The Cellular Homology of Digraphs
Xinxing Tang, Shing-Tung Yau
TL;DR
This work builds a cellular homology theory for digraphs by restricting GLMY path theory to strongly regular minimal paths that admit singular cubical realizations. Central to the construction is the admissible-pair framework $(P, ext{Supp}(P))$, which yields a quotient cellular chain complex $C_n(G;R)$ and a boundary $partial^{ ext{cell}}$ compatible with the path-boundary on supports; the resulting homology $H^{ ext{cell}}_*(G;R)$ enjoys functoriality and, in the acyclic setting, homotopy invariance. The authors establish a Künneth-type formula for digraph products and demonstrate the theory on circulant and Johnson digraphs to illustrate computations and deformation-retraction arguments. They further discuss the relationship between cellular and singular cubical homology, conjecturing an isomorphism $H^{ ext{cell}}_*(G;R)\u2261 H^c_*(G;R)$ for acyclic graphs and interpreting admissible relations as cubical-boundary conditions. Overall, the paper provides a robust bridge between discrete digraph topology and classical topological invariants, offering concrete computation tools and a geometric interpretation via CW-type structures on digraphs.
Abstract
In \cite{TY}, we investigate the pair $(P, \Supp(P))$ of minimal path $P$ and its supporting sub-digraph $\Supp(P)$ in the path complex of a digraph $G$ under the strongly regular condition. In this paper, first, we consider the special minimal path $P$ specified by the admissible condition (Definition \ref{admpair}), which means that $(P,\Supp(P))$ admits a singular cubical realization. Based on such a subset, we systematically introduce the definitions of cellular chain complex associated to $G$ and prove the well-definedness. Then we study several properties of such cellular homologies. Finally, we present several intriguing examples as well as some important observations.
