Double-permutation sequences and pseudoline transversals

J. Goodman

DOI: https://doi.org/10.1145/1064092.1064094

2005-06-06

Abstract:In joint work with Ricky Pollack about 25 years ago we defined an encoding of a configuration of points in the plane by what we called its circular sequence of permutations. This notion was subsequently used, both by us and others, to study an array of problems in combinatorial geometry.Recently we noticed that a similar technique can be used to encode combinatorial properties of a collection of mutually disjoint compact convex sets in the plane, and that we can use it, for example, to give a new proof of the Edelsbrunner-Sharir "(2n-2)-theorem" on the maximum number of geometric permutations of a collection of such sets. At the same time this new proof generalizes the Edelsbrunner-Sharir result to the case of a family of compact connected sets in the plane, separated by pseudolines, every two having four double-tangent pseudolines, with all the pseudolines together forming an arrangement. In this talk we will describe the encoding, define the notion of geometric permutation for such a collection, and discuss the proof.The same encoding, which associates to a family of n sets a sequence of permutations of 1, 1, 2, 2,..., n, n, shows promise of being applicable to a number of other problems concerning combinatorial properties of collections of convex sets.The talk is based on ongoing joint work with Ricky Pollack.