0
$\begingroup$

Hi, does anyone know if it is known what is the number of undirected graphs with the following properties:

  • Number of nodes: $N$, a big number,
  • Average degree: $z_1$,
  • Average number of outgoing edges $z_2$, where an outgoing edges are all edges leaving one of your neighbors (except the one that connects you to that neighbor).

?

I interested in asymptotic results for large N.

Similarly, the entropy of an ensemble of graphs such that the average degree is $z_1$ and the average number of outgoing edges is $z_2$ is also of interest.

EDIT: The definition of $z_2$ is here: http://www.santafe.edu/media/workingpapers/00-07-042.pdf on section F, "Numbers of neighbors and average path length". Sorry for the loose definition. $z_2$ is also the number of second neighbors.

Thanks.

$\endgroup$
4
  • $\begingroup$ Are the nodes labelled, or not? (That is, are we counting up to isomorphism, or not?) $\endgroup$ Commented Jul 4, 2011 at 19:54
  • $\begingroup$ Yes, the nodes are labelled. Sorry, I should have said this before. $\endgroup$
    – Rorsa
    Commented Jul 5, 2011 at 14:19
  • $\begingroup$ Also, suppose v,w,and u are vertices of a triangle in the graph. How are the triangle edges weighted in the definition of z_2? I can see the edges counted once each or twice each. Gerhard "Email Me About System Design" Paseman, 2011.07.05 $\endgroup$ Commented Jul 5, 2011 at 21:20
  • $\begingroup$ If (v, w, u) is a triangle, the edge (w,u) is to be counted once in the set of "outgoing edges" of the node v. $\endgroup$
    – Rorsa
    Commented Jul 6, 2011 at 23:00

1 Answer 1

1
$\begingroup$

I do not understand what your outgoing edges are but the number of undirected graphs with $N$ labeled vertices and average degree $d$ (which you call $z_1$ but $d$ seems more natural to me) is approximatively given by $${2^{N(N-1)/2}\choose dN/2}$$ since this is the number of all possible adjacency matrices.

The asymptotics for fixed $d$ and $N\rightarrow\infty$ are then given by $$2^{dN^2(N-1)/4}\frac{1}{(dN/2)!}$$ which can be made more explicit using Stirling's formula.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .