- Journal
- INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL
- Année
- 2006
- Volume
- 2
- Numéro
- 4
- Pages
- 771-788
- Mois
- AUG
Abstract
Discrete event systems (DES) involving only synchronization phenomena can be represented by linear models in some particular algebraic structures called dioids. The (max, +) algebra is a dioid over which, these systems can be modelled by linear mathematical equations with special sum and product operations. The study of the DES over this algebra is usually made through the corresponding timed event graph (TEG). In this paper, we extend the study to DES that are subject to time-varying constraints. We enrich the existing TEG formalism to characterize these systems and propose a modelling and a just-in-time control policy. The proposed approach is illustrated with a supply chain system subject to non constant logistic demand. Then, we present a simulation-based optimization technique, which implements a genetic algorithm, to evaluate the proposed just-in-time control and the model performances while minimizing the supply chain costs.