Failures semantics and deadlocking of modular Petri nets (Q1101233)

This paper studies the modular construction of Petri nets, where subnets are exchanged such that the behaviour of the whole net remains the same. Subnets and context nets are composed by merging transitions, corresponding to synchronization; but also the connection with new arcs is covered. It is investigated which subnets can be exchanged such that deadlock-freeness and maximal behaviour are preserved, and it is shown that some variations of the failure semantics, which has been developed for TCSP, give characterizations of exchangeable subnets.