Semilinear mixed problems on Hilbert complexes and their numerical approximation (Q454131)

This paper is the second of two papers extending the Arnold-Falk-Winther abstract Hilbert complex framework. In particular, the paper extends this framework to the study of semi-linear mixed variational problems. The authors also analyze variational crimes in this semi-linear setting extending their earlier analysis in the first paper. Despite the addition of nonlinear terms, this result agrees with the quasi-optimal estimate of Arnold-Falk-Winther for the linear case. Moreover, the extension of the variational crimes analysis produce convergence results agreeing with the linear case. These last results can also be used to extend the a priori estimates for Galerkin solutions to the Laplace-Beltrami equation on approximate 2- and 3-hypersurfaces, due to \textit{G. Dziuk} [Lect. Notes Math. 1357, 142--155 (1988; Zbl 0663.65114)] and \textit{A. Demlow} [SIAM J. Numer. Anal. 47, No. 2, 805--827 (2009; Zbl 1195.65168)], to the larger class of semi-linear problems involving the Hodge Laplacian on hypersurfaces of arbitrary dimension. A tool that allows the extension from linear to semi-linear case is an alternative operator-theoretic formalism for the linear problem. It is equivalent to the mixed variational formulation, but allows for a more natural extension to semi-linear problems, due to the fact that the semi-linear problems can be expressed as an abstract Hammerstein equation.