Solving the generalized minimum spanning tree problem by a branch-and-bound algorithm

M. Haouari, J. Chaouachi, Moshe Dror

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

We present an exact algorithm for solving the generalized minimum spanning tree problem (GMST). Given an undirected connected graph and a partition of the graph vertices, this problem requires finding a least-cost subgraph spanning at least one vertex out of every subset. In this paper, the GMST is formulated as a minimum spanning tree problem with side constraints and solved exactly by a branch-and-bound algorithm. Lower bounds are derived by relaxing, in a Lagrangian fashion, complicating constraints to yield a modified minimum cost spanning tree problem. An efficient preprocessing algorithm is implemented to reduce the size of the problem. Computational tests on a large set of randomly generated instances with as many as 250 vertices, 1000 edges, and 25 subsets provide evidence that the proposed solution approach is very effective.

Original languageEnglish (US)
Pages (from-to)382-389
Number of pages8
JournalJournal of the Operational Research Society
Volume56
Issue number4
DOIs
StatePublished - Apr 2005

Fingerprint

Costs
Minimum spanning tree
Branch and bound algorithm
Graph
Minimum cost spanning tree problems
Lower bounds

Keywords

  • Branch-and-bound: Lagrangian relaxation
  • Minimum spanning tree

ASJC Scopus subject areas

  • Management of Technology and Innovation
  • Strategy and Management
  • Management Science and Operations Research

Cite this

Solving the generalized minimum spanning tree problem by a branch-and-bound algorithm. / Haouari, M.; Chaouachi, J.; Dror, Moshe.

In: Journal of the Operational Research Society, Vol. 56, No. 4, 04.2005, p. 382-389.

Research output: Contribution to journalArticle

@article{e1def15b91e34a7095deeaa519a20830,
title = "Solving the generalized minimum spanning tree problem by a branch-and-bound algorithm",
abstract = "We present an exact algorithm for solving the generalized minimum spanning tree problem (GMST). Given an undirected connected graph and a partition of the graph vertices, this problem requires finding a least-cost subgraph spanning at least one vertex out of every subset. In this paper, the GMST is formulated as a minimum spanning tree problem with side constraints and solved exactly by a branch-and-bound algorithm. Lower bounds are derived by relaxing, in a Lagrangian fashion, complicating constraints to yield a modified minimum cost spanning tree problem. An efficient preprocessing algorithm is implemented to reduce the size of the problem. Computational tests on a large set of randomly generated instances with as many as 250 vertices, 1000 edges, and 25 subsets provide evidence that the proposed solution approach is very effective.",
keywords = "Branch-and-bound: Lagrangian relaxation, Minimum spanning tree",
author = "M. Haouari and J. Chaouachi and Moshe Dror",
year = "2005",
month = "4",
doi = "10.1057/palgrave.jors.2601821",
language = "English (US)",
volume = "56",
pages = "382--389",
journal = "Journal of the Operational Research Society",
issn = "0160-5682",
publisher = "Palgrave Macmillan Ltd.",
number = "4",

}

TY - JOUR

T1 - Solving the generalized minimum spanning tree problem by a branch-and-bound algorithm

AU - Haouari, M.

AU - Chaouachi, J.

AU - Dror, Moshe

PY - 2005/4

Y1 - 2005/4

N2 - We present an exact algorithm for solving the generalized minimum spanning tree problem (GMST). Given an undirected connected graph and a partition of the graph vertices, this problem requires finding a least-cost subgraph spanning at least one vertex out of every subset. In this paper, the GMST is formulated as a minimum spanning tree problem with side constraints and solved exactly by a branch-and-bound algorithm. Lower bounds are derived by relaxing, in a Lagrangian fashion, complicating constraints to yield a modified minimum cost spanning tree problem. An efficient preprocessing algorithm is implemented to reduce the size of the problem. Computational tests on a large set of randomly generated instances with as many as 250 vertices, 1000 edges, and 25 subsets provide evidence that the proposed solution approach is very effective.

AB - We present an exact algorithm for solving the generalized minimum spanning tree problem (GMST). Given an undirected connected graph and a partition of the graph vertices, this problem requires finding a least-cost subgraph spanning at least one vertex out of every subset. In this paper, the GMST is formulated as a minimum spanning tree problem with side constraints and solved exactly by a branch-and-bound algorithm. Lower bounds are derived by relaxing, in a Lagrangian fashion, complicating constraints to yield a modified minimum cost spanning tree problem. An efficient preprocessing algorithm is implemented to reduce the size of the problem. Computational tests on a large set of randomly generated instances with as many as 250 vertices, 1000 edges, and 25 subsets provide evidence that the proposed solution approach is very effective.

KW - Branch-and-bound: Lagrangian relaxation

KW - Minimum spanning tree

UR - http://www.scopus.com/inward/record.url?scp=15844376710&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=15844376710&partnerID=8YFLogxK

U2 - 10.1057/palgrave.jors.2601821

DO - 10.1057/palgrave.jors.2601821

M3 - Article

VL - 56

SP - 382

EP - 389

JO - Journal of the Operational Research Society

JF - Journal of the Operational Research Society

SN - 0160-5682

IS - 4

ER -