Nonlinear eigenvalue problems in Julia: Iterative methods and benchmarks
Clone or download
Latest commit 7599fd6 Dec 8, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.


Build Status codecov

A nonlinear eigenvalue problem is the problem to determine a scalar λ and a vector v such that


where M is an nxn-matrix depending on a parameter. This package aims to provide state-of-the-art algorithms to solve this problem, as well as a framework to formulate applications and easy access to benchmark problems. This currently includes (but is not restricted to) Newton-type methods, Subspace methods, Krylov methods, contour integral methods, block methods, companion matrix approaches. Problem transformation techniques such as scaling, shifting, deflating are also natively supported by the package.

How to use it?

On Julia 1.0.X and Julia 0.7, install it as a registered package by typing ] add ... at the REPL-prompt:

julia> ]
(v1.0) pkg> add NonlinearEigenproblems

After that, check out "Getting started" in

NEP-PACK online user's guide

or read the preprint:

Older Julia versions:

We don't support Julia 0.6, but it is possible to use it by running

julia> Pkg.clone("git://");

and switch to the git-tag v0.1.

GIT Version

If you want the cutting edge development version and not the latest release, install it with the URL:

julia> ]
(v1.0) pkg> add git://

NEP solvers

These solvers are currently available (see

  • Newton & Rayleigh type:
    • Classical Newton-Raphson
    • Augmented Newton
    • Residual inverse iteration
    • Quasi-Newton
    • Block Newton
    • Rayleigh functional iteration (RFI a, b)
    • Newton-QR
    • Implicit determinant method
    • Broyden's method
  • Arnoldi/Krylov type
    • NLEIGS
    • Infinite Arnoldi method (IAR, TIAR, Infinite bi-Lanczos)
  • Projection methods
    • Jacobi-Davidson (two versions)
    • Nonlinear Arnoldi method
  • Contour integral
    • Beyn's contour integral method
  • Deflation
    • Effenberger style deflation


The main work of NEP-PACK has been done in a closed repository at KTH, but as of May 2018 the development is carried out in a public github repo.

Core developers (alphabetical): Max Bennedich, Elias Jarlebring (, Giampaolo Mele (, Emil Ringh (, Parikshit Upadhyaya ( Thanks to A Koskela for involvement in initial version of the software.

How to cite

If you find this software useful please cite

  author = 	 {E. Jarlebring, M. Bennedich, G. Mele, E. Ringh, P. Upadhyaya},
  title = 	 {{NEP-PACK}: A {Julia} package for nonlinear eigenproblems},
  year = 	 {2018},
  note = 	 {},
  eprint = {arXiv:1811.09592},

If you use a specific method, please also give credit to the algorithm researcher. Reference to a corresponding algorithm paper can be found by in, e.g., by writing ?resinv.

Some links below are developer info on KTH. We will migrate them soon: