Domain decomposition for multiscale PDEs (Q2373400)

The authors consider additive Schwarz domain decomposition preconditioners for piecewise linear finite element approximations of elliptic partial differential equations with highly variable coefficients. These preconditioners combine local solvers on general overlapping subdomains together with a global solver on a general coarse space of functions on a coarse grid. An analysis of the preconditioned matrix is performed, which shows that its condition number depends on the variable coefficient in the partial differential equation as well as on the coarse mesh and overlap parameters. These results show that with a good choice of subdomains and coarse space basis functions, the preconditioner can still be robust even for large coefficient variation inside domains. Further, these estimates prove precisely that the previously made empirical observation that the use of low-energy coarse spaces can lead to robust preconditioners. Some numerical experiments are carried out in supporting the theoretical results.