Constraint propagation is a generic method for combining monotone operators over lattices. In this paper, a family of constraint propagation strategies are designed as instances of a single algorithm shown to be correct, finite, and strategy-independent. The main idea is to separate complex reasoning processes (that exploit some knowledge of constraints and solvers in order to accelerate the convergence) from basic ffxed-point algorithms. Some sequential interval narrowing strategies are expressed in this framework and parallel implementations in different execution models are discussed.
CITATION STYLE
Granvilliers, L., & Monfroy, E. (2000). Declarative modelling of constraint propagation strategies. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1909, pp. 201–215). Springer Verlag. https://doi.org/10.1007/3-540-40888-6_19
Mendeley helps you to discover research relevant for your work.