On a Neumann problem involving two critical Sobolev exponents: Remarks on geometrical and topological aspects (Q1357762)

Consider the boundary value problem \[ -\Delta u = u^{2^*-1} \quad\text{in }\Omega, \qquad\partial u/\partial\nu + \delta u^{q^*-1} + g(x,u) = 0\quad\text{on }\partial\Omega,\tag{P} \] where \(\Omega\) is the unit ball in \({\mathbb{R}}^n\) (\(n\geq 3\)), \(\delta\in{\mathbb{R}}\), \(2^{*}=2n/(n-2)\) and \(q^{*}=(2n-2)/(n-2)\) are the critical exponents for the embeddings \(H^1(\Omega)\hookrightarrow L^p(\Omega)\) and \(H^1(\Omega)\hookrightarrow L^q(\partial\Omega)\), and \(g\) has subcritical growth at infinity. The main result of the paper asserts that under some additional conditions on \(g\) problem (P) has a positive solution for each \(\delta\geq 0\). The proof uses a variational argument. It is shown that the functional associated with (P) has mountain pass geometry; however, since the Palais-Smale condition fails, this is not enough to conclude. Instead the authors show that either there exists a critical point at or below the mountain pass level, or it is possible to construct a family of sets with the property that the Palais-Smale condition is satisfied on the corresponding minimax level. An important role in the proof is played by a representation lemma for Palais-Smale sequences and a comparison with rescaling-invariant problems in \({\mathbb{R}}^n\) and \({\mathbb{R}}^n_+\).