Distributed knowledge for communication in decentralized discrete-event systems

An extension to a formal model for reasoning about knowledge and communication in discrete-event systems is presented. The model is based on a modal logic where propositions describe the knowledge that agents in the system possess. Decentralized discrete-event control problems, where agents must communicate in order to effect control, are examined. Previously the identification of a state where agents should communicate was expressed solely in terms of properties of a formal language. The updated knowledge model presented here yields equivalent propositional logic expressions so that agents can determine where to communicate based on whether certain propositions are true at certain states. This amounts to agents making control decisions based on their "knowledge" of the system. The key relevant logic proposition corresponds to a check on whether a group of agents, if they pooled their knowledge, would possess sufficient information about the legality of a given event sequence. The new formulation is applied to a previously-solved example.