Journal of Computations & Modelling

Krylov subspace methods for solving quadratic eigenvalue problems

  • Pdf Icon [ Download ]
  • Times downloaded: 90
  • Abstract

    In this paper, the Arnoldi-type process and symmetric Lanczos-type process for solving large scale quadratic eigenvalue problem (l^2A +lB+C)x=0 are given. One decomposition theorem about the matrices A, B and C is obtained based on the Householder transformation. The advantage of the Arnoldi-type process and symmetric Lanczos-type process is that they can preserve the matrix structure and properties of the original problems. Finally, some numerical examples are presented to show the efficiency of the proposed methods.