On correction equations and domain decomposition for computing invariant subspaces
Abstract
This paper considers a number of schemes for computing an approximate invariant subspace associated with the smallest eigenvalues of a sparse symmetric (real) matrix. The approach taken is that of the so-called ``correction equation'' which leads to such standard schemes as the Jacobi-Davidson method or Olsen's method. We consider the situation of block corrections and discuss two algorithms. The application of the viewpoint that is developed is then explored for Domain Decomposition methods.
Loading...