Isolated-vertex threshold for FS(X,Y) connectivity
Informal Conjecture (§1.2, Theorem 1.1 remarks) · arXiv:2009.07840
Status partial medium confidence
Three follow-up papers have narrowed the gap in Theorem 1.1 of the source paper. Milojevic (arXiv:2210.03864, 2022) improved the isolated-vertex lower bound and disproved two formal conjectures from the source paper about threshold probabilities. Krishnan and Li (arXiv:2410.21334, 2024) established k-connectivity for p ≥ n^{-1/2+o(1)}, bringing the connectivity upper bound and isolated-vertex lower bound to the same asymptotic order n^{-1/2+o(1)}, broadly consistent with the informal conjecture. A sharp threshold matching an exact constant has not been proved.
Cited literature (3)
-
Extends the connectivity threshold analysis to the asymmetric case X ∈ G(n,p₁), Y ∈ G(n,p₂), showing FS(X,Y) is connected with high probability when p₁p₂ ≥ n^{-1+o(1)}, consistent with the isolated-vertex-threshold heuristic.
-
Slightly improves the isolated-vertex lower bound on the disconnectedness threshold for FS(X,Y) when X,Y ∼ G(n,p), showing isolated vertices persist up to p = O(log n / n^{1/2}), and in doing so disproves two formal conjectures of Alon, Defant, and Kravitz about specific threshold probabilities.
-
Shows that for independent G(n,p₁) and G(n,p₂) with p₁p₂ ≥ p₀² where p₀ = n^{-1/2+o(1)}, FS(X,Y) is k-connected with high probability, improving the connectivity upper bound in the symmetric case to the same n^{-1/2+o(1)} order as the isolated-vertex threshold.
Reviewer notes. The two formal conjectures disproved by Milojevic (arXiv:2210.03864) appear to be distinct formal conjectures in the source paper about specific threshold values, not the informal conjecture reviewed here; the full paper text could not be retrieved so the exact disproved statements are unconfirmed. The informal conjecture (connectivity threshold determined by isolated-vertex threshold) is broadly consistent with current bounds — both lower and upper bounds now sit at n^{-1/2+o(1)} — but the sharp threshold with an exact constant matching the isolated-vertex threshold has not been established.
Context
Theorem 1.1 establishes a lower bound $p \leq (2^{-1/2}-\varepsilon)/n^{1/2}$ arising from isolated vertices and an upper bound $p \geq \exp(2(\log n)^{2/3})/n^{1/2}$ for the high-probability connectivity threshold of $\mathrm{FS}(X,Y)$. The authors suggest the gap between these bounds is an artifact of the proof method and that the two thresholds should coincide up to lower-order factors.
Notes. PDF source; prose conjecture without a labelled environment. Section 7 (explicitly titled as containing open questions and conjectures) is absent from the supplied text, so further items from that section could not be extracted.
Source paper
Typical and Extremal Aspects of Friends-and-Strangers Graphs
Noga Alon, Colin Defant, Noah Kravitz · 2021-06-15
https://arxiv.org/abs/2009.07840
PDF source