Using Stochastic Solvers in Constraint Logic Programming

This paper proposes a general framework for integrating a constraint logic programming system with a stochastic constraint solver to solve constraint satisfaction problems eeciently. Stochastic solvers can solve hard constraint satisfaction problems very eeciently, and constraint logic programming allows heuristics and problem breakdown to be encoded in the same language as the constraints. Hence their combination is attractive. Unfortunately there is a mismatch in the kinds of information a stochastic solver provides, and that which a constraint logic programming system requires. We study the semantic properties of constraint logic programming systems that make use of stochastic solvers, and give soundness and completeness results for their use. We describe an example system. We have implemented a modiied neural network simulator, GENET, as a constraint solver. We study various strategies for making use of this constraint solver, and compare the eeciency of the system on some sample problems against the propagation based solver approaches typically used in constraint logic programming.