A posteriori error estimation for DEIM reduced nonlinear dynamical systems (Q2874985)

The authors introduce a novel approach for a posteriori error estimation of nonlinear dynamical systems reduced by a subspace projection and the discrete empirical interpolation method (DEIM) approximation of the system's nonlinearities. The reduction process is based on both applying the Galerkin projection of the full nonlinear system into a suitable linear subspace and applying the DEIM method to approximate the system's nonlinearity. The computations for a posteriori error estimators are efficiently decomposed in an offline/online fashion. The effectiveness of the proposed approach for a posteriori error estimation is discussed via two numerical examples: a one-dimensional viscous Burger equation and a two-dimensional reaction-diffusion model for cell apoptosis.NEWLINENEWLINEThe reviewer believes that the proposed algorithm is effective and applicable to a wide range class of nonlinear dynamical systems.