Tabu Search for Generalized Hypertree Decompositions

Many important real world problems can be formulated as Constraint Satisfaction Problems (CSPs). A CSP consists of a set of variables each with a domain of possible values, and a set of constraints on the allowed values for specified subsets of variables. A solution to CSP is the assignment of values to variables, such that no constraint is violated. CSPs include many NP-complete problems and are in general hard to solve. However, some classes of CSPs can be solved efficiently if they have bounded treewidth (tw) or generalized hypertree width (ghw). The process of solving problems with bounded tw/ghw includes two phases. In the first phase the tree or generalized hypertree decomposition with small upper bound for tw/htw width is generated. Second phase includes solving a problem (based on the decomposition) with a particular algorithm which runs in polynomial time on the width and the size of the problem. Given that the classes of CSPs of bounded tw/ghw are solvable in polynomial time, these two concepts are very important for efficient solving of intractable problems. In this paper we consider the generation of generalized hypertree decompositions of small width.