Rare event simulation with an adaptive "RESTART" method in a Petri net modeling environment

For the performability evaluation of complex soft real time systems, simulation often remains the only feasible method. Simulation experiments tend to be very time consuming if rare events have to be considered. The paper describes a fast simulation approach for rare events implemented in a Petri net tool. The technique is based on the recently developed RESTART method, which is applicable for rare events in a wide range of simulation models and has the potential to drastically reduce the simulation overhead. The paper presents selection and refinement techniques for thresholds, which are the most important input parameters of RESTART. The results show run length reductions up to six orders of magnitude.