Regularized formulations of FETI

نویسندگان

  • Ismael Herrera
  • David E. Keyes
  • Olof B. Widlund
  • Robert Yates
  • Pavel Bochev
  • R. B. Lehoucq
چکیده

FETI 205 A conforming FEM for (2.7) results in the discrete optimality system   K1 0 B T 1 0 0 K̃2 −B2 −C2 Υ B1 −B2 0 0 0 −ΥC2 0 Υ     u1 u2 λ μ   =   f1 f2 0 0   (2.8) where K̃2 ≡ K2 +C2 ΥC2. Elimination of the primal variables in (2.8) results in the coarse grid problem [ B1K −1 1 B T 1 +B2K̃ −1 2 B T 2 B2K̃ −1 2 C T 2 Υ ΥC2K̃ −1 2 B T 2 ΥC2K̃ −1 2 C T 2 Υ−Υ ] [ λ μ ] = [ d1 d1 ] (2.9) where [ d1 d2 ] = [ B1K −1 1 f1 −B2K̃−1 2 f2 −ΥC2K̃−1 2 f2 ] . As compared with FETI-1, the columns of K̃−1 2 C T 2 Υ are approximating a basis for the rigid body modes associated with Ω2, and K̃ −1 2 is an approximation to the pseudoinverse of K2. Inserting the solution of the coarse grid problem (2.9) into (2.8) results in K1u1 = f1 −B1 λ (2.10) K̃2u2 = f2 +B T 2 λ+C T 2 Υμ. (2.11) These two linear systems have symmetric positive definite coefficient matrices and can be solved in parallel. We remark that (2.11) corresponds to the minimization problem inf v∈H1(Ω2) 1 2 ã(v, v)− 〈f̃ , v〉Ω2 where f̃ is the continuous load associated with the discrete load of (2.11). 3. FETI-SS: Regularization by space splitting. In this section we introduce a modification of FETI-1 that allows for a wider choice of well-posed primal problems for these domains. In particular, our approach results in nonsingular linear systems with properties that can be easily controlled. Our starting point is the splitting of H(Ω2) into the direct sum H(Ω2) = H 1 c (Ω2)⊕N2 where N2 is the RBM space for Ω2 and H c (Ω2) = {u ∈ H(Ω2) | c2(u) = 0}, is the complement space with respect to the moments c2. The report [1] demonstrates that such a splitting exists for any non-degenerate set of moments. As a result, any u2 ∈ H(Ω2) can be uniquely written as u2c +R2α where R2 is a basis for N2 and α ∈ R. To solve (2.1) we consider the problem of finding the saddle-point (u1, u2c, α, λ) ∈ H(Ω1, ∂Ω1)×H c (Ω2)× R p ×H(Γ) of the Lagrangian L(û1, û2c, α̂, λ̂) = 2 ∑ i=1 (1 2 a(ûi, ûi)Ωi − 〈f, ûi〉Ωi ) + 〈λ̂, û1 − (û2c +R2α̂)〉Γ. (3.1)

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

1 . Regularized formulations of FETI

FETI 3 Theorem 2.1 (u1, u2, λ) solves (2.4) if and only if (u1, u2, λ, μ = c(u2)) solves (2.7). Proof. The theorem is easily established by using the stabilized constraints (2.5) and recalling that τ = 0. The theorem demonstrates that (2.5) represents a consistent stabilization. The impact of this innocuous sleight of hand is that the resulting coarse grid problem is equivalently stabilized. We...

متن کامل

Preconditioners for the dual-primal FETI methods on nonmatching grids: Numerical study

FETI-DP method is a substructuring method that uses Lagrange multipliers to match the continuity condition on the subdomain boundaries. For the FETI-DP method on nonmatching grids, two different formulations are known with respect to how to employ the mortar matching condition. Keeping step with the developments of the FETI-DP methods, a variety of preconditioners for the FETI-DP operator have ...

متن کامل

FETI–DP, BDDC, and Block Cholesky Methods

Two popular non-overlapping domain decomposition methods, the FETI–DP and BDDC algorithms, are reformulated using Block Cholesky factorizations, an approach which can provide a useful framework for the design of domain decomposition algorithms for solving symmetric positive definite linear system of equations. Instead of introducing Lagrange multipliers to enforce the coarse level, primal conti...

متن کامل

A Practical Factorization of a Schur Complement for PDE-Constrained Distributed Optimal Control

A distributed optimal control problem with the constraint of a linear elliptic partial differential equation is considered. A necessary optimality condition for this problem forms a saddle point system, the efficient and accurate solution of which is crucial. A new factorization of the Schur complement for such a system is proposed and its characteristics discussed. The factorization introduces...

متن کامل

Some properties of regularization and penalization schemes for MPECs

Some properties of regularized and penalized nonlinear programming formulations of mathematical programs with equilibrium constraints (MPECs) are described. The focus is on the properties of these formulations near a local solution of the MPEC at which strong stationarity and a second-order sufficient condition are satisfied. In the regularized formulations, the complementarity condition is rep...

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 2003