Issue: 2017/Vol.27/No.4, Pages 27-43
CONTROL DESIGN FOR UNTIMED PETRI NETS USING MARKOV DECISION PROCESSES
Cite as: C. Daoui, D. Lefebvre. Control design for untimed Petri nets using Markov Decision Processes. Operations Research and Decisions 2017: 27(4), 27-43. DOI 10.5277/ord170402
Design of control sequences for discrete event systems (DESs) has been presented modelled by untimed Petri nets (PNs). PNs are well-known mathematical and graphical models that are widely used to describe distributed DESs, including choices, synchronizations and parallelisms. The domains of application include, but are not restricted to, manufacturing systems, computer science and transportation networks. We are motivated by the observation that such systems need to plan their production or services. The paper is more particularly concerned with control issues in uncertain environments when unexpected events occur or when control errors disturb the behaviour of the system. To deal with such uncertainties, a new approach based on discrete time Markov decision processes (MDPs) has been proposed that associates the modelling power of PNs with the planning power of MDPs. Finally, the simulation results illustrate the benefit of our method from the computational point of view.
Keywords: discrete event systems, Petri nets, control design, Markov decision process, value iteration algorithm
Received: 13 June 2017 Accepted: 9 January 2018