Existence of continuous eigenvalues for a class of parametric problems involving the \((p,2)\)-Laplacian operator (Q2304856): Difference between revisions

The authors investigate a class of parametric eigenvalue problems: \[ \begin{cases} -t\Delta_p u-(1-t)\Delta u = \lambda u, & \quad\mbox{in} \; D, \cr u = 0, & \quad\mbox{on} \; \partial D, \end{cases} \eqno{(6)} \] where \(t \in [0, 1]\), \(\lambda\in\mathbb{R}\), \(D \subseteq \mathbb{R}^n\) is a smooth bounded domain and \(\Delta_p\) denotes the \(p\)-Laplacian operator for \(p\in (1,\infty ).\) They denote \({\mathfrak C}_t= -t\Delta_p-(1-t)\Delta\) and prove the following three main theorem. Theorem 1. Let \(p\in (1,\infty )\setminus \{2\}, t \in (0, 1)\), and assume that \(\lambda_1\) denotes the first eigenvalue of \(-\Delta\) with respect to the Dirichlet boundary condition on \(\partial D\). Then \(\sigma({\mathfrak C}_t ) =((1 -t)\lambda_1,\infty ).\) The authors prove the theorem using variational methods and consider an energy functional associated with (6). They prove that the critical points of this functional will give rise to non-trivial solutions of (6). For \(p > 2\), the energy functional is coercive, hence the direct method applies, for \(p < 2\) the authors apply the fibering method. They also derive a priori bounds and regularity results on the eigenfunctions. Theorem 2. Assume that \(p \in (2,\infty )\), and let \(u \in W^{1,p}_0 (D)\) be a non-trivial solution of (6). Then the following hold: (i) \(u \in C^{2,\alpha}_{\mathrm{loc}}\cap L^\infty(D)\), and \(u \in C^\infty (D\setminus \{\nabla u = 0\}).\) (ii) \(\|u\|_{W^{1,p}_0(D)}\le \left( \frac{\lambda -(1-t)\lambda_1}{\lambda_1t}\right)^{1/(p-2)}|D|^{1/p}.\) (iii) \(\sup_D |u| \le \frac{C}{\lambda_1^{1/2}} \left( \frac{\lambda -(1-t)\lambda_1}{\lambda_1t}\right)^{1/(p-2)}|D|^{1/2}\). In particular, \(\sup_D |u|\to 0\) as \(\lambda \downarrow (1- t)\lambda_1.\) Theorem 3. Assume that \(p \in (1, 2)\) and let \(u \in W^{1,2}_0 (D)\) solve (6). Then the following hold: (i) \(u \in C^{1,\alpha}_{\mathrm{loc}} L^\infty(D)\), and \(u \in C^\infty (D\setminus \{\nabla u = 0\})\). (ii) For any non-trivial solution \(u\), there is a constant \(C = C(\lambda ,p,n,D)\) such that: \[ \|u\|_{W^{1,p}_0 (D)} \ge \left(\int_D |u|^p \,dx \right)^{1/p} \ge C \left( \frac{t\lambda_p}{\lambda -\lambda_1(1-t)}\right)^{1/(2-p)}. \] Moreover, \(C >0\) if \(\lambda >0\). (iii) \(\sup_D |u|\to \infty\) as \(\lambda \downarrow (1- t)\lambda_1\).