On the Interface and Two-Level Preconditioners in Newton-Schwarz Method

نویسندگان

  • Tony Chan
  • Takashi Kako
  • Hideo Kawarada
  • Olivier Pironneau
  • Daniel Lee
چکیده

This paper is concerned with parallel computation in solving the convection-diffusion equation and the incompressible Navier-Stokes equation via Newton-Schwarz method, a nonlinear domain decomposition (DD) method. Various preconditioners are investigated here. An interface problem is tackled as a preconditioner for nonlinear block Jacobi DD approach, with an optional fine level interface problem solved as further preconditioner. Also a (global) coarse level preconditioner is considered. Examined also is the relaxation type preconditioner. Such preconditioned nonlinear DD methods exhibit impressive improvement over the basic non-preconditioned parallel NewtonJacobi method. A general review on Newton-Schwarz method is [GEMT98]. Our setup has the advantages of both the overlapped and the nonoverlap DD approach. The subdomain variables form a (nonoverlap) partition of the whole global system of equations. Solving the interface problem is regarded as a preconditioner to all subproblems. We note that the interface variables were excluded from subproblems in [LC98]. We describe in later sections the problem and solution procedure, the test cases and results, and a brief conclusion.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Nonlinear Preconditioning: How to Use a Nonlinear Schwarz Method to Precondition Newton's Method

For linear problems, domain decomposition methods can be used directly as iterative solvers, but also as preconditioners for Krylov methods. In practice, Krylov acceleration is almost always used, since the Krylov method finds a much better residual polynomial than the stationary iteration, and thus converges much faster. We show in this paper that also for non-linear problems, domain decomposi...

متن کامل

A New Class of Parallel Two-level Nonlinear Schwarz Preconditioned Inexact Newton Algorithms

We propose and test a new class of two-level nonlinear additive Schwarz preconditioned inexact Newton algorithms (ASPIN). The two-level ASPIN combines a local nonlinear additive Schwarz preconditioner and a global linear coarse preconditioner. This approach is more attractive than the two-level method introduced in [Cai, Keyes, and Marcinkowski, Nonlinear additive Schwarz preconditioners and ap...

متن کامل

A Monolithic Ale Newton-krylov Solver with Multigrid-richardson-schwarz Preconditioning for Incompressible Fluid Structure Interaction

In this paper we study a monolithic Newton-Krylov solver with exact Jacobian for the solution of incompressible FSI problems. A main focus of this work is on the use of geometric multigrid preconditioners with modified Richardson smoothers preconditioned by an additive Schwarz algorithm. The definition of the subdomains in the Schwarz smoother is driven by the natural splitting between fluid an...

متن کامل

Non - linear additive Schwarz preconditioners and application in computational uid dynamics 3

The focus of this paper is on the numerical solution of large sparse non-linear systems of algebraic 9 equations on parallel computers. Such non-linear systems often arise from the discretization of non-linear partial di erential equations, such as the Navier–Stokes equations for uid ows, using nite element 11 or nite di erence methods. A traditional inexact Newton method, applied directly to t...

متن کامل

Parallel Performance of Some Two-Level ASPIN Algorithms

In this paper we study the parallel performance of some nonlinear additive Schwarz preconditioned inexact Newton methods for solving large sparse system of nonlinear equations arising from the discretization of partial differential equations. The main idea of nonlinear preconditioning is to replace an ill-conditioned nonlinear system by an equivalent nonlinear system that has more balanced nonl...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2001