A Morse index formula for radial solutions of Lane-Emden problems

Abstract

We consider the semilinear Lane-Emden problem: \begin{equation}\label{problemAbstract}\left\{\begin{array}{lr}-Δu= |u|^{p-1}u\qquad \mbox{ in }B u=0\qquad\qquad\qquad\mbox{ on }\partial B \end{array}\right.\tag{${\mathcal{E}}_p$} \end{equation} where $B$ is the unit ball of ${\mathbb{R}}^N$, $N\ge 3$, centered at the origin and $1<p<p_S$, $p_S=\frac{N+2}{N-2}$. We prove that for any radial solution $u_p$ of \eqref{problemAbstract} with $m$ nodal domains its Morse index $\mathsf{m}(u_p)$ is given by the formula \[\mathsf{m}(u_p)=m+N(m-1)\] if $p$ is sufficiently close to $p_S$.