Nonlinear time-harmonic Maxwell equations in an anisotropic bounded medium (Q2400360): Difference between revisions

This paper is a continuation of the first author's work with \textit{Y. Ding} [Math. Nachr. 279, No. 12, 1267--1288 (2006; Zbl 1117.58007)] studying ground state and other solutions to \[ \nabla\times(\mu(x)^{-1}\nabla\times E)-\omega^2\epsilon(x)E=\partial_E F(x,E)\quad in\;\Omega \] with perfect metallic boundary condition: \(\nu\times E=0\). The frame of study extends the one in [loc. cit.] in several ways: 1) \(\Omega\) is Lipchitz while it is smooth and simply connected in [loc. cit.], 2) \(\epsilon,\mu\) are constant in [loc. cit.]. Here they are symmetric matrices. In the axisymmetric case they are diagonal and the present results recover those of \textit{T. D'Aprile} and \textit{G. Siciliano} [Adv. Differ. Equ. 16, No. 5--6, 435--466 (2011; Zbl 1260.35224)]. 3) the class of admissible nonlinear terms is extended and made easier to characterize thanks to a simplification of a condition in [the first author and Ding, loc. cit.] bearing on the product \(F\partial_E F\). Here instead, convexity is required. This allows for Kerr nonlinearity of the form \[ F(x,u)=1_{[\delta,\infty[}(|u|)\sum_{i=1}^m\frac{1}{p_i}|\Gamma_i(x)u|^{p_i},\quad 2 < p_i < 6 \] for some \(\delta\geq 0\). The main tool for showing existence of solutions is by adapting and refining the mountain-pass method used in [the first author and Ding, loc. cit.].