Paperlane

no. 16 · Tuesday 17 November 2026

State-space symmetry in Boolean network ensembles

Yi-Kai Mo and Kevin E. Bassler

math.DS · via cond-mat.stat-mech · 17 pp

Abstract

Studies of evolving Boolean network systems have found that Boolean functions cluster into equivalence classes with a characteristic frequency distribution - a pattern governed by the Zyklenzeiger symmetry group - but the dynamical origin of this symmetry has remained unexplained. Here we show that it is a consequence of a fundamental symmetry of Boolean network dynamics itself. A Boolean network consists of nodes that update according to Boolean functions of their regulatory inputs; its global dynamics is encoded in a state-space network. We identify the dynamical symmetry group of ensembles of Boolean networks that average over random interaction topologies: for networks with $N$ nodes, it is the $N$-dimensional hyperoctahedral group, generated by node-flips and node-label permutations. These transformations are admissible for every Boolean network regardless of topology; for random-topology ensembles, including all ensembles of fixed in-degree $K<N$, they constitute the complete symmetry group. We also show that input-label permutations constitute a hidden gauge freedom of the dynamics. The Zyklenzeiger symmetry of Boolean functions follows directly from the node-flips of the hyperoctahedral group together with this gauge freedom. Boolean networks related by the hyperoctahedral group have isomorphic state-space networks and form its orbits, and Burnside's lemma gives the number of these orbits for any fixed network topology. It provides a symmetry reduction of the analysis of Boolean network dynamics and explains why the Boolean functions of evolved networks fall into Zyklenzeiger classes, and why these classes split in finite networks.

comments
17 pages, 7 figures
fields
Dynamical Systems
pdf ↗ page arXiv:2611.19639 ↗