<< Prev | - Up - |
Concurrent Constraint Programming allows a very intuitive implementation of a solver for dominance constraints.
Every variable in the dominance constraints corresponds to a node in the constraints graph. Every node of the constraint graph is associated with four sets of nodes: the sets of graph nodes that denote tree nodes equal, strictly above, strictly below, and to the side of it.
Finite set constraints can be used to axiomatize the problem; they can be taken over in Mozart with only syntactic variations.
The dominance constraint solver based on finite set constraints has been integrated into the CHORUS demo system and runs efficiently on dominance constraints from underspecified semantics.
<< Prev | - Up - |