Remarks about the geometry of the Ambrosetti-Prodi problem (Q5960863): Difference between revisions

Consider the Dirichlet problem \[ -\Delta u=f(u)+g,\qquad u\big|_{\partial\Omega}=0 \] in a smooth and bounded domain \(\Omega\subset{\mathbb R}^n,\) where \(g\in C^\alpha(\overline{\Omega},{\mathbb R});\) \(f\in C^k({\mathbb R},{\mathbb R}),\) \(k\geq 2,\) satisfies \(f''>0\) and the first eigenvalue \(\lambda_1\) of \(-\Delta\) belongs to \(f'({\mathbb R}).\) Define further the mapping \[ F:\;C^{2,\alpha}_D=\{u\in C^{2,\alpha}(\overline{\Omega}),\;u\big|_{\partial\Omega}=0\} \rightarrow C^{0,\alpha}(\overline{\Omega}) \] by \(F(u)=-\Delta u-f(u)\) and denote by \(\Sigma\) the set of singular points of \(F.\) It is proved by \textit{B. Ruf} [Topological Nonlinear Analysis II, Frascati, 1995, Progress in Nonlinear Differential Equations Applications 27, 315-395 (1997; Zbl 0892.58057)] that these assumptions give rise to a closed manifold \(\Sigma_1\subset\Sigma,\) \(\text{ codim} \Sigma_1=1,\) globally parametrized over \(H,\) which is defined as the orthogonal space (in \(L^2(\Omega)\)) to the eigenspace corresponding to \(\lambda_1.\) Imposing an additional hypothesis on \(f\) in order to have \(F\) closed, the authors show that \(N=F(\Sigma_1)\) is also a manifold of codimension 1, globally parametrized over \(H.\) Moreover, \(N\) splits \(C^{0,\alpha}(\overline{ \Omega}):\) on one side of \(N\) there is no solution \(u\in C_D^{2,\alpha}(\overline{\Omega})\) of the equation \(F(u)=g,\) while on the other side there is at least one solution.