The bunkbed conjecture for graph percolation

Let GG be a graph, and let BB(G)BB(G) be its bunkbed graph with two layers, whose corresponding vertices are denoted xx^- and x+x^+ for xobreak in obreakV(G)x obreak\text{ in } obreak V(G). Let

μ ⁣:E(BB(G))[0,1]\mu \colon E(BB(G)) \to [0,1]

be a symmetric weight on BB(G)BB(G), and write PBB(G),μ(uv)\mathbb{P}_{BB(G),\mu}(u\sim v) for the probability that uu and vv are connected after independently retaining each edge according to its weight. Bunkbed conjecture. For every x,yV(G)x,y\in V(G),

PBB(G),μ(xy)PBB(G),μ(xy+).\mathbb{P}_{BB(G),\mu}(x^-\sim y^-)\geq \mathbb{P}_{BB(G),\mu}(x^-\sim y^+).

The conjecture asserts that, under symmetric edge percolation, vertices in the same layer are at least as likely to be connected as vertices in opposite layers. The paper proves that the conjecture is preserved when graphs are glued along a vertex, with forests as an immediate consequence; explicit proven families include complete graphs, complete bipartite graphs, and complete graphs minus one edge.

Sources & referencesView supporting material

Primary source

Paul Meunier and Pegah Pournajafi, “Vertex gluing preserves the bunkbed conjecture”, arXiv:2410.08957 (2026).

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.