Revisiting state-based models for synthesizing optimal communicating decentralized discrete-event controllers

The foremost motivation for introducing communication into the decentralized discrete-event control problem is to allow controllers to distinguish between behaviors of a system that remain within a pre-defined set of so-called legal behaviors and those that lead to behaviors outside this set. After communication, updates to a controller's partial view of the system should reflect that such behaviors are now distinguishable whenever either the legal or illegal behavior occurs. In state-based models, updates occur only at information states that occur immediately after a communication, resulting in inconsistent information states in the rest of the model. A previously-defined state-based structure is expanded to include the notion of system-wide information state updates in response to expected communications. A discussion of the identification of optimal control policies in the new model is presented.