Triangle-based consistencies for cost function networks
 - LIRMM - Laboratoire d’Informatique, de Robotique et de Microélectronique de Montpellier
Article Dans Une Revue Constraints Année : 2017

Triangle-based consistencies for cost function networks


Résumé

Cost Function Networks (aka Weighted CSP) allow to model a variety of problems, such as optimization of deterministic and stochastic graphical models including Markov random Fields and Bayesian Networks. Solving cost function networks is thus an important problem for deterministic and probabilistic reasoning. This paper focuses on local consistencies which define essential tools to simplify Cost Function Networks, and provide lower bounds on their optimal solution cost. To strengthen arc consistency bounds, we follow the idea of triangle-based domain consistencies for hard constraint networks (path inverse consistency, restricted or max-restricted path consistencies), describe their systematic extension to cost function networks, study their relative strengths, define enforcing algorithms, and experiment with them on a large set of benchmark problems. On some of these problems, our improved lower bounds seem necessary to solve them.
Fichier principal
Vignette du fichier
Hiep2016.pdf (903.12 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

lirmm-01374514 , version 1 (30-09-2016)

Identifiants

Citer

Christian Bessiere, Simon de Givry, Thomas Schiex, Thi Hông Hiêp Nguyên. Triangle-based consistencies for cost function networks
. Constraints, 2017, 22 (2), pp.230-264. ⟨10.1007/s10601-016-9250-1⟩. ⟨lirmm-01374514⟩
168 Consultations
244 Téléchargements

Altmetric

Partager

More