Efficient Reachability Analysis for Time Petri Nets

We propose in this paper some efficient approaches, based on the state class graph method, to construct abstractions for the Time Petri Net (TPN) model, suitable to verify its linear or reachability properties. Experimental results have shown that these abstractions are very appropriate as both time and size are considerably reduced. For some tested models, abstractions that preserve reachability properties can be as many as 2,051 times smaller and more than 592 times faster to compute. For abstractions, which are overapproximations (useful to prove that certain states are not reachable), gains can overpass 10,000 for both time and size.