Constraint programming often involves global constraints, for which various custom filtering algorithms have been published. This work presents a semi-automatic generation of CHR solvers for the subset of global constraints defineable by specific automata. The generation is based on a constraint logic program modelling an automaton and an improved version of the Prim-Miner algorithm. The solvers only need to be generated once and achieve arc-consistency for over 40 global constraints.
展开▼
