Models from Scenarios

Synthesis of Petri nets from behavioral descriptions has important applications in the design of systems in different application areas. In this paper we present a survey on the technique of region based synthesis of Petri nets from languages. Each word in a given language specifies one run of the searched Petri net, i.e. represents one observable scenario of the system. We concentrate on recent developments for languages of different kinds of causal structures (such as partial orders and stratified order structures). Causal structures represent causal relationships between events of one run. Expressible causal relationships are for example direct and indirect causal dependency, concurrency and synchronicity of events. Concerning infinite languages, several possibilities of a finite representation are discussed. As the goal of synthesis, place/transition nets and inhibitor nets as well as several restrictions of these net classes are used. The presented framework integrates all classical results on sequential languages.