Skip to main content
françaisFR
European Commission logo print header
Contenu archivé le 2024-04-19

Causal Calculi Based on Nets

Objectif

CALIBAN contributes towards the combination and the integration of causality-based models of concurrency (such as Petri nets) and calculus-oriented models of concurrency (such as process algebras). This involves:

- On the foundational level: (1) investigations towards a unified behavioural semantics of causality, taking into account all major approaches (especially Petri net based ones), strengthening existing and deriving new results based on the connections between them; and (2) the investigation of the connections between the structure and the behaviour of a concurrent system.
- On the practical level: (3) the development of the Box algebra, a causality-based algebra (defined in Action 3148, DEMON) into a practical method supporting the design of concurrent systems, emphasising the benefits of an explicit representation of causality; and (4) the integration of high-level nets into a calculus-oriented systematic specification method. The integration of logical specification techniques and model checking. Further, the demonstration of the advantages of this method by its application to substantial case studies.
The problem of synthesizing Petri nets from their behavioural description has been solved from various points of view. Their construction from logical specifications, and the logical formulation of truly concurrent behaviour, has been studied. The theory of finite and infinite traces has been investigated deeply. Basic properties of causality based models have been studied. Work on truly concurrent equivalences has been continued. Milner's pi-calculus has been given a Petri net rendering.

Structure theory has been consolidated and the first draft of the book has been completed. Generalizations of structure theory have been developed. A survey on performance aspects has been produced. A very fast model checking algorithm based on partial orders has been developed and proved correct, while other algorithms known from the literature have been corrected.

Denotational Petri net semantics for refinement and recursion have been found. General covering properties have been proved as a results from the semantics. The behavioural semantics of the Box algebra have been defined in 3 ways: by processes, by branching processes, and by structured operational semantics. A programming notation has been defined whose compositional semantics are in terms of the Box algebra and hence also by nets.

Work on the algebraization of high level nets, and on the modularization of algebraic high level nets, has been continued. Links with object orientation have been firmly established. The Box algebra has been given finite semantics in these terms. Two major case studies (replicated processing and distributed sorting) have been reported.
CALIBAN disseminates its results through established conferences and journal papers. It also plans to produce two monographs. The partners hold a general meeting once a year. Workshops involving at least two partners are being held. Amongst others, a workshop on the High Level Box Algebra has been held in May 1993, and a workshop on Causality Semantics and the Box Algebra is planned for November 1993. The partners are also in the process of realising a programme of visits between researchers at cooperating sites. CALIBAN cooperates with several national and international projects on concurrency. The group will organise at least one open workshop. All activities (including a list of publications) are documented in the CALIBAN Newsletter that is issued in irregular intervals by its coordinator. Copies of the newsletter can be obtained upon request.

POTENTIAL

CALIBAN's results provide a better knowledge of causality-based models, a collection of net analysis algorithms including model checking, a powerful process algebra with compositional net semantics, and a sound high-level net model for the specification of net systems. These results may eventually be integrated into a commercial tool which will allow parallel programs to be written in a specially designed language, obtain and exploit their net semantics, and verify some of their properties using a range of particularly efficient algorithms. A prototype of this kind is being implemented in a concomitant national project in Germany.

Champ scientifique

Programme(s)

Thème(s)

Data not available

Appel à propositions

Data not available

Régime de financement

Data not available

Coordinateur

Universität Hildesheim
Contribution de l’UE
Aucune donnée
Adresse
Marienburger Platz 22
31141 Hildesheim
Allemagne

Voir sur la carte
Coût total
Aucune donnée

Participants (11)

Partager cette page

Télécharger

Dernière mise à jour: 17 Juin 1994

Mon livret

Permalink: https://cordis.europa.eu/project/id/6067/fr

European Union, 2024