Some remarks on the Dirichlet problem for semi-linear elliptic equations with the Ambrosetti-Prodi conditions (Q912283): Difference between revisions

A variant of the Ambrosetti-Prodi problem for the so-called jumping nonlinearities is studied in this article. More precisely, the author considers the semilinear elliptic problem \(Lu=f(u)+s\theta +h\) in D, \(u=t\phi\) on \(\partial D\), where D is a smooth bounded domain, L is a uniformly elliptic second order differential operator, s and t are real parameters, \(\theta\) is the eigenfunction corresponding to the first eigenvalue \(\lambda_ 1\) of L with zero boundary conditions, h is orthogonal to \(\theta\), and f satisfies a jumping nonlinearity condition. The main result in the paper is that if \(\phi \in L^{\infty}(\partial D)\cap H^{1/2}(\partial D)\) and \(\phi\geq 0\), \(\phi\) \(\not\equiv 0\), there exists a number \(s_ 0\) such that for any \(s\leq s_ 0\) there is \(t^*=t^*(s)\) such that the above problem has at least one solution in \(H^ 1(D)\) for \(t\leq t^*\) and no solution for \(t>t^*\). A weaker version is obtained if \(\phi \in L^{\infty}(\partial D)\) only. Some variants are included, for example if \(\phi\) changes sign. The fact that \(\phi\) is not smooth raises several problems: a notion of weak solution is needed, and this involves weighted Sobolev spaces and sub and supersolutions in the sense of \(H^{1/2}(\partial D)\).