Damien Fay, McGill University
(damien_fay (AT) yahoo.co.uk)[*]
Hamed Haddadi, Max Planck Institute for
Software Systems (hamed (AT) mpi-sws.org)[**]
The Weighted Spectral Toolbox is a matlab toolbox for use with undirected
graphs. There are several functions in the toolbox but the main ones obtain the
best fit (in the weighted spectral sense) of a topology generator to a target
graph. See the readme file for more information.
The WSD [1] toolbox primarily
provides matlab based scripts to perform the following:
1. Produce a weighted spectral
distribution for a given network (adjacency matrix).
2. Tune the parameters of a of
topology generator to match a given target graph.
3. Auxiliary functions:
1. Calculate the entropy rate of
a network,
2. Perform coarse graining of a network using a (adjacency matrix)
spectral approach [3],
3. Provide a matlab interface for BRITE AB, GLP, and
INET and PFP model generation,
4. Code for a fast and memory
efficient implementation of the waxman model,
5. Code for generation of a
wealth information topology [2],
6. Code for A Strawman (i.e.
toy) topology generator,
7. Native code to generate the
degree correlations in a network and including a fast algorithm to calc the
assortativity and graph auto-correlation function.
8. A script to analyse a graph
and produce multiple plots and tables of common graph features (requires
matlabBGL library)
[1] Weighted spectral
distribution,Damien Fay, Hamed Haddadi, Steve Uhlig, Andrew W. Moore, Richard
Mortier, Almerima Jamakovic, Cambridge university tech report, UCAM-CL-TR-729
ISSN 1476-2986, http://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-729.pdf
[2] Wealth-Based Evolution Model
for the Internet AS-Level Topology, Wang, X.; Loguinov, D., INFOCOM 2006. 25th
IEEE International Conference on Computer Communications. Proceedings, April
2006 Page(s):1 - 11
[3] D. Gfeller, P. De Los Rios,
Spectral Coarse Graining of complex networks, Physical review letters, 99,
038701, 2007
[*]
This work was started while Damien was a Research Associate at the Computer
Laboratory, University of Cambridge.
[**]
This work was started while Hamed was a visiting PhD student at the Computer
Laboratory, University of Cambridge.
