The sandpile Betti-number conjecture via connected partitions

At least 14 years old · documented by

Let G=(V,E,s)G=(V,E,s) be an undirected sandpile graph. For U⊆VU\subseteq V, let G∣UG|_U be the subgraph induced by UU. A connected kk-partition of GG is a partition Π=⨆i=1kVi\Pi=\bigsqcup_{i=1}^k V_i such that each G∣ViG|_{V_i} is connected. Let GΠG_\Pi be the graph whose vertices are V1,…,VkV_1,\dots,V_k, with edge weights

wt⁡(Vi,Vj)=#{e∈E:one endpoint of e is in Vi and the other is in Vj}.\operatorname{wt}(V_i,V_j)=\#\{e\in E:\text{one endpoint of }e\text{ is in }V_i\text{ and the other is in }V_j\}.

Treat GΠG_\Pi as a sandpile graph with sink vertex ViV_i for the unique ii such that s∈Vis\in V_i, and let Pk\mathcal{P}_k denote the set of connected kk-partitions of GG. The sandpile Betti-number conjecture.

βk=∑Π∈Pk+1#{c:c a minimal recurrent configuration on GΠ}.\beta_k=\sum_{\Pi\in\mathcal{P}_{k+1}}\#\{c:\text{$c$ a minimal recurrent configuration on }G_\Pi\}.

This conjecture identifies the Betti numbers of the relevant sandpile-related algebra with counts of minimal recurrent configurations on partition graphs. It appears as Corollary 3.29 in work of Wilmes and has been verified using Sage for all undirected, unweighted graphs with fewer than 77 vertices; the general assertion remains open.

References

Primary source

David Perkinson, Jacob Perlman and John Wilmes, “Primer for the algebraic geometry of sandpiles”, arXiv:1112.6163 (2011).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.