Answers are grounded in this paper and cite their sources.
We develop and assess an exponential integration generalized multiscale virtual element method (EI-GMsVEM) for time-dependent parabolic problems in heterogeneous and high-contrast media. The method combines an H 1 {H¹} -conforming fine-scale virtual element discretization on polygonal meshes with a generalized multiscale reduction based on local spectral problems over coarse neighborhoods. The resulting reduced system retains coefficient-dependent fine-scale information through multiscale basis functions while substantially decreasing the number of dynamic unknowns. Time integration is performed by an exponential Euler scheme that treats the stiff diffusion operator through matrix exponential actions and evaluates the nonlinear reaction term explicitly. The paper focuses on the design, construction, implementation, and computational behavior of the method. We describe the offline and online stages, the assembly of coarse operators, and the practical workflow required to implement the EI-GMsVEM. Numerical experiments for linear and semilinear problems demonstrate accuracy, robustness with respect to high contrast, convergence under spectral enrichment, and the computational advantages of the reduced exponential formulation.
No takes yet. Share an insight, caveat, or question.
Contreras et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: