thesis Nils Timm author Heike Wehrheim supervisor 77 department SFB 901 project SFB 901 - Subprojekt B4 project SFB 901 - Project Area B project Software systems are playing an increasing role in our everyday life, and as the amount of software applications grows, so does their complexity and the relevance of their computations. Software components can be found in many systems that are charged with safety-critical tasks, such as control systems for aviation or power plants. Hence, software verification techniques that are capable of proving the absence of critical errors are becoming more and more important in the field software engineering. A well-established approach to software verification is model checking. Applying this technique involves an exhaustive exploration of a state space model corresponding to the system under consideration. The major challenge in model checking is the so-called state explosion problem: The state space of a software system grows exponentially with its size. Thus, the straightforward modelling of real-life systems practically impossible. A common approach to this problem is the application of abstraction techniques, which reduce the original state space by mapping it on a significantly smaller abstract one. Abstraction inherently involves a loss of information, and thus, the resulting abstract model may be too imprecise for a definite result in verification. Therefore, abstraction is typically combined with abstraction refinement: An initially very coarse abstract model is iteratively refined, i.e. enriched with new details about the original system, until a level of abstraction is reached that is precise enough for a definite outcome. Abstraction refinement-based model checking is fully automatable and it is considered as one of the most promising approaches to the state explosion problem in verification. However, it is still faced with a number of challenges. There exist several types of abstraction techniques and not every type is equally well-suited for all kinds of systems and verification tasks. Moreover, the selection of adequate refinement steps is nontrivial and typically the most crucial part of the overall approach: Unfavourable refinement decisions can compromise the state space-reducing effect of abstraction, and as a consequence, can easily lead to the failure of verification. It is, however, hard to predict which refinement steps will eventually be expedient for verification – and which not. https://ris.uni-paderborn.de/download/478/1324/478-Dissertation-Timm.pdf application/pdf Universität Paderborn2013 Timm, N. (2013). <i>Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems</i>. Universität Paderborn. @book{Timm_2013, title={Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems}, publisher={Universität Paderborn}, author={Timm, Nils}, year={2013} } N. Timm, Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems, Universität Paderborn, 2013. Timm, Nils. <i>Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems</i>. Universität Paderborn, 2013. N. Timm, <i>Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems</i>. Universität Paderborn, 2013. Timm, Nils. <i>Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems</i>. Universität Paderborn, 2013. Timm N. <i>Three-Valued Abstraction and Heuristic-Guided Refinement for Verifying Concurrent Systems</i>. Universität Paderborn; 2013. 4782017-10-17T12:42:25Z2022-01-06T07:01:22Z