Generalized arc consistency for global cardinality constraint

  • Authors:
  • Jean-Charles Régin

  • Affiliations:
  • ILOG S.A., Gentilly Cedex, France

  • Venue:
  • AAAI'96 Proceedings of the thirteenth national conference on Artificial intelligence - Volume 1
  • Year:
  • 1996

Quantified Score

Hi-index 0.00

Visualization

Abstract

A global cardinality constraint (gcc) is specified in terms of a set of variables X = {x1,..., xp} which take their values in a subset of V = {v1,...,vd}. It constrains the number of times a value vi ∈ V is assigned to a variable in X to be in an interval [li, ci. Cardinality constraints have proved very useful in many real-life problems, such as scheduling, timetabling, or resource allocation. A gcc is more general than a constraint of difference, which requires each interval to be [0,1]. In this paper, we present an efficient way of implementing generalized arc consistency for a gcc. The algorithm we propose is based on a new theorem of flow theory. Its space complexity is O(|X| × |V|) and its time complexity is O(|X|2 × |V|). We also show how this algorithm can efficiently be combined with other filtering techniques.