1,721,081 research outputs found
Event Structures with Disabling/Enabling relation and Event Automata
No full textIn recent years the consideration that events in evolutions of concurrent systems can happen with different histories has received ground. In particular the possibility that part of the history can be abstracted away or identified, like in the collective tokens philosophy for Petri Nets, has gained the stage. The various brands of event structures considered in literature are tailored to a fixed interpretation with respect to the history of an event. We investigate the adequateness of event structures with a disabling/enabling relation, to settle a common ground for the history dependent and history independent interpretations, and we establish a relationship between event automata and these notions of event structures
Petri Nets Unfoldings and the individual/collective token philosophy
No full textThe notion of unfolding plays a major role in the so called non sequential semantics of Petri nets. In literature various approaches to this notion have been proposed, taking into account the so called individual token philosophy (the whole history is relevant) and the so called collective token philosophy (part of the history can be forgotten). In this paper we compare and relate two notions of unfolding (one for the individual and the other for the collective token philosophy) and we investigate on the relations between these two approaches and the non sequential behaviour of Petri Nets
Process discovery and Petri nets
No full textThe aim of the research domain known as process mining is to use process discovery to construct a process model as an abstract representation of event logs. The goal is to build a model (in terms of a Petri net) that can reproduce the logs under consideration, and does not allow different behaviours compared with those shown in the logs. In particular, process mining aims to verify the accuracy of the model design (represented as a Petri net), basically checking whether the same net can be rediscovered. However, the main mining methods proposed in the literature have some drawbacks: the classical α-algorithm is unable to rediscover various nets, while the region-based approach, which can mine them correctly, is too complex.
In this paper, we compare different approaches and propose some ideas to counter the weaknesses of the region-based approach
Catalytic and communicating Petri nets are Turing complete
Full text linkIn most studies about the expressiveness of Petri nets, the focus has been put either on adding suitable arcs or on assuring that a complete snapshot of the system can be obtained. While the former still complies with the intuition on Petri nets, the second is somehow an orthogonal approach, as Petri nets are distributed in nature. Here, inspired by membrane computing, we study some classes of Petri nets where the distribution is partially kept and which are still Turing complete
Component-based Verification in a Synchronous Setting
No full textFormal verification of properties in reactive real-time systems is crucial, as these systems are often safety-critical. Such systems are successfully implemented using synchronous languages, where refinement is a relevant operation. This paper investigates the interplay between this operation and formal verification. It turns out that, while for the refined program component-based verification of properties expressed using suitable temporal logics is easily achieved, component-based verification from the point of view of the refining program is best achieved with observers. Our results are based on a translation of synchronous programs into Boolean automata. Their practical relevance is illustrated with a protocol case study
- …
