Abstract
In the context of rare-event simulation, splitting and importance sampling (IS) are the primary approaches to make important rare events happen more frequently in a simulation and yet recover an unbiased estimator of the target performance measure, with much smaller variance than a straightforward Monte Carlo (MC) estimator. Randomized quasi-Monte Carlo (RQMC) is another class of methods for reducing the noise of simulation estimators, by sampling more evenly than with standard MC. It typically works well for simulations that depend mostly on very few random numbers. In splitting and IS, on the other hand, we often simulate Markov chains whose sample paths are a function of a long sequence of independent random numbers generated during the simulation. In this article, we show that RQMC can be used jointly with splitting and/or IS to construct better estimators than those obtained by either of these methods alone. We do that in a setting where the goal is to estimate the probability of reaching B before reaching (or returning to) A when starting A from a distinguished state not in B, where A and B are two disjoint subsets of the state space, and B is very rarely reached. This problem has several practical applications. The article is in fact a two-in-one: the first part provides a guided tour of splitting techniques, introducing along the way some improvements in the implementation of multilevel splitting. At the end of the article, we also give examples of situations where splitting is not effective. For these examples, we compare different ways of applying IS and combining it with RQMC.
- Andradóttir, S., Heyman, D. P., and Ott, T. J. 1995. On the choice of alternative measures in importance sampling with Markov chains. Oper. Res. 43, 3, 509--519.Google ScholarDigital Library
- Asmussen, S. 2002. Large deviations in rare events simulation: Examples, counterexamples, and alternatives. In Monte Carlo and Quasi-Monte Carlo Methods 2000, K.-T. Fang, F. J. Hickernell, and H. Niederreiter, Eds. Springer-Verlag, Berlin, 1--9.Google Scholar
- Booth, T. E. 1982. Automatic importance estimation in forward Monte Carlo calculations. Trans. Amer. Nuc. Soc. 41, 308--309.Google Scholar
- Booth, T. E. 1985. Monte Carlo variance comparison for expected-value versus sampled splitting. Nuc. Sci. Eng. 89, 305--309.Google ScholarCross Ref
- Booth, T. E. and Hendricks, J. S. 1984. Importance estimation in forward Monte Carlo estimation. Nucl. Tech./Fusion 5, 90--100.Google ScholarCross Ref
- Booth, T. E. and Pederson, S. P. 1992. Unbiased combinations of nonanalog Monte Carlo techniques and fair games. Nucl. Sci. Eng. 110, 254--261.Google ScholarCross Ref
- Bucklew, J. A. 2004. Introduction to Rare Event Simulation. Springer-Verlag, New York. Google ScholarDigital Library
- Cancela, H., Rubino, G., and Tuffin, B. 2002. MTTF estimation by Monte Carlo methods using Markov models. Monte Carlo Meth. Appl. 8, 4, 312--341.Google ScholarCross Ref
- Cancela, H., Rubino, G., and Tuffin, B. 2005. New measures of robustness in rare event simulation. In Proceedings of the 2005 Winter Simulation Conference, M. E. Kuhl, N. M. Steiger, F. B. Armstrong, and J. A. Joines, Eds. 519--527. Google ScholarDigital Library
- Cérou, F. and Guyader, A. 2005. Adaptive multilevel splitting for rare event analysis. Tech. Rep. 5710, INRIA. Oct.Google Scholar
- Cérou, F., LeGland, F., Del Moral, P., and Lezaud, P. 2005. Limit theorems for the multilevel splitting algorithm in the simulation of rare events. In Proceedings of the 2005 Winter Simulation Conference, F. B. A. M. E. Kuhl, N. M. Steiger and J. A. Joines, Eds. 682--691. Google ScholarDigital Library
- Chang, C. S., Heidelberger, P., Juneja, S., and Shahabuddin, P. 1994. Effective bandwidth and fast simulation of ATM intree networks. Perf. Eval. 20, 45--65. Google ScholarDigital Library
- Del Moral, P. 2004. Feynman-Kac Formulae. Genealogical and Interacting Particle Systems with Applications. Probability and its Applications. Springer, New York.Google Scholar
- Ermakov, S. M. and Melas, V. B. 1995. Design and Analysis of Simulation Experiments. Kluwer Academic, Dordrecht, The Netherlands.Google Scholar
- Fox, B. L. 1999. Strategies for Quasi-Monte Carlo. Kluwer Academic, Boston, MA.Google Scholar
- Garvels, M. J. J. 2000. The splitting method in rare event simulation. Ph.D. thesis, Faculty of mathematical Science, University of Twente, The Netherlands.Google Scholar
- Garvels, M. J. J. and Kroese, D. P. 1998. A comparison of RESTART implementations. In Proceedings of the 1998 Winter Simulation Conference. IEEE Press, 601--609. Google ScholarDigital Library
- Garvels, M. J. J., Kroese, D. P., and Van Ommeren, J.-K. C. W. 2002. On the importance function in splitting simulation. Euro. Trans. Telecomm. 13, 4, 363--371.Google ScholarCross Ref
- Glasserman, P., Heidelberger, P., Shahabuddin, P., and Zajic, T. 1998. A large deviations perspective on the efficiency of multilevel splitting. IEEE Trans. Auto. Cont. AC-43, 12, 1666--1679.Google ScholarCross Ref
- Glasserman, P., Heidelberger, P., Shahabuddin, P., and Zajic, T. 1999. Multilevel splitting for estimating rare event probabilities. Oper. Res. 47, 4, 585--600. Google ScholarDigital Library
- Glynn, P. W. 1994. Efficiency improvement techniques. Annals Oper. Res. 53, 175--197.Google ScholarCross Ref
- Glynn, P. W. and Iglehart, D. L. 1989. Importance sampling for stochastic simulations. Manage Sci. 35, 1367--1392. Google ScholarDigital Library
- Goyal, A., Shahabuddin, P., Heidelberger, P., Nicola, V. F., and Glynn, P. W. 1992. A unified framework for simulating Markovian models of highly reliable systems. IEEE Tran. Comput. C-41, 36--51. Google ScholarDigital Library
- Hammersley, J. M. and Handscomb, D. C. 1964. Monte Carlo Methods. Methuen, London.Google Scholar
- Harris, T. 1963. The Theory of Branching Processes. Springer-Verlag, New York.Google Scholar
- Heidelberger, P. 1995. Fast simulation of rare events in queueing and reliability models. ACM Trans. Model. Comput. Simul. 5, 1, 43--85. Google ScholarDigital Library
- Heidelberger, P., Shahabuddin, P., and Nicola, V. F. 1994. Bounded relative error in estimating transient measures of highly dependable non-Markovian systems. ACM Trans. Model. Comput. Simul. 4, 2, 137--164. Google ScholarDigital Library
- Hickernell, F. J. 2002. Obtaining o(n−2+ε) convergence for lattice quadrature rules. In Monte Carlo and Quasi-Monte Carlo Methods 2000, K.-T. Fang, F. J. Hickernell, and H. Niederreiter, Eds. Springer-Verlag, Berlin, 274--289.Google Scholar
- Juneja, S. and Shahabuddin, P. 2001. Fast simulation of Markov chains with small transition probabilities. Manage Sci. 47, 4, 547--562. Google ScholarDigital Library
- Kahn, H. and Harris, T. E. 1951. Estimation of particle transmission by random sampling. National Bureau of Standards Applied Mathematical Series 12, 27--30.Google Scholar
- Keilson, J. 1979. Markov Chain Models---Rarity and Exponentiality. Springer-Verlag, New York.Google Scholar
- Lécot, C. and Tuffin, B. 2004. Quasi-Monte Carlo methods for estimating transient measures of discrete time Markov chains. In Monte Carlo and Quasi-Monte Carlo Methods 2002, H. Niederreiter, Ed. Springer-Verlag, Berlin, 329--343.Google Scholar
- L'Ecuyer, P., Lécot, C., and Tuffin, B. 2005. A randomized quasi-Monte Carlo simulation method for Markov chains. INRIA Res. Rep. 5545 (April), Operations Research, to appear.Google Scholar
- L'Ecuyer, P., Lécot, C., and Tuffin, B. 2006. Randomized quasi-Monte Carlo simulation of Markov chains with an ordered state space. In Monte Carlo and Quasi-Monte Carlo Methods 2004, H. Niederreiter and D. Talay, Eds. 331--342.Google Scholar
- L'Ecuyer, P. and Lemieux, C. 2000. Variance reduction via lattice rules. Manage. Sci. 46, 9, 1214--1235. Google ScholarDigital Library
- L'Ecuyer, P. and Lemieux, C. 2002. Recent advances in randomized quasi-Monte Carlo methods. In Modeling Uncertainty: An Examination of Stochastic Theory, Methods, and Applications, M. Dror, P. L'Ecuyer, and F. Szidarovszky, Eds. Kluwer Academic, Boston, 419--474.Google Scholar
- L'Ecuyer, P. and Tuffin, B. 2006. Splitting with weight windows to control the likelihood ratio in importance sampling. In Proceedings of ValueTools 2006: International Conference on Performance Evaluation Methodologies and Tools. ACM Publications, Pisa, Italy. Google ScholarDigital Library
- Melas, V. B. 1997. On the efficiency of the splitting and roulette approach for sensitivity analysis. In Proceedings of the 1997 Winter Simulation Conference. IEEE Press, Piscataway, NJ, 269--274. Google ScholarDigital Library
- Nakayama, M. K. 1996. General conditions for bounded relative error in simulations of highly reliable Markovian systems. Advances in Applied Probability 28, 687--727.Google ScholarCross Ref
- Nakayama, M. K. and Shahabuddin, P. 1998. Likelihood ratio derivative estimation for finite-time performance measures in generalized Semi-Markov processes. Manage. Sci. 44, 10, 1426--1441. Google ScholarDigital Library
- Nicola, V. F., Shahabuddin, P., and Heidelberger, P. 1993. Techniques for fast simulation of highly dependable systems. In Proceedings of the 2nd International Workshop on Performability Modelling of Computer and Communication Systems.Google Scholar
- Niederreiter, H. 1992. Random Number Generation and Quasi-Monte Carlo Methods. SIAM CBMS-NSF Regional Conference Series in Applied Mathematics, vol. 63. SIAM, Philadelphia. Google ScholarDigital Library
- Owen, A. B. 1998. Latin supercube sampling for very high-dimensional simulations. ACM Trans. Model. Comput. Simul. 8, 1, 71--102. Google ScholarDigital Library
- Owen, A. B. 2003. Variance with alternative scramblings of digital nets. ACM Trans. Model. Comput. Simul. 13, 4, 363--378. Google ScholarDigital Library
- Papadopoulos, C. 1998. A new technique for MTTF estimation in highly reliable Markovian systems. Monte Carlo Methods and Applications 4, 2, 95--112.Google ScholarCross Ref
- Parekh, S. and Walrand, J. 1989. A quick simulation method for excessive backlogs in networks of queues. IEEE Trans. Auto. Cont. AC-34, 54--56.Google ScholarCross Ref
- Shahabuddin, P. 1994a. Fast transient simulation of Markovian models of highly dependable systems. Performance Eval. 20, 267--286. Google ScholarDigital Library
- Shahabuddin, P. 1994b. Importance sampling for the simulation of highly reliable Markovian systems. Manage. Sci. 40, 3, 333--352. Google ScholarDigital Library
- Shahabuddin, P. 1995. Rare event simulation in stochastic models. In Proceedings of the 1995 Winter Simulation Conference. IEEE Press, 178--185. Google ScholarDigital Library
- Shahabuddin, P., Nicola, V. F., Heidelberger, P., Goyal, A., and Glynn, P. W. 1988. Variance reduction in mean time to failure simulations. In Proceedings of the 1988 Winter Simulation Conference. IEEE Press, 491--499. Google ScholarDigital Library
- Taylor, H. M. and Karlin, S. 1998. An Introduction to Stochastic Modeling, third ed. Academic Press, San Diego.Google Scholar
- Tuffin, B. 1999. Bounded normal approximation in simulations of highly reliable Markovian systems. J. Appl. Probab. 36, 4, 974--986.Google ScholarCross Ref
- Tuffin, B. 2004. On numerical problems in simulations of highly reliable Markovian systems. In Proceedings of the 1st International Conference on Quantitative Evaluation of SysTems (QEST). IEEE CS Press, University of Twente, Enschede, The Netherlands, 156--164. Google ScholarCross Ref
- Vasicek, O. 1977. An equilibrium characterization of the term structure. J. Financ. Econ. 5, 177--188.Google ScholarCross Ref
- Villén-Altamirano, J. 2006. Rare event RESTART simulation of two-stage networks. Europ. J. Oper. Res. to appear.Google Scholar
- Villén-Altamirano, M., Martinez-Marrón, A., Gamo, J., and Fernández-Cuesta, F. 1994. Enhancement of the accelerated simulation method restart by considering multiple thresholds. In Proceedings of the 14th International Teletraffic Congress. Elsevier Science, 797--810.Google Scholar
- Villén-Altamirano, M. and Villén-Altamirano, J. 1994. RESTART: A straightforward method for fast simulation of rare events. In Proceedings of the 1994 Winter Simulation Conference. IEEE Press, 282--289. Google ScholarDigital Library
- Villén-Altamirano, M. and Villén-Altamirano, J. 2002. Analysis of RESTART simulation: Theoretical basis and sensitivity study. Europ Trans. Telecomm. 13, 4, 373--386.Google ScholarCross Ref
- Villén-Altamirano, M. and Villén-Altamirano, J. 2006. On the efficiency of RESTART for multidimensional systems. ACM Trans. Model. Comput. Simul. 16, 3, 251--279. Google ScholarDigital Library
- Wilson, J. R. 1983. Antithetic sampling with multivariate inputs. Amer. J. Math. Manage. Sci. 3, 121--144.Google ScholarCross Ref
Index Terms
- Rare events, splitting, and quasi-Monte Carlo
Recommendations
Coupling from the past with randomized quasi-Monte Carlo
The coupling-from-the-past (CFTP) algorithm of Propp and Wilson permits one to sample exactly from the stationary distribution of an ergodic Markov chain. By using it n times independently, we obtain an independent sample from that distribution. A more ...
Pseudo-marginal Markov Chain Monte Carlo for Nonnegative Matrix Factorization
A pseudo-marginal Markov chain Monte Carlo (PMCMC) method is proposed for nonnegative matrix factorization (NMF). The sampler jointly simulates the joint posterior distribution for the nonnegative matrices and the matrix dimensions which indicate the ...
Efficient Monte Carlo simulation via the generalized splitting method
We describe a new Monte Carlo algorithm for the consistent and unbiased estimation of multidimensional integrals and the efficient sampling from multidimensional densities. The algorithm is inspired by the classical splitting method and can be applied ...
Comments