An Abstraction Framework for Mixed Non-deterministic and Probabilistic Systems

Abstract

AbstractWe study abstraction techniques for model checking systems that combine non-deterministic with probabilistic behavior, emphasizing the discrete case. Existing work on abstraction offers a host of isolated techniques which we discuss uniformly through the formulation of abstracted model-checking problems (MCPs). Although this conceptualization is primarily meant to be a useful focal point for surveying the literature on abstraction-based model checking even beyond such combined systems, it also opens up new research opportunities and challenges for abstract model checking of mixed systems. In particular, we sketch how quantitative domain theory may be used to specify the precision of answers obtained from abstract model checks.KeywordsModel CheckMarkov Decision ProcessAbstract InterpretationLabel Transition SystemKripke StructureThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.