Typically analysis goals either refer to qualitative properties, e.g. absence of deadlocks, liveness or boundedness, or to the determination of performance measures. Within QPN-Tool qualitative properties are investigated by a so-called qualitative analysis which is based on Petri net theory. Performance measures, on the other hand, are computed by analysis of the corresponding Markov chain, which is called quantitative analysis. A major advantage of a hierarchical model description is that isolated LLQPNs can be analysed assuming a well-behaving environment and that the HLQPN can be analysed assuming well-behaving LLQPNs.
Ahead of any qualitative analysis a consistency check is performed, where naming inconsistencies and mismatches between different levels of a hierarchical model are recognized. Subsequent to the consistency check, the embedded CPN of an isolated subnet or the HLQPN itself is unfolded and classified in terms of an uncoloured Place/Transition net: marked graph, state machine, free choice, etc. This classification neglects timing information, replaces a subnet place by its input and output place and connects these places via a virtual transition. It supports the choice of a suitable analysis algorithm for a local qualitative analysis, as special algorithms are available for certain net classes, see Fig. 5.