A tool for comparing the spectra of two matrices, e.g., the
eigenvalues of a matrix and its preconditioned form. Six ways
of plotting them are shown, pick whichever one best shows
the features you want.

This assumes the eigenvalues come from real-valued matrics.
In that case the eigenvalues for a matrix occur in conjugate
pairs. I.e., if x + iy is an eigenvalue, so is x - iy. In
that case only the upper plane for one matrix and the lower
plane for the second matrix need be plotting, helping reduce
the amount by which they overlay each other. For real-valued
eigenvalues, they overlap unfortunately.

If one or both of the matrices have almost all real-valued
eigenvalues, then this is not so useful and you're better off
to use a standard eigenvalue versus eigenvalue number plot, 
with two curves, one for each matrix.

    Randall Bramley
    Department of Computer Science
    Indiana University, Bloomington

    Started      :       23 Oct 1997, 05:38 AM
    Modified     : Thu 08 May 2008, 06:28 PM
    Last Modified: Mon 09 Mar 2009, 11:37 AM


=======
Files:
=======

A_eigenvalues.mat
B_eigenvalues.mat
    Two binary files with the eigenvalues of one matrix each:

eigenvalue_plot.m
    The main function. This demonstrates more techniques, esp.
    using polar plots and how to create a multiline text object
    using cells. Also, how to use graphics handles to get finer
    control over text and placement (see the legend() commands
    in the file).

    Read this file carefully - small things like the use of
    cells for multiline text are embedded, but unlike that 
    example are not commented upon. Without the comments lines
    before that multiline cell, would you have even noticed
    that braces {} where used instead of parentheses () or
    brackets [] ?

------------------------------------
Small drivers for the main function
------------------------------------
evdriver.m
compare.m