MINISYMPOSIUM 4: Domain Decomposition Methods Motivated by the Physics of the Underlying Problem

Domain decomposition methods are a powerful tool to handle very large systems of equations. They can however also be used to couple different physical models or approximations, which one might want to do for various reasons: in fluid structure coupling for example, the physical laws in the fluid differ from the physical laws in the structure, and a domain decomposition method could naturally take this into account. Even if the physical model is the same, one might want to use a simplified equation in part of the domain, where certain effects are negligible, like for example in aerodynamics, to save computation time. Or one could simply want to use a much coarser mesh, like in combustion away from the flame front, which again could be taken naturally into account by a domain decomposition method that can handle non-matching grids, possibly in space and time.