The evolution of the permutahedron

Abstract

In their seminal paper introducing the theory of random graphs, Erd\H{o}s and R\'{e}nyi considered the evolution of the structure of a random subgraph of $K_n$ as the density increases from $0$ to $1$, identifying two key points in this evolution -- the \emph{percolation threshold}, where the order of the largest component seemingly jumps from logarithmic to linear in size, and the \emph{connectivity threshold}, where the subgraph becomes connected. Similar phenomena have been observed in many other random graph models, and in particular, works of Ajtai, Koml\'{o}s and Szemer\'{e}di and of Spencer and Erd\H{o}s determine corresponding thresholds for random subgraphs of the hypercube. We study similar questions on the \emph{permutahedron}. The permutahedron, like the hypercube, has many different equivalent representations, and arises as a natural object of study in many areas of combinatorics. In particular, as a highly-symmetric simple polytope, like the $n$-simplex and $n$-cube, this percolation model naturally generalises the Erd\H{o}s-R\'{e}nyi random graph and the percolated hypercube. We determine the percolation threshold and the connectivity threshold for random subgraphs of the permutahedron. Along the way we develop a novel graph exploration technique which can be used to find exponentially large clusters after percolation in high-dimensional geometric graphs and we initiate the study of the isoperimetric properties of the permutahedron.