Truncated \(QZ\) methods for large scale generalized eigenvalue problems (Q1279542)

This paper presents three methods for the large scale generalized eigenvalue problem \(Ax= \lambda Bx\). The methods are developed within a Krylov subspace projection framework as truncations of the QZ-algorithm for dense problems. These techniques provide natural extensions of the implicitly restarted Arnoldi method and the truncated RQ method for large scale generalized eigenvalue problems. These two methods require the accurate solution of linear systems at each step of the iteration. A two-block method is developed that incorporates \(k\) approximate Newton corrections at each iteration. Computational experience is presented to compare the three new methods.