- Accepted Paper
Coarse-graining directed networks with ergodic sets preserving diffusive dynamics
Phys. Rev. E - Accepted 12 November, 2026
DOI: https://doi.org/10.1103/g5ch-1hdg
Phys. Rev. E - Accepted 12 November, 2026
DOI: https://doi.org/10.1103/g5ch-1hdg
In this paper, we introduce a network compression algorithm based on ergodic sets, subsets of nodes of directed networks that are dynamically disjoint from the rest of the network. These subsets are either unreachable from the rest of the network or cannot be escaped once entered. We connect the definition of ergodic sets to purely structural considerations of the network and study some of their basic properties, as well as the prevalence of such structures in the Erdos-Renyi random graph and several networks taken from various areas of applications. Finally, we use these structures to develop a compression algorithm for directed networks that combines an exact ergodic decomposition with a low-rank representation of transport between source and sink sets.
If the author has provided any supplemental materials with this article they will be available upon publication of the version of record.