Fractionally Log-Concave and Sector-Stable Polynomials: Counting Planar Matchings and More

TL;DR

This paper develops new polynomial properties called sector-stability and fractional log-concavity to analyze mixing times of Glauber dynamics, enabling efficient approximate counting and sampling of matchings and other combinatorial structures in planar graphs.

Contribution

It introduces sector-stability and fractional log-concavity as new tools for analyzing polynomial-based distributions, leading to FPRAS for counting matchings in planar graphs.

Findings

Multi-site Glauber dynamics mix rapidly for monomer-dimer systems.

Efficient sampling from partition-constrained strongly Rayleigh distributions.

Polynomials avoiding roots in a sector are fractional log-concave.

Abstract

We show fully polynomial time randomized approximation schemes (FPRAS) for counting matchings of a given size, or more generally sampling/counting monomer-dimer systems in planar, not-necessarily-bipartite, graphs. While perfect matchings on planar graphs can be counted exactly in polynomial time, counting non-perfect matchings was shown by [Jer87] to be #P-hard, who also raised the question of whether efficient approximate counting is possible. We answer this affirmatively by showing that the multi-site Glauber dynamics on the set of monomers in a monomer-dimer system always mixes rapidly, and that this dynamics can be implemented efficiently on downward-closed families of graphs where counting perfect matchings is tractable. As further applications of our results, we show how to sample efficiently using multi-site Glauber dynamics from partition-constrained strongly Rayleigh distributions, and nonsymmetric determinantal point processes. In order to analyze mixing properties of the multi-site Glauber dynamics, we establish two notions for generating polynomials of discrete set-valued distributions: sector-stability and fractional log-concavity. These notions generalize well-studied properties like real-stability and log-concavity, but unlike them robustly degrade under useful transformations applied to the distribution. We relate these notions to pairwise correlations in the underlying distribution and the notion of spectral independence introduced by [ALO20], providing a new tool for establishing spectral independence based on geometry of polynomials. As a byproduct of our techniques, we show that polynomials avoiding roots in a sector of the complex plane must satisfy what we call fractional log-concavity; this extends a classic result established by [Gar59] who showed homogeneous polynomials that have no roots in a half-plane must be log-concave over the positive orthant.