A filter diagonalization for generalized eigenvalue problems based on the Sakurai–Sugiura projection method

摘要

The Sakurai–Sugiura projection method, which solves generalized eigenvalue problems to find certain eigenvalues in a given domain, was reformulated by using the resolvent theory. A new interpretation based on filter diagonalization was given, and the corresponding filter function was derived explicitly. A block version of the method was also proposed, which enabled not only resolution of degenerated eigenvalues, but also an improvement in numerical accuracy. Three numerical examples were provided to illustrate the method.