Solutions of elliptic equations with indefinite nonlinearities via Morse theory and linking (Q1912017): Difference between revisions

The paper is concerned with the boundary value problem \[ - \Delta u= \lambda u+ h(x) f(u)\quad \text{in} \quad \Omega,\quad u= 0\quad \text{on} \quad \partial\Omega,\tag{\(*\)} \] where \(\Omega\) is a bounded domain in \(\mathbb R^N\), \(h\) changes sign in \(\Omega\), \(f(u)= o(u)\) as \(u\to 0\) and \(f\) is superlinear (but subcritical) as \(|u|\to \infty\). It is shown that, depending on the choice of \(\lambda\) (and some additional hypotheses), \((*)\) has at least one, three and five nontrivial solutions. The proofs use a combination of variational and Morse-theoretical arguments. In particular, in the last case two solutions are obtained as local minima of the associated functional (by the method of sub- and supersolutions), another two by the mountain pass theorem and the last one by computing the critical groups and employing the Morse inequalities. It is also shown, via a linking argument, that if \(\lambda\in (\lambda_1, \lambda_2)\) (\(\lambda_i\) are the eigenvalues of \(-\Delta\)), then \((*)\) has at least one nontrivial solution under a somewhat different (and weaker) set of hypotheses.