Error analysis of an implicit-explicit time discretization scheme for semilinear wave equations with application to multiscale problems
Daniel Eckhardt, Marlis Hochbruck, Barbara Verfürth
TL;DR
This work develops and analyzes an implicit–explicit (IMEX) time discretization for semilinear damped wave equations, achieving unconditional stability by implicitly treating the stiff linear part and explicitly handling the nonlinear nonstiff part. A rigorous error framework is established for full discretization, combining the IMEX time stepping with abstract space discretizations, and a second-order temporal convergence result is proved under suitable regularity. The theory is then extended to a multiscale setting by applying it to the FE-Heterogeneous Multiscale Method (FE-HMM) for wave equations with highly oscillatory coefficients, deriving an $\varepsilon$-independent homogenized model and quantifying the spatial and temporal discretization errors. The paper also provides a detailed multiscale implementation and numerical evidence confirming the predicted convergence rates, illustrating the practical relevance of IMEX schemes for semilinear wave equations and HMM-based multiscale simulations. Overall, the contribution offers a robust, efficient framework for accurate time integration of semilinear damped waves in multiscale environments and delivers concrete error guarantees for fully discrete FE-HMM discretizations.
Abstract
We present an implicit-explicit (IMEX) scheme for semilinear wave equations with strong damping. By treating the nonlinear, nonstiff term explicitly and the linear, stiff part implicitly, we obtain a method which is not only unconditionally stable but also highly efficient. Our main results are error bounds of the full discretization in space and time for the IMEX scheme combined with a general abstract space discretization. As an application, we consider the heterogeneous multiscale method for wave equations with highly oscillating coefficients in space for which we show spatial and temporal convergence rates by using the abstract result.
