On Jun 8, 7:47 pm, Jason Grout <jason-s...@creativetrax.com> wrote:
> I've never had to use the left eigenvectors of a graph.

And I've never needed a directed graph so my matrices are always
symmetric.  (Just kidding.)  ;-)

The above is now:  http://trac.sagemath.org/sage_trac/ticket/6258

A related question:  matrices and graphs have eigenspaces() methods,
but matrices have eigenvalues() and graphs have spectrum().  The
spectrum() method seems identical in the style of its output to the
eigenvalues() method - a list of eigenvalues repeated according to
multiplicity.  Furthermore, in algebraic graph theory the term
"spectrum" is closely tied to both the values of the eigenvalues and
their multiplicities.

For example, graphs.HoffmanSingletonGraph() has 50 eigenvalues but
just three are distinct.  The usual shorthand is to list the
eigenvalues with their multiplicities as an exponent - for the Hoffman-
Singleton graph this would look like

7^1  2^28 (-3)^21

Yes, you can count the multiple eigenvalues, or write a routine to do
it for you, but I think this is something that should be built into
Sage.

So a proposal for graphs:

1.  Add a new eigenvalues() method for a graph,  identical to the
current spectrum(), more in line with the use for matrices.

2.  Have  spectrum() return something like a list of pairs (e_i, m_i)
with m_i being the multiplicity of eigenvalue e_i, more in line with
expectations for the term.  For example, "Algebraic Graph
Theory" (Godsil and Royle) says the spectrum is the "list of the
eigenvalues together with their multiplicities."  Biggs' "Algebraic
Graph Theory" says the spectrum "is the set of numbers which are
eigenvalues, together with their multiplicities."

Rob
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to 
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to