Zentrum für Informationsdienste und Hochleistungsrechnen ( ZIH ) A primal-dual Jacobi – Davidson-like method for nonlinear eigenvalue problems

We propose a method for the solution of non-Hermitian ill-conditioned nonlinear eigenvalue problems T (λ)x = 0, which is based on singularity theory of nonlinear equations. The singularity of T (λ) is characterized by a scalar condition μ(λ) = 0 where the singularity function μ is implicitly defined by a nonsingular linear system with the appropriately bordered matrix T (λ). One Newton step for μ(λ) = 0 is performed which leads to a generalized Rayleigh quotient as new eigenvalue approximation, and the bordering vectors are updated, too. Quadratic convergence is shown. Rearranging of the system leads to equivalent Jacobi–Davidson-like correction equations. An important difference to standard JD is that the correction equations do not depend on the derivative of T and that only orthogonal projectors occur. Numerical examples demonstrate a considerably better performance of the new method compared to standard Jacobi–Davidson when applied to highly nonnormal eigenproblems.

[1]  Karl Meerbergen,et al.  The Quadratic Eigenvalue Problem , 2001, SIAM Rev..

[2]  Gerard L. G. Sleijpen,et al.  A Jacobi-Davidson Iteration Method for Linear Eigenvalue Problems , 1996, SIAM J. Matrix Anal. Appl..

[3]  A. Neumaier RESIDUAL INVERSE ITERATION FOR THE NONLINEAR EIGENVALUE PROBLEM , 1985 .

[4]  Axel Ruhe ALGORITHMS FOR THE NONLINEAR EIGENVALUE PROBLEM , 1973 .

[5]  L. B. Rall,et al.  The solution of characteristic value-vector problems by Newton's method , 1968 .

[6]  M. Hochstenbach,et al.  Two-sided and alternating Jacobi-Davidson , 2001 .

[7]  Jack Dongarra,et al.  Templates for the Solution of Algebraic Eigenvalue Problems , 2000, Software, environments, tools.

[8]  Peter Lancaster,et al.  On the numerical solution of nonlinear eigenvalue problems , 1995, Computing.

[9]  Heinrich Voß,et al.  Numerical methods for sparse nonlinear eigenvalue problems , 2004 .

[10]  Alastair Spence,et al.  Photonic band structure calculations using nonlinear eigenvalue techniques , 2005 .

[11]  V. Mehrmann,et al.  Nonlinear eigenvalue problems: a challenge for modern eigenvalue methods , 2004 .

[12]  H. Schwetlick,et al.  A generalized Rayleigh quotient iteration for computing simple eigenvalues of nonnormal matrices , 2000 .

[13]  Peter Lancaster,et al.  Lambda-matrices and vibrating systems , 2002 .

[14]  Eugene L. Allgower,et al.  A General View of Minimally Extended Systems for Simple Bifurcation Points , 1997 .

[15]  Timo Betcke,et al.  A Jacobi-Davidson-type projection method for nonlinear eigenvalue problems , 2004, Future Gener. Comput. Syst..