Generating function exponential random graphs with expected number of edges ⟨m⟩

34 Views Asked by At

We refers to Statistical mechanics of networks by Park and Newman in the section. Random graphs Considering exponential random graphs with fixed number of vertices n we know only the expected number of edges $\langle m \rangle$. We have the adjacency matrix: $$\sigma_{ij}=\begin{cases} 1 & \text{if i connected to j} \\ 0 & \text{otherwise} \end{cases}$$

To write the degree distribution we can introduce the generating function: $$g(\alpha)= \sum_{k}p_ke^{\alpha k}$$ After some passages: $$g(\alpha)= \frac{\sum_{G \in \mathcal{G}}\prod_{k<l}exp[-(\theta+\alpha)\sigma_{kl}]}{(1+ e^{-\theta})^{\binom{n}{2}}} \stackrel{?}{=} \left(\frac{1+ e^{-(\theta+\alpha)}}{1+ e^{-\theta}}\right)^{n-1}$$ Does someone know how the last equality is possible? In my opinion should be: $$\frac{\sum_{G \in \mathcal{G}}\prod_{k<l}exp[-(\theta+\alpha)\sigma_{kl}]}{(1+ e^{-\theta})^{\binom{n}{2}}} \stackrel{?}{=} \left(\frac{1+ e^{-(\theta+\alpha)}}{1+ e^{-\theta}}\right)^{\binom{n}{2}}$$