Real analyticity and non-degeneracy (Q1566339): Difference between revisions

The following theorem is proved: Let \(U\) be a real Banach space and \(F: U\times \mathbb{R}\to U\) an analytic Fredholm map of index one with \(F(u_0,0)= 0\), let \(\partial_uF(u_0,0)\) be invertible, and let the set \(\{(u,t)\in U\times [0,1]: F(u,t)= 0\}\) be compact. Then there exists a subset \(T\) of \([0,1)\) consisting of isolated points such that for all \(t\in [0,1)\setminus T\) the equation \(F(u,t)= 0\) has an odd number of non-degenerated solutions. Note that this theorem is false for \(C^\infty\) maps. Applications are given to positive solutions of elliptic PDEs of the type \(-\Delta u= u^p\) on bounded domains \(\Omega\subset \mathbb{R}^N\), on strips \(S\subset \mathbb{R}^N\) and on the whole space \(\mathbb{R}^N\). For that the following surprising result is shown: The superposition operator \(u\mapsto u^p\) is analytic on certain open sets in certain subspaces of the space of continuous functions \(u:\Omega\to \mathbb{R}\) with \(u= 0\) on \(\partial\Omega\), even if \(p\) is non-integer.