Polynomial mixing time of the triangle-solitaire Markov chain

Prove that the Markov chain on triangle bases induced by the triangle-solitaire dynamics has polynomial mixing time, thereby enabling uniform sampling of triangle bases and, via the established bijection, uniform sampling of 3-permutations avoiding the patterns 12 and 312.

Background

The triangle-solitaire dynamics defines a Markov chain on triangle bases. The paper cites that the diameter of the associated reconfiguration graph is O(n3) and notes that the graph has strong connectivity properties.

The authors conjecture that the mixing time is polynomial. Establishing this would provide a method for uniformly sampling triangle bases; the bijection in the paper would then transfer such a sampling procedure to 3-permutations avoiding 12 and 312.

References

The diameter of the reconfiguration graph is $O(n3)$ , and we conjecture that its mixing time is also polynomial since the graph has strong connectivity properties.

3D permutations and triangle solitaire  (2502.14657 - Schabanel, 20 Feb 2025) in Section 5.2, “Random sampling”