Re: [igraph] Sneppen-Maslov Randomization

[Tamas Nepusz]

Hi Lorenzo,

> However, I would like to know if there is already an implementation  
> (in particular for igraph under R) of the Sneppen-Maslov algorithm,  
> a randomization technique that preserves P(k).
Yes, there is:

> library(igraph)
> ?rewire

> (3) Given e.g. a graph with 1000 edges, how many i--j and k--l  ---- 
> > i--l and k--j valid edge transformations should I perform before I  
> can consider the graph randomized?
The rewire() function accepts a number which specifies how many  
rewiring trials should be made. (A rewiring trial may fail if the two  
selected edges i--j and k--l have a common endpoint). If you pass the  
number of edges here, chances are that every edge will be chosen for  
rewiring twice on average as the edges are chosen uniformly. I think  
there are more sophisticated bounds on the desired value of the number  
of iterations, though.

--
Tamas