Cutting planes in constraint logic programming

In this paper, we show how recently developed techniques from combinatorial optimization can be embedded into constraint logic programming. We develop a constraint solver for the constraint logic programming language CLP($\cal PB$) for logic programming with pseudo-Boolean constraints. Our approach is based on the generation of polyhedral cutting planes and the concept of branch-and-cut. In the case of 0-1 constraints, this can improve or replace the finite domain techniques used in existing constraint logic programming systems