Time and space parallelization of the Navier-Stokes equations

Abstract

In this paper, we will be mainly concerned with a parallel algorithm (in time and space) which is used to solve the incompressible Navier-Stokes problem. This relies on two main ideas: (a) a splitting of the main differential operator which permits to consider independently the most important difficulties (nonlinearity and incompressibility) and (b) the approximation of the resulting stationary problems by a family of second-order one-dimensional linear systems. The same strategy can be applied to two-dimensional and three-dimensional problems and involves the same level of difficulty. It can be also useful for the solution of other more complicate systems like Boussinesq or turbulence models. The behavior of the method is illustrated with some numerical experiments.