Boundary value problems with jumping nonlinearities (Q1094583): Difference between revisions

Let \(\Omega\) be a bounded domain in \(R^ n\) with sufficient smooth boundary \(\partial \Omega\) and consider the boundary value problem \[ (1)\quad \Delta u+f(u)=g(x),\quad x\in \Omega;\quad u=0,\quad x\in \partial \Omega, \] where \(f\in C^ 1(R)\) and g is Hölder continuous on \({\bar \Omega}=\Omega \cup \partial \Omega.\) Let \(\lambda_ 0<\lambda_ 1\leq \lambda_ 2\leq..\). denote the set of all eigenvalues of the homogeneous problem \(\Delta u+u=0\), \(x\in \Omega\), \(u=0\), \(u\in \partial \Omega\). If \(\alpha =_{s\to \infty} f'(s),\beta =_{s\to \infty} f'(s)\) exist and satisfy \(\alpha <\lambda_ 0<\beta <\lambda_ 1\). In the present paper, the author uses techniques from bifurcation theory to study the conditions under which there may or may not exist solutions for the problem (1). If a solution exists, then how many are there ? The author has given a detailed analysis for the case of ordinary differential equations. There are still many questions left open. It would be interesting to see whether the method of this paper can be extended with some modifications to study similar problems for partial differential equations.