On a class of quasilinear eigenvalue problems on unbounded domains (Q5956000)

The authors consider the quasilinear elliptic problem, \[ -\text{div} (a(x) |\nabla u|^{p-2} \nabla u) +h(x) |u|^{r-2} = \lambda (1+|x|)^{\alpha_1} |u|^{q-2} u, \tag{I} \] on an unbounded domain \(\Omega \subset \mathbb{R}^N\) with the nonlinear boundary conditions \[ a(x) |\nabla u|^{p-2} \nabla u \cdot n + b(x) |u|^{p-2} u = \theta g(x,u), \quad x \in \Gamma = \partial \Omega, \tag{II} \] involving the two parameters \(\lambda\) and \(\theta\). Under conditions \(1<p<N\), \(\max\{p,2\} < q< r < pN/(N-p)\), \(\alpha \in (-N,q(N-p)/p -N)\), \(a(x) \geq 0\); \(c(1+|x|)^{p-1} \leq b(x) \leq C(1+|x|)^{p-1}\); \(\int_\Omega w_1^{r/(r-q)}/h^{q/((r-q)} dx < \infty\), where \(w_1 = (1+|x|)^{\alpha_1}\), together with conditions on the Carathéodory function \(g(x,u)\), the authors prove the following theorem: Theorem 1: There exist numbers \(\theta_* < \theta^*\) and \(\lambda^* >0\) such that the problem (I)-(II) has no nontrivial solutions for \(\lambda \in (0,\lambda^*)\) and \(\theta \in (\theta_*, \theta^*)\). In the above theorem the quantities \(\theta_*\), \(\theta^*\) are defined in terms of the function \(g(x,u)\). If \(G(x,u)\) is the primitive of \(g(x,u)\), one can define two additional values of \(\theta\), \(\theta_-\) and \(\theta^+\). Under the condition that the interval \[ J = ( \max \{ \theta_*, \theta_- \}, \min \{ \theta^*, \theta^+ \}) \] is non-empty, the authors also prove Theorem 2: (i) (I)-(II) admits a nontrivial solution for all \(\theta \in J\) and \( \lambda \geq \lambda _0\) for some \(\lambda_0 > 0\); (ii) If \(\theta \in J\), \(\lambda < \lambda_0\), (I)-(II) does not admit a nontrivial solution. The proofs proceed by considering the properties of the energy functional on a suitably defined weighted Sobolev space that is associated with (I)-(II).