The tropical polytope is the set of all weighted tropical Fermat-Weber points
Shelby Cox, Mark Curiel
TL;DR
The paper generalizes tropical Fermat-Weber theory to incorporate positive weights, showing that weighted Fermat-Weber sets FW$(V,\mathbf{w})$ are covector cells of the min-tropical convex hull $\mathrm{tconv}^{\min}(V)$ and that every covector cell can be realized by suitable weights. It reframes the optimization as minimization of a homogeneous tropical signomial $f_{V,\mathbf{w}}$, connects covector decompositions to tropical convexity via the Cayley trick and Newton polytopes, and establishes a constructive method to realize any covector cell through fractional matchings. The results extend to phylogenetics by defining a weighted tropical consensus method that is regular and Pareto/co-Pareto on rooted triples, with FW points corresponding to equidistant trees in tropical tree space. This provides a geometry-driven, weight-aware framework for consensus in tree space and suggests practical avenues for weighted bootstrapping and efficient computation using transportation-like structures.
Abstract
Let $\mathbf{v}_1,\ldots,\mathbf{v}_m$ be points in a metric space with distance $d$, and let $w_1,\ldots,w_m$ be positive real weights. The weighted Fermat-Weber points are those points $\mathbf{x}$ which minimize $\sum w_i d(\mathbf{v}_i, \mathbf{x})$. We extend a result of Comăneci and Joswig, that the set of unweighted Fermat-Weber points agrees with the "central" covector cell of the tropical convex hull of $\mathbf{v}_1,\ldots,\mathbf{v}_m$, to the weighted setting. In particular, we show that for any fixed data points $\mathbf{v}_1, \ldots, \mathbf{v}_m$, and any covector cell of the tropical convex hull of the data, there is a choice of weights that makes that cell the Fermat-Weber set. We similarly extend the method of Comăneci and Joswig for computing consensus trees in phylogenetics.
