DINLE (v0.3): A MATLAB Package for Large-Scale Nonlinear Eigenvalue Problems (based on a Derivative Interpolating Subspace Framework)

This is a Matlab implementation of the subspace frameworks for large-scale nonlinear eigenvalue problems based on Hermite interpolation. At every subspace iteration, a projected eigenvalue problem of small-size is solved. Then the projection subspaces are expanded so as to attain Hermite interpolation between the full matrix-valued function and its projected counterpart at the eigenvalues of the projected problem before the subspace expansion.

Implementations specifically for polynomial and rational eigenvalue problems are available. But a general implementation for a nonlinear eigenvalue problem that is not necessarily polynomial nor rational is also available.