The research domain of ARC OCOQS is the optimal control of stochastic processes, with applications to the control of networks and manufacturing systems. The principal aim is to widen the set of mathematical techniques that can be used to prove that optimal policies are of threshold type, thereby widening the set of classes of models that can be effectively solved exactly or numerically handled in practice. To that end, the collaborative action will bring together specialists: of discrete-event systems and stochastic orders on the one hand, and of queueing, networking and manufacturing applications on the other hand. By analyzing several concrete problems which cannot presently be handled by known methods, it is hoped that new proof techniques, presumably based on stochastic ordering concepts, will emerge.

Detailed description...

- Ana Busic, coordinator (TREC, INRIA Paris-Rocquencourt)
- Emmanuel Hyon (University of Paris Ouest Nanterre La Defense and LIP6)
- Alain Jean-Marie (MAESTRO, INRIA Sophia-Antipolis)
- Ingrid Vliegen (University of Twente)