RIM is a general eigensolver for computing all eigenvalues in a region on the complex plane. The method tests if a region contains eigenvalues using an indicator, which is based on approximate spectral projection. Regions that contain eigenvalues are subdivided and tested until eigenvalues are isolated with a specified precision. It needs no a priori spectral information and overcomes some difficulties of existing methods. RIM is highly parallel and scalable.

Usage: lambda=RIMCv0(A,B,s)
The code computes generalized eigenvalues of
Ax = lambda Bx
in a rectangle s on the complex plane.
Input:
A -- N x N matrix
B -- N x N matrix
s -- 4 x 1 vector [xmin ymin xmax ymax]
Output: lambda -- generalized eigenvalues in s
Jiguang Sun, 05/09/2017, jiguangs@mtu.edu
Please report bugs to jiguangs@mtu.edu

Copyright (c) 2017, Jiguang Sun, all rights reserved.
THIS SOFTWARE IS PROVIDED "AS IS".
Redistribution and use in source and binary forms, with or without modification, for academic purpose only are permitted.