Constraint Satisfaction by Means of Dynamic.pdfVIP

Constraint Satisfaction by Means of Dynamic Polyhedra Miguel A. Salido, Adriana Giret, Federico Barber Universidad Politécnica de Valencia, 46071, Valencia, Spain Abstract. Nowadays, many real problems in the Artificial Intelligence environments, can be efficiently modelled as Constraint Satisfaction Problems (CSP’s) and can be solved by means of Operational Research techniques. It is well known that any non-binary CSP can be transformed into an equivalent binary one, using some of the current techniques. However, this transformation may not be practical in problems with certain properties. Therefore, it is necessary to manage these non-binary constraints directly. In this paper, we propose an algorithm called HSA≠ that solves non-binary constraint satisfaction problems in a natural way as an incremental and non-binary CSP solver. This non-binary CSP solver carries out the search through a polyhedron that maintains in its vertices those solutions that satisfy all non-binary constraints. 1 Introduction Nowadays, many researchers are working on combined methods of Artificial Intelligence and Operational Research (AI/OR) because many real problems can be modelled as constraint satisfaction problems (CSPs) and can be solved using linear programming techniques. These include problems from fields such as artificial intelligence, operational research, expert systems, databases, etc. Most of these problems can be naturally modelled using non-binary (or n-ary) constraints. In the constraint satisfaction literature, the need to address issues regarding non- binary constraints has only recently started to be widely recognised. Researchers have traditionally focused on binary constraints. The basic reasons were the simplicity of dealing with binary constraints compared to n