English

Cutting corners cheaply, or how to remove Steiner points

Data Structures and Algorithms 2015-08-21 v2

Abstract

Our main result is that the Steiner Point Removal (SPR) problem can always be solved with polylogarithmic distortion, which answers in the affirmative a question posed by Chan, Xia, Konjevod, and Richa (2006). Specifically, we prove that for every edge-weighted graph G=(V,E,w)G = (V,E,w) and a subset of terminals TVT \subseteq V, there is a graph G=(T,E,w)G'=(T,E',w') that is isomorphic to a minor of GG, such that for every two terminals u,vTu,v\in T, the shortest-path distances between them in GG and in GG' satisfy dG,w(u,v)dG,w(u,v)O(log5T)dG,w(u,v)d_{G,w}(u,v) \le d_{G',w'}(u,v) \le O(\log^5|T|) \cdot d_{G,w}(u,v). Our existence proof actually gives a randomized polynomial-time algorithm. Our proof features a new variant of metric decomposition. It is well-known that every nn-point metric space (X,d)(X,d) admits a β\beta-separating decomposition for β=O(logn)\beta=O(\log n), which roughly means for every desired diameter bound Δ>0\Delta>0 there is a randomized partitioning of XX, which satisfies the following separation requirement: for every x,yXx,y \in X, the probability they lie in different clusters of the partition is at most βd(x,y)/Δ\beta\,d(x,y)/\Delta. We introduce an additional requirement, which is the following tail bound: for every shortest-path PP of length d(P)Δ/βd(P) \leq \Delta/\beta, the number of clusters of the partition that meet the path PP, denoted ZPZ_P, satisfies Pr[ZP>t]2eΩ(t)\Pr[Z_P > t] \le 2e^{-\Omega(t)} for all t>0t>0.

Keywords

Cite

@article{arxiv.1304.1449,
  title  = {Cutting corners cheaply, or how to remove Steiner points},
  author = {Lior Kamma and Robert Krauthgamer and Huy L. Nguyen},
  journal= {arXiv preprint arXiv:1304.1449},
  year   = {2015}
}
R2 v1 2026-06-21T23:54:03.294Z