Dynamic GSPNs: formal definition, transformation towards GSPNs and formal verification

In their basic form, Petri nets face several shortcomings in the design/analysis of reconfigurable discrete-event systems. Hence, researchers have enriched Petri nets with dynamic structure. Such formalisms allow explicitly modeling system features and its structure evolution in two separate levels. To reuse available tools in verifying dynamic-structure Petri nets, several formalisms flatten dynamic nets into basic Petri nets. To the best of our knowledge, this kind of approaches copes only with dynamic topology, i.e. sets of places and transitions cannot be changed. However, this restriction limits severely the modeling power of these formalisms. In this paper, we propose a new formalism, called dynamic generalized stochastic Petri nets (D-GSPNs), that allows modeling dynamic sets of places and transitions, as well as, keeping the possibility to transform D-GSPNs into semantic-equivalent GSPNs that preserve their stochastic behaviors. Hence, D-GSPNs preserve the property decidability of GSPNs, as well as, the panoply of available verification tools of GSPNs is still usable in D-GSPNs analysis. The feasibility of the proposed formalism is demonstrated on an illustrative example of a reconfigurable manufacturing system.