In this talk, we present a method for solving linear
and nonlinear eigenvalue problems using contour integration.
First, we talk about the case for generalized eigenvalue
problems (GEPs). The major advantage of our method is that it
is suitable for distributed computation because iterative

process for constructing subspaces is not required.
A parallel implementation and the estimation of an integral
path using the algebraic sub-structuring method are mentioned.
Next, we present an extension of our method for nonlinear

eigenvalue problems (NEPs). Some numerical experiments
illustrate numerical properties of the proposed method.