Regenerative Simulation of Stochastic Petri Nets with Discrete and Continuous Timing

We present a regenerative simulation method applicable to a special class of non-Markovian stochastic Petri net (SPN) that may be useful in practice for constructing and solving performability models. Introduced in [1], this special SPN is called a phased delay Petri net (PDPN) and was conceived to allow efficient numerical analysis while increasing the modeling fidelity. The PDPN permits phasetype firing delays in both discrete and continuous time simultaneously in the same model. However, exact numerical analysis requires that the discrete-time transitions be synchronized when active. We propose here, instead, a new regenerative simulation technique that offers efficient analysis of PDPNs where numerical solution by exact means or other (known) types of regenerative simulation is too diffi-