Timing Analysis of Cyclic Time Petri Net Using Relaxed Unfolding and Global Time Technique

Time Petri nets are a powerful formalism for representing real-time systems. State Class Graph is a useful tool for timing analysis of time Petri nets. However, this tool is inaccurate when performing timing analysis of sequences of transitions firings. This paper proposes to overcome this limitation, taking advantage of two techniques: global time and relaxed unfolding. Global time technique can perform timing analysis in safe time Petri nets with more accuracy, but only on acyclic nets. This limitation can be overcome by using the unfolding technique to generate an acyclic occurrence net containing all reachable markings of the original safe Petri net. Although the unfolding process enumerates all reachable markings, it does not enumerate all possible transition firings sequences. For this reason, we propose a relaxed process that unfolds further than the original, by redefining cut-off events to identify existing temporal classes instead of existing markings. The process transforms a cyclic net in an acyclic net, with all classes and paths preserved, in which accurate timing analysis is feasible.