속산 시뮬레이션을 위한 적응형 비모수 중요 샘플링 기법

Non-parametric Adaptive Importance Sampling for Fast Simulation Technique

  • 김윤배 (성균관대학교 시스템경영공학부)
  • 발행 : 1999.09.01

초록

Simulating rare events, such as probability of cell loss in ATM networks, machine failure in highly reliable systems, requires huge simulation efforts due to the low chance of occurrence. Importance Sampling (IS) has been applied to accelerate the occurrence of rare events. However, it has a drawback of effective biasing scheme to make the estimator of IS unbiased. Adaptive Importance Sampling (AIS) employs an estimated sampling distribution of IS to the system of interest during the course of simulation. We propose Nonparametric Adaptive Importance Sampling (NAIS) technique which is nonparametrical version of AIS. We test NAIS to estimate a probability of rare event in M/M/1 queueing model. Comparing with classical Monte Carlo simulation, the computational efficiency and variance reductions gained via NAIS are substantial. A possible extension of NAIS regarding with random number generation is also discussed.

키워드

참고문헌

  1. IEEE Trans. Communications v.43 Stochastic gradient optimization of importance sampling for the efficient simulation of digital communication systems Al-Qaq, W. A.;Devetsikiotis,M;Townsend, J. K.
  2. Discrete-Event Simulation Banks J;Carson II, J. S.;Nelson, B. L.
  3. Structural Safety v.5 Adaptive Sampling - An Iterative Fast Monet Carlo Procedure Bucher, C. G.
  4. Large Deviation Techniques in Decision Simulation and Estimation Bucklew, J. A
  5. Proceedings of the 1993 Winter Simulation Conference Simulation of Rare Queueing Events by Switching Arrival and Service Rates Cheng, R. C. H;Taylor T;Sztrik J
  6. IEEE Trans. Communications v.41 An algorithmic approach to the optimization of importance sampling parameters in digital communication system simulation Devetsikiotis, M;Townsend, J. K
  7. IEEE Communication Magazine v.32 Traffic Modeleling for Telecommunications Network Frost, V. S.;Melamed, B
  8. Management Science v.35 Importance Sampling for Stochastic Simulations Glynn, P.W.;Iglehart, D. L.
  9. IEEE Journal of Selected Areas in Communications v.SAC-2 Techniques for Estimating Bit Error Rate in The Simulation of Digital Communication systems Jerchim, M. C.
  10. Journal of America Statistical Association v.41 Adaptive Importance Sampling in Monte Carlo Importance Sampling Oh, M. S.;Berger, J
  11. Managemet Science v.40 no.3 Importance Sampling for the Simulation of Highly Reliable Markovian Systems Shahabuddin, P
  12. Proceedings of the 1994 Winter Simulation Conference Fast Simulation of Packet Loss Rates In Communication Network with Priorities Shahabuddin, P
  13. The Annals of Statistics v.4 no.4 Importance Sampling in Monte Carlo Study of Sequential Tests Siegmud, D
  14. Density Estimation Silverman, B. W.
  15. IEEE Journal on Selected Areas in Communications v.11 no.3 Adaptive Importance Sampling Stadler, J. S;Roy, S
  16. Nonparametric Function Estimation, Modeling, and Simulation, Society for Industrial and Applied Mathematics Thompson, J. R.;Tapia, R. A.
  17. in Baysian Statistics(4rd ed) West, M
  18. Journal of Royal Society, Ser. B v.55 Approximation Posterior Distribuions by Mixtures West, M
  19. Jouranl of America Statistical Association v.91 no.432 Nonparametric Importance Sampling Zhang, P