POOC- a Platform for Object-Oriented Constraint Programming

Abstract

Abstract. In this paper, we describe an implementation-independent object-oriented interface for commercial and academic Constraint Solvers. This serves as a basis for evaluating different Constraint Solvers and for developing solver-independent applications. We show, how applications can use the interface, which solvers are already integrated into the framework and how additional solvers can be added. Furtermore, we provide to the community the described system as real Java packages via Internet, that even includes a basic but powerful Constraint Solver. 1 Motivation Constraint Programming (CP, [9, 8, 5]) has emerged an adequate means to implement and solve many problems known from Artificial Intelligence. Constraint Satisfaction Problems (CSP) can be solved by stating the problem itself instead of defining an algorithm to find a solution. The constraints are posted in a domain-specific solver that infers a solution of the CSP from the constraints and the possible values of constraint variables. Thanks to this declarativity, constraint-based methods are used in many academic applications and in some commercial software systems. Another reason for the success of CP is its efficiency. Many combinatorial NP-hard problems can be solved with constraintbased methods.