English

Pattern Matching in Doubling Spaces

Data Structures and Algorithms 2020-12-22 v1 Computational Geometry

Abstract

We consider the problem of matching a metric space (X,dX)(X,d_X) of size kk with a subspace of a metric space (Y,dY)(Y,d_Y) of size nkn \geq k, assuming that these two spaces have constant doubling dimension δ\delta. More precisely, given an input parameter ρ1\rho \geq 1, the ρ\rho-distortion problem is to find a one-to-one mapping from XX to YY that distorts distances by a factor at most ρ\rho. We first show by a reduction from kk-clique that, in doubling dimension log23\log_2 3, this problem is NP-hard and W[1]-hard. Then we provide a near-linear time approximation algorithm for fixed kk: Given an approximation ratio 0<ε10<\varepsilon\leq 1, and a positive instance of the ρ\rho-distortion problem, our algorithm returns a solution to the (1+ε)ρ(1+\varepsilon)\rho-distortion problem in time (ρ/ε)O(1)nlogn(\rho/\varepsilon)^{O(1)}n \log n. We also show how to extend these results to the minimum distortion problem in doubling spaces: We prove the same hardness results, and for fixed kk, we give a (1+ε)(1+\varepsilon)-approximation algorithm running in time ((dist(X,Y)/ε)O(1)n2logn(X,Y)/\varepsilon)^{O(1)}n^2\log n, where dist(X,Y)(X,Y) denotes the minimum distortion between XX and YY.

Keywords

Cite

@article{arxiv.2012.10919,
  title  = {Pattern Matching in Doubling Spaces},
  author = {Corentin Allair and Antoine Vigneron},
  journal= {arXiv preprint arXiv:2012.10919},
  year   = {2020}
}

Comments

23 pages, 4 figures

R2 v1 2026-06-23T21:06:29.861Z