A Bijection Between the Recurrent Configurations of a Hereditary Chip-Firing Model and Spanning Trees

Hereditary chip-firing models generalize the Abelian sandpile model and the cluster firing model to an exponential family of games induced by covers of the vertex set. This generalization retains some desirable properties, e.g. stabilization is independent of firings chosen and each chip-firing equivalence class contains a unique recurrent configuration. In this paper we present an explicit bijection between the recurrent configurations of a hereditary chip-firing model on a graph and its spanning trees.

💡 Research Summary

The paper introduces a broad generalization of chip‑firing dynamics called the Hereditary Chip‑Firing Model (HCFM). In this framework a finite undirected graph G=(V,E) is equipped with a hereditary family 𝔽 of vertex subsets; “hereditary’’ means that whenever a set A belongs to 𝔽, every subset of A also belongs to 𝔽. Each set A∈𝔽 is interpreted as a cluster that may fire simultaneously, provided every vertex v∈A holds at least as many chips as the number of edges it has inside A (denoted deg_A(v)). When A fires, each v∈A loses deg_A(v) chips while each neighbor outside A receives one chip for every incident edge to A.

A central observation is that every firing operation can be expressed as an integer linear combination of columns of the graph Laplacian L. Consequently, the final stable configuration obtained after any sequence of legal firings is independent of the order in which the clusters are chosen. This “stabilization independence’’ mirrors the Abelian property of the classical sandpile model and guarantees that each chip‑distribution belongs to a well‑defined equivalence class modulo the lattice generated by the columns of L.

Within each equivalence class there exists a unique recurrent configuration. A recurrent configuration is defined as a stable chip distribution that returns to itself after adding a sufficiently large number of chips to every vertex and then stabilizing. The authors extend Dhar’s burning algorithm to the hereditary setting, producing a “burning test’’ that decides whether a given stable configuration is recurrent. The test proceeds by fixing a distinguished root vertex q, adding one chip to the configuration, and then tracing the reverse of the stabilization process: each vertex records the moment it first becomes “burned.’’

The main contribution of the paper is an explicit bijection between recurrent configurations of HCFM and spanning trees of G. The construction works in two directions.

1. From recurrent configurations to spanning trees: Starting from a recurrent configuration c, the authors add a chip at the root q and run the stabilization process backward. Whenever a vertex v becomes burned for the first time, they select the edge that connected v to the already‑burned component at that moment. Collecting all such edges yields a set T of |V|−1 edges that is acyclic and connects all vertices, i.e., a spanning tree.

2. From spanning trees to recurrent configurations: Given a spanning tree T rooted at q, the authors perform a “prufer‑like’’ traversal (a leaf‑removal order) that processes vertices from the leaves toward the root. For each vertex v in this order they compute the minimal number of chips that v must hold so that, when the remaining vertices of its incident subtree fire according to the hereditary rules, v can be burned at the appropriate step. The resulting chip vector c_T is stable and, by construction, is recurrent.

The authors prove that these two maps are inverses of each other, establishing a one‑to‑one correspondence. As a corollary, the number of recurrent configurations equals the number of spanning trees, which is known to be det(L̂), the cofactor of the Laplacian matrix, in accordance with Kirchhoff’s Matrix‑Tree Theorem.

Algorithmically, both directions can be implemented in polynomial time. The burning test and the leaf‑removal order are both realizable with standard breadth‑first or depth‑first search, leading to an overall O(|V|·|E|) complexity. This makes the bijection practical for large graphs and opens the door to uniform random sampling of recurrent configurations via random spanning trees.

The paper also situates HCFM within the landscape of existing models. When 𝔽 consists solely of singletons {v}, HCFM collapses to the classical Abelian sandpile model, and the bijection reduces to Dhar’s original bijection. When 𝔽 contains all connected subsets, HCFM becomes the cluster‑firing model, and the construction recovers the known bijection for that setting. Thus HCFM provides a unified framework that simultaneously generalizes both models.

Finally, the authors discuss several avenues for future work. Extending the hereditary framework to weighted graphs or directed graphs would lead to new combinatorial objects beyond ordinary spanning trees. The bijection suggests a method for reconstructing possible hereditary families from observed spanning trees, an inverse problem with potential applications in network inference. Moreover, the ability to sample recurrent configurations efficiently could be leveraged in the study of self‑organized criticality, electrical network reliability, and Markov chain mixing times.

In summary, the paper delivers a rigorous definition of hereditary chip‑firing dynamics, proves key structural properties (stabilization independence and unique recurrence), and constructs an explicit, algorithmically efficient bijection between recurrent configurations and spanning trees. This work not only unifies previous sandpile‑type models but also broadens the combinatorial toolkit available for analyzing complex firing processes on graphs.