New Search Strategies for the Petri Net CEGAR Approach

Abstract

Abstract. Petri nets are a successful formal method for the modeling and verification of asynchronous, concurrent and distributed systems. Reachability analysis can provide important information about the behavior of the model. However, reachability analysis is a computationally hard problem, especially when the state space is infinite. Abstractionbased techniques are often applied to overcome complexity. In this paper we analyze an algorithm, which uses counterexample guided abstraction refinement. This algorithm proved its efficiency on the model checking contest. We examine the algorithm from a theoretical and practical point of view. On the theoretical side, we show that the algorithm cannot decide reachability for relatively simple instances. We propose a new iteration strategy to explore the invariant space, which extends the set of decidable problems. We also give proofs on the theoretical limits of our approach. On the practical side, we examine different search strategies and we present our new, complex strategy with superior performance compared to traditional strategies. Measurements show that our new contributions perform well for traditional benchmark models as well.

PetriDotNet 1.5: Extensible Petri Net Editor and Analyser for Education and Research

“…The CEGAR algorithm for Petri nets introduced in [26] was the base of our work. Our implementation includes various search strategies, adapted to the characteristics of the different models [12,13]. Stochastic Analysis Algorithms.…”

BME VIK Annual Research Report on Electrical Engineering and Computer Science 2015

“…Resilience of cloud-based cyber-physical systems [113]: risk assessment [112], cloudification, analytics of resource transients, empirical assessment; cost prediction. Model based software and systems engineering [114,115,121,122]: techniques and tools for requirements engineering, design, analysis, optimization, deployment, testing [116], maintenance, domain-specific modeling languages, model transformations [123] and code generation). Formal methods [118,119,124] The Intelligent Systems Group is competent and dedicated to research in ambient intelligent systems [127], heterogeneous information processing, data analysis, medical image analysis [126], modeling complex systems [128], bioinformatics, biomedical informatics [125].…”

Industrial applications of the PetriDotNet modelling and analysis tool

“…The CEGAR algorithm for Petri nets was first introduced in [40] and it served as the basis of our work [41,42]. This section presents the main concept of the original algorithm and our contributions as well.…”

“…We also investigated the completeness of the algorithm and presented some subclasses of nets where the algorithm could not decide the problem (yielding inconclusive answer or non-termination) due to its iteration strategy [41,42]. We introduced the concept of distant invariants [42] and some search space pruning techniques [41] to overcome this deficiency for some of the undecidable subclasses. We also investigated the theoretical limits of our new approaches [42].…”

“…We introduced the concept of distant invariants [42] and some search space pruning techniques [41] to overcome this deficiency for some of the undecidable subclasses. We also investigated the theoretical limits of our new approaches [42].…”