,
Markus Chimani
,
Henning Jasper
Creative Commons Attribution 4.0 International license
Given an undirected graph G with edge weights and lengths, a minimum α-spanner is a least-weight subgraph H ⊆ G that preserves distances w.r.t. the lengths between all node pairs up to a factor of α. Literature often takes the simplifying assumption of a single (coupled) edge function for weights and lengths. For such instances, several exact and non-exact algorithms are known and have been thoroughly evaluated in practice. However, many practical instances have decoupled form, as their weights and lengths are generally independent. Due to the increased complexity, only few (and even fewer practical) algorithms are able to guarantee low-weight solutions. This prompts practitioners to force their naturally decoupled instances into a coupled format, forsaking any quality guarantee. We implement several exact, approximative and heuristic algorithms for decoupled α-spanners, and use algorithm engineering to speed them up in practice. Our hypothesis-driven experiments evaluate their performance w.r.t. solution quality and speed. Generally, many practical instances can indeed be solved exactly within reasonable time, while LP-based approximation algorithms are not worthwhile. We find that standard greedy algorithms often yield acceptable results, but there are also practical instances for which they yield arbitrarily poor solutions. Here, augmented greedy variations offer a good compromise between solution quality and speed.
@InProceedings{bokler_et_al:LIPIcs.SEA.2026.8,
author = {B\"{o}kler, Fritz and Chimani, Markus and Jasper, Henning},
title = {{General Multiplicative Spanners in Practice}},
booktitle = {24th International Symposium on Experimental Algorithms (SEA 2026)},
pages = {8:1--8:21},
series = {Leibniz International Proceedings in Informatics (LIPIcs)},
ISBN = {978-3-95977-422-2},
ISSN = {1868-8969},
year = {2026},
volume = {371},
editor = {Aum\"{u}ller, Martin and Finocchi, Irene},
publisher = {Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
address = {Dagstuhl, Germany},
URL = {https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.SEA.2026.8},
URN = {urn:nbn:de:0030-drops-260120},
doi = {10.4230/LIPIcs.SEA.2026.8},
annote = {Keywords: Graph spanners, ILP, experimental study, algorithm engineering}
}