Math-1715 · mathematics · Probability

Math-1715

Isoperimetric dimension and nontrivial percolation threshold

Status Open · unsolvedmath.comLLM status Open unverified

Top Open Problems Not reported cataloguedOpenOpen

Statement

Definitions and mathematical statement

Let $G=(V,E)$ be an infinite connected graph with uniformly bounded degree. For a finite vertex set $A$, let $\partial_G A$ be its external vertex boundary. Its isoperimetric dimension is \[ \operatorname{Dim}(G)=\sup\left\{s\ge1:\exists c>0\ \forall A\subset V, \ 0<|A|<\infty,\quad |\partial_G A|\ge c|A|^{(s-1)/s}\right\}. \] For Bernoulli bond percolation with parameter $p$, independently retain each edge with probability $p$. If $o$ is any fixed vertex, define \[ p_c(G)=\inf\{p\in[0,1]:\mathbb P_p(o\text{ lies in an infinite open component})>0\}. \] Connectedness makes this threshold independent of the root. Prove the implication $\operatorname{Dim}(G)>1\Longrightarrow p_c(G)<1$.

The isoperimetric inequality is required for every finite set, with a constant independent of its position and cardinality. No transitivity, polynomial growth, planarity, or symmetry is assumed. The desired parameter below one may depend on the entire graph; no universal upper bound for all such graphs is requested.

The source separately proposes the weaker case $\operatorname{Dim}(G)=\infty$. That case is included in the established theorem of Duminil-Copin, Goswami, Raoufi, Severo and Yadin: dimension greater than four suffices. Thus the unresolved target retained here is the full implication above one, in particular the range $1<\operatorname{Dim}(G)\le4$. This distinction separates the full target from the already established weaker high-dimensional case.

Short English statement

Does any uniform isoperimetric dimension strictly above one guarantee percolation after deleting a sufficiently small positive fraction of edges?

Sources

Catalogue entry

Statement licence: CC-BY-4.0 · Source file

Contributions 0

No attempts yet. Significant partial results and substantial reductions are welcome. Contribute an attempt.