_ 0\) denote a linear operator, \(E_ 0=\ker D_ 0\), and fix a subspace \(M_ 0\) of dom \(D_ 0\) with \(E_ 0\cap M_ 0=\{0\}\). Assume both domain and range of \(D_ 0\) are in \(C_ Q\), the space of real continuous functions on a compact set Q. An example is an elliptic differential operator in a compact domain \(Q\subseteq {\mathbb{R}}^ n\), taking \(M_ 0\) to be the space of all functions in \(C^ 2(int Q)\) vanishing on \(\partial Q\). The paper is concerned with questions of boundedness and invertibility of \(D_ 0\) on \(D_ 0M_ 0\), and abstract boundary problems (of a type to be described below) involving such operators. Theorem 4: Assume there exists a strictly positive function in \(D_ 0M_ 0\) and assume that every \(v\in M_ 0\) such that \(D_ 0v\) is nonnegative on Q is itself nonnegative on Q. Then \(D_ 0\) has a bounded inverse \((D_{M_ 0})^{-1}\) on \(D_ 0M_ 0\) and we have an integral representation \((D_{M_ 0})^{-1}x(q)=\int_{Q}x(s)d\sigma (s,q).\) Here q runs over \(Q\setminus \Gamma^ 0\), where \(\Gamma^ 0\) is the set of those \(q\in Q\) such that \((D_{M_ 0})^{-1}p(q)=0\) for every \(p\in D_ 0M_ 0\) strictly positive on Q, \(Q\neq \Gamma^ 0\) contains the functional boundary, and for each q, \(\sigma\) (\(\cdot,q)\) is a finite measure \(\{\) concentrated on the functional boundary?\(\}\). Most of the paper is devoted to abstract boundary problems, by which is meant an equation \(D_ 1^{n_ 1}D_ 2^{n_ 2}...D_ n^{n_ N}y=By+f\) with each \(D_ i\) of the type described above, and where these \(D_ i\) are not assumed to commute with each other or with B. In Theorem 3, the key hypothesis is the existence of an operator with spectral radius less than 1 which majorizes \(A=B(D^{-1}_{M_ n})^{n_ N}...(D^{-1}_{M_ 1})^{n_ 1}\); and the conclusion is the existence of a unique solution y which can be expressed in terms of \((1- A)^{-1}.\) \{If the author had lived to publish this paper, he probably would have supplied more references. The spirit of this theorem seems to be allied to that of \textit{D. Przeworska-Rolewicz} [Equations with transformed argument, Chapter IV, PWN, Warsaw (1973; Zbl 0271.47008)].\}