Title: NB-DEVS: a hybrid approach for modelling and simulation of imperfect systems

Authors: Kadda Mostefaoui; Youcef Dahmani

Addresses: Ecole nationale Supérieure d'Informatique, BP 68M, 16309, Oued-Smar, Alger, Algeria ' Department of Computer Science, University of Ibn Khaldoun, Tiaret, Algeria

Abstract: This article presents the NB-DEVS approach for the modelling and simulation of discrete event systems (DES), our objectives are to propose a new formal approach for describing the complex systems involving imperfectly. In our research we propose to integrate the naïve Bayesian network (NB) into DEVS formalism. The association of these two formalisms we have allow to extend the classic DEVS formalism to new domains. This new approach permits to study the system when its behaviour is uncertain. Finally, we simulated the evolution of the production system using our NB-DEVS approach. In simulation, we used several different models to replicate the empirical data. From these simulation iterations, we were able to verify and validate our model.

Keywords: modelling; simulation; discrete event systems; DEVS; Bayesian networks; uncertain.

DOI: 10.1504/IJSPM.2019.103588

International Journal of Simulation and Process Modelling, 2019 Vol.14 No.4, pp.377 - 388

Received: 03 Aug 2018
Accepted: 07 Mar 2019
Published online: 12 Nov 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article