Treewidth in Verification: Local vs. Global
The treewidth of a graph measures how close the graph is to a tree. Many problems that are intractable for general graphs, are tractable when the graph has bounded treewidth. Recent works study the complexity of model checking for state transition systems of bounded treewidth. There is little reason to believe, however, that the treewidth of the state transition graphs of real systems, which we refer to as global treewidth, is bounded. In contrast, we consider in this paper concurrent transition systems, where communication between concurrent components is modeled explicitly. Assuming boundedness of the treewidth of the communication graph, which we refer to as local treewidth, is reasonable, since the topology of communication in concurrent systems is often constrained physically.
In this work we study the impact of local treewidth boundedness on the complexity of verification problems. We first present a positive result, proving that a CNF formula of bounded treewidth can be represented by an OBDD of polynomial size. We show, however, that the nice properties of treewidth-bounded CNF formulas are not preserved under existential quantification or unrolling. Finally, we show that the complexity of various verification problems is high even under the assumption of local treewidth boundedness. In summary, while global treewidth boundedness does have computational advantages, it is not a realistic assumption; in contrast, local treewidth boundedness is a realistic assumption, but its computational advantages are rather meager.
KeywordsModel Check Transition System Transition Relation Tree Decomposition Communication Graph
Unable to display preview. Download preview PDF.
- 2.Beer, I., Ben-David, S., Geist, D., Gewirtzman, R., Yoeli, M.: Methodology and system for practical formal verification of reactive hardware. In: Proc. 6th Conf. on Computer Aided Verification, Stanford, June 1994, pp. 182–193 (1994)Google Scholar
- 3.Biere, A., Cimatti, A., Clarke, E.M., Fujita, M., Zhu, Y.: Symbolic model checking using SAT procedures instead of BDDs. In: DAC 1999, pp. 317–320 (1999)Google Scholar
- 7.Bodlaender, H.L.: A partial k-arboretum of graphs with bounded treewidth. Technical report, Universiteit Utrecht (1998)Google Scholar
- 8.Bryant, R.E.: Graph-based algorithms for boolean-function manipulation. IEEE Trans. on Computers C-35(8) (1986)Google Scholar
- 10.Clarke, E.M., Grumberg, O., Peled, D.A.: Model Checking. MIT Press, Cambridge (2000)Google Scholar
- 12.Diestel, R.: Graph Theory. Graduate Texts in Mathematics, vol. 173. Springer, Heidelberg (2000)Google Scholar
- 14.Freuder, E.C.: Complexity of k-tree structured constraint satisfaction problems. In: Proc. AAAI 1990, pp. 4–9 (1990)Google Scholar
- 21.Milner, R.: An algebraic definition of simulation between programs. In: Proc. 2nd Int. Joint Conf. on Artif. Int., pp. 481–489. British Computer Society (September 1971)Google Scholar
- 27.Schiex, T.: A note on CSP graph parameters. Technical Report 1999/03, INRIA (1999)Google Scholar
- 29.Wang, D., Clarke, E.M., Zhu, Y., Kukula, J.: Using cutwidth to improve symbolic simulation and boolean satisfiability. In: IEEE International High Level Design Validation and Test Workshop (HLDVT 2001), p. 6 (2001)Google Scholar