,
Jacobus Conradi
,
Sariel Har-Peled
,
Antonia Kalb
,
Abhiruk Lahiri
,
Lukas Plätz
,
Carolin Rehs
,
Sampson Wong
Creative Commons Attribution 4.0 International license
We study the problem of computing a minimum-size Well-Separated Pair Decomposition (WSPD) of a given point set. We obtain the following results for the minimum-size WSPD: (1) a constant-factor approximation in doubling metrics, (2) a simple 3-approximation in ℝ, and (3) an NP-hardness proof in ℝ². We also provide an optimal output-sensitive runtime for the algorithm in doubling metrics and an implementation of the 3-approximation algorithm in ℝ. Furthermore, we introduce a new pair-decomposition for point sets in a metric space. It is defined using a relaxed requirement that, for all pairs {X,Y} in the decomposition, all the distances of pairs of points in X × Y are equal up to a factor in [1 ± ε]. Surprisingly, we show that in a general metric space, one can compute such a decomposition of size O(n/ε log n), which is dramatically smaller than the Θ(n²) bound for WSPDs. For a point set in ℝ^d, the bound improves to O(d n/ε log 1/ε).
@InProceedings{buchin_et_al:LIPIcs.ESA.2026.102,
author = {Buchin, Kevin and Conradi, Jacobus and Har-Peled, Sariel and Kalb, Antonia and Lahiri, Abhiruk and Pl\"{a}tz, Lukas and Rehs, Carolin and Wong, Sampson},
title = {{On Small Pair Decompositions for Point Sets}},
booktitle = {34th Annual European Symposium on Algorithms (ESA 2026)},
pages = {102:1--102:22},
series = {Leibniz International Proceedings in Informatics (LIPIcs)},
ISBN = {978-3-95977-445-1},
ISSN = {1868-8969},
year = {2026},
volume = {388},
editor = {Bille, Philip and Pettie, Seth and Storandt, Sabine},
publisher = {Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
address = {Dagstuhl, Germany},
URL = {https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ESA.2026.102},
URN = {urn:nbn:de:0030-drops-272380},
doi = {10.4230/LIPIcs.ESA.2026.102},
annote = {Keywords: Well-separated pair decomposition, Semi-separated pair decomposition, Approximation algorithms, Doubling metrics, Computational geometry}
}