Algebraic characterization of binary graphs

One of the fundamental concepts in the statistical mechanics field is that of ensemble. Ensembles of graphs are collections of graphs, defined according to certain rules. The two most used ensembles in network theory are the microcanonical and the grandcanonical (whose definitions mimick the classical ones, originally proposed by Boltzmann and Gibbs), even if the latter is far more used than the former to carry on the analytical calculations. For binary (undirected or directed) networks, the grandcanonical ensemble is defined by considering all the graphs with the same number of vertices and a variable number of links, ranging from 0 to the maximum: N(N-1)/2 for binary, undirected graphs and N(N-1) for binary, directed graphs. Even if it is commonly used almost exclusively as a tool to calculate the average of some topological quantity of interest, its structure is so rich to deserve an analysis on its own. In this paper a logic-algebraic characterization of the grandcanonical ensemble of binary graphs is provided.

💡 Research Summary

**
The paper by Tiziano Squartini offers a rigorous, logic‑algebraic description of the grand‑canonical ensemble of binary (undirected or directed) graphs, a cornerstone object in statistical‑mechanics‑based network theory. While the grand‑canonical ensemble is routinely employed as a calculational device—allowing the number of links to fluctuate between zero and the maximal possible value—it has rarely been examined as a mathematical structure in its own right. The author fills this gap by showing that the ensemble can be understood simultaneously as a set of binary relations and as a Boolean algebra.

1. From statistical ensembles to graph ensembles
The paper begins by recalling the analogy between physical ensembles (micro‑canonical, canonical, grand‑canonical) and graph ensembles. In the grand‑canonical setting the number of vertices (N) is fixed, while the number of edges is allowed to vary from 0 up to the combinatorial maximum (N(N-1)/2) for undirected graphs and (N(N-1)) for directed graphs. This freedom mirrors the role of volume, chemical potential, and temperature in traditional thermodynamics.

2. Explicit construction via a binary tree
A constructive algorithm is presented: list all unordered vertex pairs ((i,j)) with (i<j) (or ordered pairs for directed graphs) and, for each pair, branch into two possibilities—edge absent (0) or present (1). Repeating this for all pairs yields a binary tree whose leaves correspond one‑to‑one with every possible adjacency matrix. For (N=4) undirected graphs the tree has depth 6 and produces (2^{6}=64) leaves; for directed graphs the depth is (N(N-1)) and the leaf count is (2^{N(N-1)}). This explicit enumeration demonstrates that the grand‑canonical ensemble is simply the support of a uniform distribution over a finite, discrete configuration space.

3. Graphs as binary relations
The author then recasts each adjacency matrix as a binary relation (R\subseteq V\times V). An entry (a_{ij}=1) means (v_i R v_j); (a_{ij}=0) means the ordered pair does not belong to the relation. This viewpoint allows the use of classical relational properties—reflexivity, symmetry, transitivity—to classify graphs.

• Reflexivity is measured by the trace (\mathrm{Tr}(A)=\sum_i a_{ii}). If (\mathrm{Tr}(A)=0) the relation is antireflexive (no self‑loops); if (\mathrm{Tr}(A)=N) it is reflexive (all self‑loops present); intermediate values indicate partial reflexivity.
• Symmetry corresponds to undirected graphs: (a_{ij}=a_{ji}) for all (i\neq j). The degree of symmetry is quantified by the reciprocity coefficient
  \