Home > Research > Publications & Outputs > A hybrid breakout local search and reinforcemen...

Electronic data

  • 1-s2.0-S0377221717300589-main

    Rights statement: This is the author’s version of a work that was accepted for publication in European Journal of Operational Research. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in European Journal of Operational Research, 261, 3, 2017 DOI: 10.1016/j.ejor.2017.01.023

    Accepted author manuscript, 1.81 MB, PDF document

    Available under license: CC BY-NC-ND: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

Links

Text available via DOI:

View graph of relations

A hybrid breakout local search and reinforcement learning approach to the vertex separator problem

Research output: Contribution to journalJournal article

Published

Standard

A hybrid breakout local search and reinforcement learning approach to the vertex separator problem. / Benlic, Una; Epitropakis, Michael G.; Burke, Edmund K.

In: European Journal of Operational Research, Vol. 261, No. 3, 16.09.2017, p. 803-818.

Research output: Contribution to journalJournal article

Harvard

APA

Vancouver

Author

Benlic, Una ; Epitropakis, Michael G. ; Burke, Edmund K. / A hybrid breakout local search and reinforcement learning approach to the vertex separator problem. In: European Journal of Operational Research. 2017 ; Vol. 261, No. 3. pp. 803-818.

Bibtex

@article{76547a1402d7485d8cd805711527c7c0,
title = "A hybrid breakout local search and reinforcement learning approach to the vertex separator problem",
abstract = "The Vertex Separator Problem (VSP) is an NP-hard problem which arises from several important domains and applications. In this paper, we present an improved Breakout Local Search for VSP (named BLS-RLE). The distinguishing feature of BLS-RLE is a new parameter control mechanism that draws upon ideas from reinforcement learning theory for an interdependent decision on the number and on the type of perturbation moves. The mechanism complies with the principle “intensification first, minimal diversification only if needed”, and uses a dedicated sampling strategy for a rapid convergence towards a limited set of parameter values that appear to be the most convenient for the given state of search. Extensive experimental evaluations and statistical comparisons on a wide range of benchmark instances show significant improvement in performance of the proposed algorithm over the existing BLS algorithm for VSP. Indeed, out of the 422 tested instances, BLS-RLE was able to attain the best-known solution in 93.8{\%} of the cases, which is around 20{\%} higher compared to the existing BLS. In addition, we provide detailed analyses to evaluate the importance of the key elements of the proposed method and to justify the degree of diversification introduced during perturbation.",
keywords = "Heuristics, Iterated local search, Vertex separator, Parameter control, Reinforcement learning",
author = "Una Benlic and Epitropakis, {Michael G.} and Burke, {Edmund K.}",
note = "This is the author’s version of a work that was accepted for publication in European Journal of Operational Research. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in European Journal of Operational Research, 261, 3, 2017 DOI: 10.1016/j.ejor.2017.01.023",
year = "2017",
month = "9",
day = "16",
doi = "10.1016/j.ejor.2017.01.023",
language = "English",
volume = "261",
pages = "803--818",
journal = "European Journal of Operational Research",
issn = "0377-2217",
publisher = "Elsevier Science B.V.",
number = "3",

}

RIS

TY - JOUR

T1 - A hybrid breakout local search and reinforcement learning approach to the vertex separator problem

AU - Benlic, Una

AU - Epitropakis, Michael G.

AU - Burke, Edmund K.

N1 - This is the author’s version of a work that was accepted for publication in European Journal of Operational Research. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in European Journal of Operational Research, 261, 3, 2017 DOI: 10.1016/j.ejor.2017.01.023

PY - 2017/9/16

Y1 - 2017/9/16

N2 - The Vertex Separator Problem (VSP) is an NP-hard problem which arises from several important domains and applications. In this paper, we present an improved Breakout Local Search for VSP (named BLS-RLE). The distinguishing feature of BLS-RLE is a new parameter control mechanism that draws upon ideas from reinforcement learning theory for an interdependent decision on the number and on the type of perturbation moves. The mechanism complies with the principle “intensification first, minimal diversification only if needed”, and uses a dedicated sampling strategy for a rapid convergence towards a limited set of parameter values that appear to be the most convenient for the given state of search. Extensive experimental evaluations and statistical comparisons on a wide range of benchmark instances show significant improvement in performance of the proposed algorithm over the existing BLS algorithm for VSP. Indeed, out of the 422 tested instances, BLS-RLE was able to attain the best-known solution in 93.8% of the cases, which is around 20% higher compared to the existing BLS. In addition, we provide detailed analyses to evaluate the importance of the key elements of the proposed method and to justify the degree of diversification introduced during perturbation.

AB - The Vertex Separator Problem (VSP) is an NP-hard problem which arises from several important domains and applications. In this paper, we present an improved Breakout Local Search for VSP (named BLS-RLE). The distinguishing feature of BLS-RLE is a new parameter control mechanism that draws upon ideas from reinforcement learning theory for an interdependent decision on the number and on the type of perturbation moves. The mechanism complies with the principle “intensification first, minimal diversification only if needed”, and uses a dedicated sampling strategy for a rapid convergence towards a limited set of parameter values that appear to be the most convenient for the given state of search. Extensive experimental evaluations and statistical comparisons on a wide range of benchmark instances show significant improvement in performance of the proposed algorithm over the existing BLS algorithm for VSP. Indeed, out of the 422 tested instances, BLS-RLE was able to attain the best-known solution in 93.8% of the cases, which is around 20% higher compared to the existing BLS. In addition, we provide detailed analyses to evaluate the importance of the key elements of the proposed method and to justify the degree of diversification introduced during perturbation.

KW - Heuristics

KW - Iterated local search

KW - Vertex separator

KW - Parameter control

KW - Reinforcement learning

U2 - 10.1016/j.ejor.2017.01.023

DO - 10.1016/j.ejor.2017.01.023

M3 - Journal article

VL - 261

SP - 803

EP - 818

JO - European Journal of Operational Research

JF - European Journal of Operational Research

SN - 0377-2217

IS - 3

ER -