← All problems

Phase transition conjecture for the lcm-density of edge ideals

19 years old · traced to

Let GG range over all graphs on nn vertices, and let the \lcm\lcm-density refer to the lcm-density of the edge ideal of GG. Consider the behavior of this quantity across the set of all such graphs as nn tends to infinity.

Phase transition conjecture. The \lcm\lcm-density of edge ideals of graphs exhibits a phase transition phenomenon when considering the set of all graphs on nn vertices as n→∞n\to\infty.

The conjecture is motivated by computations showing low-, transition-, and high-graph-density regimes for graphs on six and seven vertices. The precise nature of the phase transition and its asymptotic behavior are not specified here, so the conjecture remains open.

References

7 papers in this index state this conjecture (2007–2026). The statement above is taken from the most recent of them; the others are arXiv:2108.04112, arXiv:2104.14009, arXiv:1811.10530, arXiv:1705.07448, arXiv:1009.3723, arXiv:math/0702369.

Source: Fatemeh Mohammadi, Sonja Petrović and Eduardo Sáenz-de-Cabezón, “Asymptotic properties of random monomial ideals”, arXiv:2605.03464 (2026).

Progress summary

Refreshed
Open

The conjecture remains open: computations suggest a sharp change in behavior, and only a version for random graphs is currently reported.

The conjecture asks whether the \lcm\lcm-density of edge ideals undergoes a phase transition across all graphs on nn vertices as n→∞n\to\infty. The May 2026 paper Asymptotic properties of random monomial ideals records the conjecture and reports computational evidence for three regimes.

Known results

  • Mohammadi et al. (2025) prove the analogous phase-transition statement for graphs drawn from the Erdős–Rényi model ER(n,p)ER(n,p), not for the full set of graphs.

May 2026 update

The latest paper reports low-, transition-, and high-density regimes in computations through 77 vertices, with transition values gd=0.4gd=0.4 for 66 vertices and gd=10/3gd=10/3 for 77 vertices. It explicitly gives no proof, counterexample, or resolution for all graphs.

Current status (as of October 2026): the restricted ER(n,p)ER(n,p) result is reported, but the conjecture for all graphs on nn vertices remains open.

Sources

Solutions 0

No solutions have been posted yet.