Inhaltszusammenfassung

Timed dynamic systems can be modeled and simulated with Petri nets using very different approaches. In Clock Pulse Models (CPM), one marking represents exactly one moment in time and all enabled transitions fire simultaneously according to a global clock. This allows for real-time observation of the modeled system during a simulation run. Since simulation time is proportional to observed real-time, such an approach barely scales concerning the observed time horizon. If not the observable beha...Timed dynamic systems can be modeled and simulated with Petri nets using very different approaches. In Clock Pulse Models (CPM), one marking represents exactly one moment in time and all enabled transitions fire simultaneously according to a global clock. This allows for real-time observation of the modeled system during a simulation run. Since simulation time is proportional to observed real-time, such an approach barely scales concerning the observed time horizon. If not the observable behavior over time is of interest but the final simulation result, Event Triggered Models (ETM) can overcome this limitation as has been shown in former publications. A time-lapse model accelerates simulation runs by focusing on the moments state chances occur. CPM and ETM, however, share one disadvantage: the models' sizes tend to be proportional to the number of events that may occur simultaneously in the real world. Hence, they scale barely with the number of modeled entities. This unsolved scaling problem is addressed in this paper and pursues the idea of folding similar subnets in temporal layers called strata. The result is a Stratified Simulation Model (SSM), a small, flexible model that scales well concerning the number of modeled entities. SSM are derivates of CPM, since they still allow for nearly-real-time observation of the real world. Moreover, these models ease the visualization of the simulation in dashboards. The approach is explained at the example of a small production process.» weiterlesen» einklappen

Autoren

Klassifikation

DFG Fachgebiet:
Informatik

DDC Sachgruppe:
Informatik