Forward bounding on pseudo-trees for DCOPs and ADCOPs

Omer Litov, Amnon Meisels

Research output: Contribution to journalArticlepeer-review

29 Scopus citations

Abstract

Complete search algorithms for solving Distributed constraint optimization problems (DCOPs) can be divided into two groups: algorithms that use a pseudo tree and algorithms that do not use one. The best performing algorithms that do not use a pseudo tree use some form of forward bounding. In order to try and gain from both worlds, a new algorithm, which incorporates a hybrid approach of the two groups is presented. The proposed algorithm – Pseudo-Tree Forward-bounding (PT-FB) – is shown to perform very well. PT-FB is next extended to be able to solve Asymmetric DCOPs (ADCOPs). Here again, its performance is better than its competitors. An extensive experimental evaluation of the performance of each of the proposed algorithms is presented.

Original languageEnglish
Pages (from-to)83-99
Number of pages17
JournalArtificial Intelligence
Volume252
DOIs
StatePublished - Nov 2017
Externally publishedYes

Bibliographical note

Publisher Copyright:
© 2017 Elsevier B.V.

Funding

This work was supported by the Lynn and William Frankel center for Computer Sciences .

FundersFunder number
Research Center of the College of Computer and Information Sciences, King Saud University

    Keywords

    • Asymmetric DCOPs
    • Distributed constraints optimization

    Fingerprint

    Dive into the research topics of 'Forward bounding on pseudo-trees for DCOPs and ADCOPs'. Together they form a unique fingerprint.

    Cite this