On E. E. Levi's method of constructing a fundamental solution for second- order elliptic equations (Q1202752)

The author is concerned with the existence of fundamental solutions, in bounded domains of \(\mathbb{R}^ n\), for second-order uniformly elliptic linear differential operators of the general form \(L = \sum^ n_{j,k=1} a_{j,k} (x)D_ j D_ k + \sum^ n_{k=1} b_ k (x)D_ k + c(x)\). The basic assumptions are the following: i) \(a_{j,k} \in C^ \alpha (\overline \Omega) \cap C^ 1 (\Omega)\) \((j,k = 1, \dots,n; \alpha \in (0,1])\); ii) \(b_ k\), \(c \in C^ \alpha (\overline \Omega)\) \((k = 1, \dots, n)\); iii) \(\sum^ n_{k=1} D_ k a_{j,k} - b_ j \in C^ \alpha (\overline \Omega) \cap C^ 1 (\Omega)\) \((j = 1, \dots, n)\). Conditions i), ii), iii) insure the existence of a fundamental solution for ``small'' domains \(\Omega\), if \(\alpha \in (0,1)\) and for any subdomain \(\Omega_ 1 \) such that \(\Omega_ 1 \subset \Omega\), if \(\alpha = 1\). For such solutions the author examines the continuity and Hölder continuity (in a suitable sense, of course), giving estimates usually not easily available. The paper is a very interesting survey on the subject of fundamental solutions, collecting an up-to-date and rich bibliography and providing explicit (and sometimes original) proofs of many technical results, which are of use to everyone working with elliptic partial differential equations.