Acceleration of the Multiple-Try Metropolis algorithm using antithetic and stratified sampling

  • Radu V. CraiuEmail author
  • Christiane Lemieux


The Multiple-Try Metropolis is a recent extension of the Metropolis algorithm in which the next state of the chain is selected among a set of proposals. We propose a modification of the Multiple-Try Metropolis algorithm which allows for the use of correlated proposals, particularly antithetic and stratified proposals. The method is particularly useful for random walk Metropolis in high dimensional spaces and can be used easily when the proposal distribution is Gaussian. We explore the use of quasi Monte Carlo (QMC) methods to generate highly stratified samples. A series of examples is presented to evaluate the potential of the method.


Antithetic variates Markov Chain Monte Carlo Extreme antithesis Korobov rule Latin Hypercube sampling Quasi Monte Carlo Sobol’ sequence Multiple-Try Metropolis Random-Ray Monte Carlo 


  1. Atchade Y.F. and Perron F. 2005. Improving on the independent Metropolis-Hastings algorithm. Statistica Sinica 5: 3–18.MathSciNetGoogle Scholar
  2. Atchade Y.F. and Rosenthal J. 2005. On adaptive Markov Chain Monte Carlo algorithms. Bernoulli 11: 815–828.zbMATHMathSciNetCrossRefGoogle Scholar
  3. Bliss C.I. 1935. The calculation of the dosage-mortality curve. Ann. Appl. Biology 22: 134–167.CrossRefGoogle Scholar
  4. Caflisch R.E., Morokoff W., and Owen A.B. 1997. Valuation of mortgage-backed securities using Brownian bridges to reduce effective dimension. The Journal of Computational Finance 1: 27–46.Google Scholar
  5. Carlin B.P. and Louis T.A. 2000. Bayes and Emprirical Bayes Methods for Data Analysis, Chapman and Hall, London.Google Scholar
  6. Chen M.-H. and Schmeiser B.W. 1993. Performances of the Gibbs, hit-and-run, and Metropolis samplers. J. Computat. Graph. Statist. 2: 251–272.CrossRefMathSciNetGoogle Scholar
  7. Craiu R.V. and Meng X.L. 2005. Multi-process parallel antithetic coupling for backward and forward MCMC. Ann. Statist. 33: 661–697.zbMATHCrossRefMathSciNetGoogle Scholar
  8. Cranley R. and Patterson T. 1976. Randomization of number theoretic methods for multiple integration. SIAM J. Numer. Anal. 13: 904–914.zbMATHCrossRefMathSciNetGoogle Scholar
  9. Draper N.R. and Smith H. 1981. Applied Regression Analysis and Its Applications, Wiley, New York.Google Scholar
  10. Frigessi A., Gåsemyr J., and Rue H. 2000. Antithetic coupling of two Gibbs sampler chains. Ann. Statist. 28: 1128–1149.zbMATHCrossRefMathSciNetGoogle Scholar
  11. Gelman A. and Meng X.L. 1991. A note on bivariate distributions that are conditionally normal. Am. Statistician 125–126.Google Scholar
  12. Gilks W.R., Roberts R.O., and George E.I. 1994. Adaptive direction sampling. Statistician 43: 179–189.CrossRefGoogle Scholar
  13. L'Ecuyer P. and Lemieux C. 2002. Recent advances in randomized quasi-Monte Carlo methods. In: Dror M., L'Ecuyer P., and F. Szidarovszki (Eds.), Modelling Uncertainty: An Examination of Stochastic Theory, Methods, and Applications, Kluwer Academic Publishers, Boston, pp. 419–474.Google Scholar
  14. Lemieux C. and Sidorsky P. 2006. Exact sampling with highly-uniform point sets. Mathematical and Computer Modelling 43: 339–349.CrossRefMathSciNetzbMATHGoogle Scholar
  15. Lindstrom M.J. and Bates D.M. 1990. Nonlinear mixed effects models for repeated measures data. Biometrics 46: 673–687.CrossRefMathSciNetGoogle Scholar
  16. Liu J.S., Liang F., and Wong W.H. 2000. The use of Multiple-Try method and local optimization in Metropolis sampling. J. Amer. Statist. Assoc. 95: 121–134.zbMATHCrossRefMathSciNetGoogle Scholar
  17. Loh W.L. 1996. On Latin hypercube sampling. Ann. Statist. 24: 2058–2080.zbMATHCrossRefMathSciNetGoogle Scholar
  18. Marsaglia G. 1962. Random variables and computers. In: Koseznik J. (Ed.), Information Theory and Statistical Decision Functions Random Processes: Transactions of the Third Prague Conference, Czechoslovak Academy of Sciences, Prague, pp. 499–510.Google Scholar
  19. McKay M.D., Beckman R.J., and Conover W.J. 1979. A comparison of three methods for selecting values of input variables in the analysis of output from a computer code. Technometrics 21: 239–245.zbMATHCrossRefMathSciNetGoogle Scholar
  20. 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 Scholar
  21. Owen A.B. 1992. A central limit theorem for Latin hypercube sampling. J. Roy. Statist. Soc. Ser. B 54: 541–551.zbMATHMathSciNetGoogle Scholar
  22. Owen A.B. 2003. Quasi Monte Carlo. Presented at the First Cape Cod Workshop on Monte Carlo Methods.Google Scholar
  23. Owen A.B. and Tribble S.D. 2004. A quasi-Monte Carlo Metropolis algorithm. Technical Report 2004-35, Department of Statistics, Stanford University.Google Scholar
  24. Paskov S. and Traub J. 1995. Faster valuation of financial derivatives. Journal of Portfolio Management 22: 113–120.CrossRefGoogle Scholar
  25. Ritter C. and Tanner M.A. 1992. Facilitating the Gibbs sampler: the Gibbs stopper and the griddy-Gibbs sampler. J. Amer. Statist. Assoc. 87: 861–868.CrossRefGoogle Scholar
  26. Stein M. 1987. Large sample properties of simulations using Latin hypercube sampling. Technometrics 29: 143–151, (Correction: 32, 367).Google Scholar
  27. Sobol' I.M. 1967. The distribution of points in a cube and the approximate evaluation of integrals. U.S.S.R. Comput. Math. and Math. Phys. 7: 86–112.CrossRefMathSciNetGoogle Scholar
  28. Tan Z. 2006. Monte Carlo integration with acceptance-rejection. J. Comput. Graphical Statist., to appear.Google Scholar
  29. van Dyk D. and Meng X.L. 2001. The art of data augmentation (with discussion). J. Comput. Graphical Statist. 10: 1–111.CrossRefMathSciNetGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  1. 1.Department of StatisticsUniversity of TorontoTorontoCanada
  2. 2.Department of Statistics and Actuarial ScienceUniversity of Waterloo, 200 University Avenue WestWaterlooCanada

Personalised recommendations