Reduced measures for obstacle problems (Q2504018): Difference between revisions

In this interesting article the authors study the problem: \[ \begin{cases} -\Delta u+\beta(u)\ni\mu\quad &\text{in }\Omega,\\ u= 0\quad &\text{on }\partial\Omega\end{cases}\tag{1} \] where \(\Omega\) is a smooth, bounded domain in \(\mathbb R^N\), \(N\geq 1\), \(\mu\) a finite measure in \(\Omega\) and \(\beta\) a maximal monotone graph such that \((0,0)\in\text{graph}\,\beta\). Here \(\text{dom}\,\beta\neq\mathbb R\) and \(\text{dom}\,\beta\) is closed, the case \(\text{dom}\,\beta=\mathbb R\) being studied in a paper by \textit{H. Brezis, M. Marcus} and \textit{A. C. Ponce} [Ann. Math. Stud. 163, 55--109 (2007; Zbl 1151.35034), see also C. R., Math., Acad. Sci. Paris 339, No. 3, 169--174 (2004; Zbl 1101.35029)] and the case \(\text{dom}\,\beta\) open set in a paper by \textit{L. Dupaigne, A. C. Ponce} and \textit{A. Porretta} [J. Anal. Math. 98, 349--396 (2006; Zbl 1132.35366)]. The main concern is twofold: the first results identify the ``good'' measures, i.e. the measures for which problem (1) has one (actually only one solution) (see Theorem 1 and Theorem 7). Moreover, the authors, if \(\mu\) is not a ``good'' measure, define some kind of ``generalized'' solution and prove that it is the limit of all natural approximation schemes (see Theorem 2, Theorem 3 and Theorem 9). Finally, the authors establish the connetion between the concept of ``generalized'' solution, introduced in this paper, and the concept of ``generalized'' solution based on duality principle introduced in a previous paper, by \textit{H. Brezis} and \textit{S. Serfaty} [Commun. Contemp. Math. 4, No. 2, 357--374 (2002; Zbl 1146.35372)] (see Theorem 10).