Skip to Main content Skip to Navigation
Theses

Hybridation de méthodes complètes et incomplètes pour la résolution de CSP

Abstract : Hybridization of incomplete and constraint programming techniques for solving Constraint Satisfaction Problems is generally restricted to some kind of master-slave combinations for specific classes of problems. In this PhD thesis, we are concerned with the design of a hybrid resolution framework based on K.R. Apt's chaotic iterations. In this framework, basic resolution processes
are abstracted by functions over an ordered structure. This allows us to consider the different resolution agents at a same level and to study more precisely various strategies for hybridization of local search, genetic algorithms and constraint propagation.
Hybrid resolution can be achieved as the computation of a fixed point of some specific reduction functions. Our framework opens up new and finer possibilities for hybridization/combination strategies. Our prototype implementation gave experimental results showing the interest of the model to design such hybridizations.
Document type :
Theses
Complete list of metadata

Cited literature [103 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00130790
Contributor : Tony Lambert <>
Submitted on : Tuesday, February 13, 2007 - 8:14:29 PM
Last modification on : Monday, June 1, 2020 - 4:54:02 PM
Long-term archiving on: : Friday, September 21, 2012 - 11:25:39 AM

Identifiers

  • HAL Id : tel-00130790, version 1

Citation

Tony Lambert. Hybridation de méthodes complètes et incomplètes pour la résolution de CSP. Modélisation et simulation. Université de Nantes, 2006. Français. ⟨tel-00130790⟩

Share

Metrics

Record views

665

Files downloads

1304