Generalized Newton potential and its applications (Q2367808)

It is known that if \(E(x)\) is the fundamental solution of the Laplacian in \(\mathbb{R}^ n\), \(n\geq 3\), and if \(\Omega\subset\mathbb{R}^ n\) is an open set, the convolution product \(E*g\) is a solution of \(\Delta u=-g\), \(g\in L^ \infty(\Omega)\). As Newton's potential cannot be defined for an arbitrary unbounded domain \(\Omega\), the author modifies the fundamental solution \(E(x)\), using a method initiated by \textit{W. K. Hayman} and \textit{P. B. Kennedy}, quoted in the bibliography [Subharmonic functions. Vol. I (1976; Zbl 0419.31001)]. It consists of the subtraction of finitely many terms of its Taylor expansion. On this basis, the author proves that \(\text{span} \{K(x,y),\partial K(x,y)/\partial y\); \(y\not\in\Omega\), \(i=1,\dots,n\}\) is a dense set in \(L^ 1_ h(\Omega)\), where \(L^ 1_ h(\Omega)=L^ 1(\Omega)\cap\{u\): \(\Delta u=0\) in \(\Omega\}\) and \(K(x,y)\) is the modified fundamental solution. By means of the generalized Newton's potential, it is possible to treat unbounded quadrature domains. Finally, he proves that the equation \(\Delta u=f\), \(f\in L^ \infty(\mathbb{R}^ n)\) has a solution satisfying the inequality \[ | u(x)|\leq C\| f\|_ \infty(1+ | x|)^ 2\log(2+| x|), \] where the constant \(C\) does not depend on \(f\).