Numerical approximation of the first eigenpair of the<i>p</i>-laplacian using finite elements and the penalty method (Q4351397)

For \(1<p <\infty\) and for bounded convex domains \(\Omega\), the authors are interested in the first eigenvalue \(\lambda_1\) of \[ \text{div} \bigl(|\nabla |^{p-2} \nabla u\bigr) =\lambda |u|^{p-2} u\quad \text{in } \Omega, \quad u=0 \quad \text{on } \partial \Omega, \] where \(|\cdot |\) denotes the norm in \(L^2(\Omega)\). The eigenvalue \(\lambda_1\) is simple, and the eigenfunction \(u_1\) of one sign. \(\lambda_1\) is the minimum of \(F(u)= \int_\Omega |\nabla u|^p dx\) over \(W_0^{1,p} (\Omega)\) under the constraint \(G(u) =1- |u|^p =0\). To evaluate \(\lambda_1\) and \(u_1\), the sequence of penalty functionals \(L(u,\gamma_k) =F(u) +\gamma_k [G(u)]^2\) is introduced, and \(u_1\) is approximated by finite elements. Convergence is proved, and numerical results are presented for various values of \(p\) and for \(\Omega= (0,1)\), \(\Omega= (0,1) \times(0,1)\), and \(\Omega =\) unit ball in \(\mathbb{R}^2\).