Nonequivalence deflation for the solution of matrix latent value problems (Q1906758): Difference between revisions

Newton's method is usually used to solve nonlinear latent value problems. The deflation technique is required for finding another latent pair starting from a computed latent pair. In this paper, several deflation strategies are examined, and a nonequivalence deflation technique is developed. The major idea is transforming the computed latent root in the given region into infinity through a suitable nonequivalent transformation, while the other latent roots are preserved. Analysis and numerical examples show that the deflation combined with a suitable restarting initial guess is reliable and effective.