Rare Event Simulation with Fully Automated Importance Splitting
Probabilistic model checking is a powerful tool for analysing probabilistic systems but it can only be efficiently applied to Markov models. Monte Carlo simulation provides an alternative for the generality of stochastic processes, but becomes infeasible if the value to estimate depends on the occurrence of rare events. To combat this problem, intelligent simulation strategies exist to lower the estimation variance and hence reduce the simulation time. Importance splitting is one such technique, but requires a guiding function typically defined in an ad hoc fashion by an expert in the field. We present an automatic derivation of the importance function from the model description. A prototypical tool was developed and tested on several Markov models, compared to analytically and numerically calculated results and to results of typical ad hoc importance functions, showing the feasibility and efficiency of this approach. The technique is easily adapted to general models like GSMPs.
KeywordsRare Event Goal State Importance Sampling Importance Function Tandem Queue
Unable to display preview. Download preview PDF.
- 2.Garvels, M.J.J.: The splitting method in rare event simulation. PhD thesis, University of Twente (2000)Google Scholar
- 6.Kahn, H., Harris, T.E.: Estimation of particle transmission by random sampling. National Bureau of Standards Applied Mathematics Series 12, 27–30 (1951)Google Scholar
- 9.Law, A.M., Kelton, W.D., Kelton, W.D.: Simulation modeling and analysis, vol. 2. McGraw-Hill, New York (1991)Google Scholar
- 10.L’Ecuyer, P., Demers, V., Tuffin, B.: Rare events, splitting, and quasi-Monte Carlo. ACM Trans. Model. Comput. Simul. 17(2) (April 2007)Google Scholar
- 11.L’Ecuyer, P., Le Gland, F., Lezaud, P., Tuffin, B.: Splitting techniques. In: Rare Event Simulation using Monte Carlo Methods, pp. 39–61. J. Wiley & Sons (2009)Google Scholar
- 12.L’Ecuyer, P., Mandjes, M., Tuffin, B.: Importance sampling in rare event simulation. In: Rare Event Simulation using Monte Carlo Methods, pp. 17–38. J. Wiley & Sons (2009)Google Scholar
- 15.Villén-Altamirano, J.: RESTART method for the case where rare events can occur in retrials from any threshold. Int. J. Electron. Commun. (AEÜ) 52, 183–189 (1998)Google Scholar
- 17.Villén-Altamirano, M., Martínez-Marrón, A., Gamo, J., Fernández-Cuesta, F.: Enhancement of the accelerated simulation method restart by considering multiple thresholds. In: Proc. 14th Int. Teletraffic Congress, pp. 797–810 (1994)Google Scholar
- 19.Villén-Altamirano, M., Villén-Altamirano, J.: The Rare Event Simulation Method RESTART: Efficiency Analysis and Guidelines for Its Application. In: Kouvatsos, D.D. (ed.) Next Generation Internet: Performance Evaluation and Applications. LNCS, vol. 5233, pp. 509–547. Springer, Heidelberg (2011) CrossRefGoogle Scholar