Finite H-family distinguishing random graphs
Problem 12 · arXiv:2402.04237
Status open high confidence
Problem 12 from arXiv:2402.04237 asks whether a finite family of graphs $H_1,\dotsc,H_t$ exists such that for almost every $G\in\mathcal{G}(n,\tfrac{1}{2})$, the generalized chromatic polynomials $\pi_G^{(H_i)}$ jointly determine $G$ up to isomorphism. This is posed as a probabilistic weakening of the Bollobás–Pebody–Riordan conjecture (that almost every graph is determined by its chromatic polynomial), in light of the paper's Theorem 5 showing no finite family of generalized chromatic polynomials determines all graphs. No resolution or partial progress on this specific problem was found in the literature as of May 2026.
Reviewer notes. The source paper was published in the Electronic Journal of Combinatorics (2024, v31i4p48). No follow-up paper addressing Problem 12 specifically was found. The conjecture is recent (2024) and the absence of follow-up is consistent with open status at high confidence. The related Bollobás–Pebody–Riordan conjecture (chromatic polynomial determines almost every graph) itself remains open.
Context
Theorem 5 of the paper shows that no finite family of generalised chromatic polynomials can distinguish all graphs; this problem asks whether the situation is better for typical (Erdős–Rényi random) graphs. It is posed as a weakening of the Bollobás–Pebody–Riordan conjecture that almost every graph is determined by its chromatic polynomial.
Source paper
A note on graphs of $k$-colourings
Emma Hogan, Alex Scott, Youri Tamitegama, Jane Tan · 2024-02-29
https://arxiv.org/abs/2402.04237