A self-stabilizing 3-approximation for the maximum leaf spanning tree problem in arbitrary networks

Abstract : The maximum leaf spanning tree (MLST) is a good candidate for constructing a virtual backbone in self-organized multihop wireless networks, but is practically intractable (NP-complete). Self-stabilization is a general technique that permits to recover from catastrophic transient failures in self-organized networks without human intervention. We propose a fully distributed self-stabilizing approximation algorithm for the MLST problem in arbitrary topology networks. Our algorithm is the first self-stabilizing protocol that is specifically designed to approximate an MLST. It builds a solution whose number of leaves is at least 1/3 of the maximum possible in arbitrary graphs. The time complexity of our algorithm is O(n^2) rounds.
Complete list of metadatas

https://hal.sorbonne-universite.fr/hal-00930035
Contributor : Sébastien Tixeuil <>
Submitted on : Tuesday, January 14, 2014 - 11:22:05 AM
Last modification on : Tuesday, May 14, 2019 - 10:12:46 AM

Links full text

Identifiers

Citation

Sayaka Kamei, Hirostugu Kakugawa, Stéphane Devismes, Sébastien Tixeuil. A self-stabilizing 3-approximation for the maximum leaf spanning tree problem in arbitrary networks. Journal of Combinatorial Optimization, Springer Verlag, 2013, 25 (3), pp.430-459. ⟨10.1007/s10878-011-9383-5⟩. ⟨hal-00930035⟩

Share

Metrics

Record views

714