Concurrently decomposable constraint systems

Authors:
Cees Witteveen;Wiebe Van Der Hoek;Nico Roos
Affiliations:
Delft University of Technology, Dept of Software Technology, Delft, The Netherlands;University of Liverpool, Dept of Computer Science, Liverpool, United Kingdom;Maastricht University, Maastricht ICT Competence Center, Maastricht, The Netherlands
Venue:
MATES'09 Proceedings of the 7th German conference on Multiagent system technologies
Year:
2009

Citing 9
Cited 1

Tree clustering for constraint networks (research note)

Artificial Intelligence
On the Desirability of Acyclic Database Schemes

Journal of the ACM (JACM)
Algorithms for a temporal decoupling problem in multi-agent planning

Eighteenth national conference on Artificial intelligence
Constraint Processing

Constraint Processing
A domain decomposition algorithm for constraint satisfaction

Journal of Experimental Algorithmics (JEA)
Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations

Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations
A comparison of structural CSP decomposition methods

IJCAI'99 Proceedings of the 16th international joint conference on Artifical intelligence - Volume 1
Constraint partitioning for solving planning problems with trajectory constraints and goal preferences

IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
Constraint partitioning in penalty formulations for solving temporal planning problems

Artificial Intelligence

Optimal temporal decoupling in multiagent systems

Proceedings of the 9th International Conference on Autonomous Agents and Multiagent Systems: volume 1 - Volume 1

Quantified Score

Hi-index	0.00

Visualization

Abstract

In constraint satisfaction, decomposition is a common technique to split a problem in a number of parts in such a way that the global solution can be efficiently assembled from the solutions of the parts. In this paper, we study the decomposition problem from an autonomous agent perspective. Here, a constraint problem has to be solved by different agents each controlling a disjoint set of variables. Such a problem is called concurrently decomposable if each agent is (i) capable to solve its own part of the problem independently of the others, and (ii) the individual solutions always can be merged to a complete solution of the total problem. First of all, we investigate how difficult it is to decide whether or not a given constraint system and agent partitioning allows for such a concurrent decomposition. Secondly, we investigate how difficult it is to find suitable reformulations of the original constraint problem that allow for concurrent decomposition.