Cone-constrained eigenvalue problems: Theory and algorithms (Q2268921)

This work deals with the analysis of cone-constrained eigenvalue problems. Special attention is paid to the Paretian case. The authors introduce and study two algorithms for solving numerically such type of eigenvalue problems: one is the scaling-and-projection algorithm (SPA), the other is a variant of the SPA called CSPA. Both algorithms can be applied to cone-constrained eigenvalue problems involving an arbitrary closed convex cone, and not just the Pareto cone. Some theoretical issues of the cone-constrained eigenvalue problems are also discussed.