SAT-Based Local Improvement for Finding Tree Decompositions of Small Width

Many hard problems can be solved efficiently for problem instances that can be decomposed by tree decompositions of small width. In particular for problems beyond NP, such as #P-complete counting problems, tree decomposition-based methods are particularly attractive. However, finding an optimal tree decomposition is itself an NP-hard problem. Existing methods for finding tree decompositions of small width either (a) yield optimal tree decompositions but are applicable only to small instances or (b) are based on greedy heuristics which often yield tree decompositions that are far from optimal. In this paper, we propose a new method that combines (a) and (b), where a heuristically obtained tree decomposition is improved locally by means of a SAT encoding. We provide an experimental evaluation of our new method.

[1]  Nysret Musliu,et al.  Metaheuristic Algorithms and Tree Decomposition , 2015, Handbook of Computational Intelligence.

[2]  Matti Järvisalo,et al.  SAT-Based Approaches to Treewidth Computation: An Evaluation , 2014, 2014 IEEE 26th International Conference on Tools with Artificial Intelligence.

[3]  Stefan Szeider,et al.  Parameterized Complexity Results for Exact Bayesian Network Structure Learning , 2014, J. Artif. Intell. Res..

[4]  Carsten Sinz,et al.  Towards an Optimal CNF Encoding of Boolean Cardinality Constraints , 2005, CP.

[5]  Adnan Darwiche,et al.  A differential approach to inference in Bayesian networks , 2000, JACM.

[6]  Joachim Gudmundsson,et al.  Turbocharging Treewidth Heuristics , 2016, IPEC.

[7]  Olivier Roussel,et al.  Controlling a Solver Execution with the runsolver Tool , 2011, J. Satisf. Boolean Model. Comput..

[8]  Georg Gottlob,et al.  Bounded treewidth as a key to tractability of knowledge representation and reasoning , 2006, Artif. Intell..

[9]  Rina Dechter,et al.  Tractable Structures for Constraint Satisfaction Problems , 2006, Handbook of Constraint Programming.

[10]  Sebastian Berndt,et al.  Jdrasil: A Modular Library for Computing Tree Decompositions , 2017, SEA.

[11]  Ton Kloks.  Treewidth, Computations and Approximations , 1994, Lecture Notes in Computer Science.

[12]  Markus Chimani,et al.  Improved Steiner Tree Algorithms for Bounded Treewidth , 2011, IWOCA.

[13]  Bruno Courcelle,et al.  On the fixed parameter complexity of graph enumeration problems definable in monadic second-order logic , 2001, Discret. Appl. Math..

[14]  Derek G. Corneil,et al.  Complexity of finding embeddings in a k -tree , 1987 .

[15]  Stefan Woltran,et al.  htd - A Free, Open-Source Framework for (Customized) Tree Decompositions and Beyond , 2017, CPAIOR.

[16]  Eugene C. Freuder.  A sufficient condition for backtrack-bounded search , 1985, JACM.

[17]  Helmut Veith,et al.  Encoding Treewidth into SAT , 2009, SAT.

[18]  Stefan Szeider,et al.  A SAT Approach to Branchwidth , 2016, SAT.

[19]  Vibhav Gogate,et al.  A Complete Anytime Algorithm for Treewidth , 2004, UAI.

[20]  Russell L. Malmberg,et al.  Tree decomposition based fast search of RNA structures including pseudoknots in genomes , 2005, 2005 IEEE Computational Systems Bioinformatics Conference (CSB'05).

[21]  Andrew Gelfand,et al.  Pushing the Power of Stochastic Greedy Ordering Schemes for Inference in Graphical Models , 2011, AAAI.

[22]  Arie M. C. A. Koster,et al.  Treewidth computations I. Upper bounds , 2010, Inf. Comput..

[23]  Arie M. C. A. Koster,et al.  Combinatorial Optimization on Graphs of Bounded Treewidth , 2008, Comput. J..

[24]  David J. Spiegelhalter,et al.  Local computations with probabilities on graphical structures and their application to expert systems , 1990 .