Special Session 82: Dissipative Systems and Applications

On the Uniqueness and Multiplicity of positive solutions to an elliptic spectral problem with concave and convex nonlinearity
Peter Takac
University of Rostock
Germany
Co-Author(s):    J. Benedikt, P. Girg, and L. Kotrla
Abstract:
We will discuss the question of {\it\bfseries existence\/} and {\it\bfseries multiplicity\/} of {\it positive solutions\/} to the semilinear elliptic Dirichlet problem % \begin{equation} \label{e:semi-lin} \begin{alignedat}{3} {}- \Delta u = \lambda\, u(x)^{q(x) - 1} + f(x,u(x)) \quad\mbox{ for }\, x\in \Omega \,;\qquad u = 0 \quad\mbox{ on }\, \partial\Omega \,, \end{alignedat} \end{equation} % where $\Omega\subset \RR^N$ is a bounded domain with the boundary of class $C^{1,\alpha}$, $\lambda\in \RR$ a spectral parameter, and $f(x,u) = |u|^{r-1}\, u$ is a \underline{\bf signed $r$\--power} ($r > 0$) of the unknown function of (a positive variable) $u\in (0,\infty)$ which depends on the point $x\in \Omega$; $r = q(x) - 1$, for instance. We will briefly present basic methods for treating the semilinear problem \eqref{e:semi-lin} with a {\it\bfseries convex\/} and {\it\bfseries concave\/} non\-linear reaction % \begin{math} f(x, \,\cdot\,)\colon s\longmapsto |s|^{q(x) - 2} s\colon \RR_+\subset \RR\to \RR \end{math} % which (for $s\geq 0$) is {\it\bfseries convex\/} in a nonempty open subset % \begin{math} \Omega_{+}\eqdef \{ x\in \Omega\colon q(x) > 2\} \end{math} % and {\it\bfseries concave\/} in another nonempty open subset % \begin{math} \Omega_{-}\eqdef \{ x\in \Omega\colon q(x) < 2\} \end{math} % of a bounded domain $\Omega\subset \mathbb{R}^N$. Here, $\lambda\in \RR_+$ is a non\-negative spectral parameter which decides about the existence and multiplicity of positive weak solutions (at least two) to problem \eqref{e:semi-lin} in case we take $f\equiv 0$. Our main contribution is a method how to handle the interplay between {\it convex\/} and {\it concave\/} non\-linearities in two disjoint nonempty open subsets of a domain $\Omega$ (connected in $\RR^N$), as opposed to the classical works assuming a non\-linearity $f(s)$ being {\it concave\/} for small values of $s\in \RR_+$ and {\it convex\/} for large $s\in \RR_{+}$, uniformly in $\Omega$.