Vector team automata (Q418753)

Team automata are composed out of sequential automata interacting through synchronizations on common actions, with a synchronous execution of a global action changing the local states of the automata that take part in the synchronization. The paper explores the relationship between team automata and Petri nets which are a model of concurrent systems, where the execution of an action is local in the sense that it only depends and affects local states in the immediate neighborhood of the action. Synchronization in a team automaton, on the other hand, in general depends on the global state which is referred to as state-sharing, and sometimes it is impossible to determine which automata take part in a given synchronization. Vector team automata are a class of team automata with an explicit representation of synchronizations in the form of vectors of component actions from which one can deduce automata involved in synchronizations. The paper shows how one can capture the behavior of non-state-sharing vector team automata by individual token net controllers. The latter are a class of state machine decomposable labeled Petri nets with a synchronization mechanism based on vector labels. It becomes therefore possible to transfer and apply the concurrency semantics developed for Petri nets to the vector team automata model.