Gurfinkel A thesis submitted in conformity with the requirements for the degree of Master of Science

Multi-Valued Symbolic Model-Checking: Fairness, Counter-Examples, Running Time Arie Gurfinkel Master of Science Graduate Department of Computer Science University of Toronto 2003 Multi-valued model-checking is an effective technique for reasoning about systems with incomplete or inconsistent information. In particular, it is well suited for reasoning about abstract, partial, and feature-based system descriptions. The technique is based on extending the classical model-checking algorithm over two-valued logic to arbitrary finite logics whose truth values form a distributive De Morgan lattice. In this thesis we address several issues surrounding the usability of multi-valued modelchecking. Firstly, we provide an improved analysis of the worst-case complexity of the symbolic multi-valued model-checking algorithm, and show that it is independent of the height of the lattice. Secondly, we extend the notion of fairness to a multi-valued models, thus enabling application of multi-valued model-checking to asynchronous concurrent systems. Thirdly, we introduce multi-valued witnesses and counter-examples that aid in interpreting the results of the model-checker. Finally, we describe the design and implementation of a multi-valued modelchecker χChek.