The Annals of Applied Probability

Randomized Hamiltonian Monte Carlo

Nawaf Bou-Rabee and Jesús María Sanz-Serna

Full-text: Access denied (no subscription detected)

We're sorry, but we are unable to provide you with the full text of this article because we are not able to identify you as a subscriber. If you have a personal subscription to this journal, then please login. If you are already logged in, then you may need to update your profile to register your subscription. Read more about accessing full-text

Abstract

Tuning the durations of the Hamiltonian flow in Hamiltonian Monte Carlo (also called Hybrid Monte Carlo) (HMC) involves a tradeoff between computational cost and sampling quality, which is typically challenging to resolve in a satisfactory way. In this article, we present and analyze a randomized HMC method (RHMC), in which these durations are i.i.d. exponential random variables whose mean is a free parameter. We focus on the small time step size limit, where the algorithm is rejection-free and the computational cost is proportional to the mean duration. In this limit, we prove that RHMC is geometrically ergodic under the same conditions that imply geometric ergodicity of the solution to underdamped Langevin equations. Moreover, in the context of a multidimensional Gaussian distribution, we prove that the sampling efficiency of RHMC, unlike that of constant duration HMC, behaves in a regular way. This regularity is also verified numerically in non-Gaussian target distributions. Finally, we suggest variants of RHMC for which the time step size is not required to be small.

Article information

Source
Ann. Appl. Probab., Volume 27, Number 4 (2017), 2159-2194.

Dates
Received: November 2015
Revised: October 2016
First available in Project Euclid: 30 August 2017

Permanent link to this document
https://projecteuclid.org/euclid.aoap/1504080029

Digital Object Identifier
doi:10.1214/16-AAP1255

Mathematical Reviews number (MathSciNet)
MR3693523

Zentralblatt MATH identifier
1373.60129

Subjects
Primary: 60J25: Continuous-time Markov processes on general state spaces
Secondary: 62D05: Sampling theory, sample surveys 60J25: Continuous-time Markov processes on general state spaces 60H30: Applications of stochastic analysis (to PDE, etc.) 37A50: Relations with probability theory and stochastic processes [See also 60Fxx and 60G10]

Keywords
Randomization Markov chain Monte Carlo Hamiltonian Monte Carlo hybrid Monte Carlo Lyapunov functions geometric ergodicity integrated autocorrelation time equilibrium mean squared displacement

Citation

Bou-Rabee, Nawaf; Sanz-Serna, Jesús María. Randomized Hamiltonian Monte Carlo. Ann. Appl. Probab. 27 (2017), no. 4, 2159--2194. doi:10.1214/16-AAP1255. https://projecteuclid.org/euclid.aoap/1504080029


Export citation

References

  • [1] Akhmatskaya, E. and Reich, S. (2008). GSHMC: An efficient method for molecular simulation. J. Comput. Phys. 227 4934–4954.
  • [2] Andersen, H. C. (1980). Molecular dynamics simulations at constant pressure and/or temperature. J. Chem. Phys. 72 2384.
  • [3] Asmussen, S. and Glynn, P. W. (2007). Stochastic Simulation: Algorithms and Analysis. Stochastic Modelling and Applied Probability 57. Springer, New York.
  • [4] Beskos, A., Pinski, F. J., Sanz-Serna, J. M. and Stuart, A. M. (2011). Hybrid Monte-Carlo on Hilbert spaces. Stochastic Process. Appl. 121 2201–2230.
  • [5] Blanes, S., Casas, F. and Sanz-Serna, J. M. (2014). Numerical integrators for the hybrid Monte Carlo method. SIAM J. Sci. Comput. 36 A1556–A1580.
  • [6] Bou-Rabee, N. and Owhadi, H. (2010). Long-run accuracy of variational integrators in the stochastic context. SIAM J. Numer. Anal. 48 278–297.
  • [7] Bou-Rabee, N. and Vanden-Eijnden, E. (2016). Continuous-time random walks for the numerical solution of stochastic differential equations. Memoirs of the AMS. To appear.
  • [8] Cancès, E., Legoll, F. and Stoltz, G. (2007). Theoretical and numerical comparison of some sampling methods for molecular dynamics. ESAIM Math. Model. Numer. Anal. 41 351–389.
  • [9] Davis, M. H. A. (1993). Markov Models and Optimization. Monographs on Statistics and Applied Probability 49. Chapman & Hall, London.
  • [10] Da Prato, G. and Zabczyk, J. (1996). Ergodicity for Infinite-Dimensional Systems. London Mathematical Society Lecture Note Series 229. Cambridge Univ. Press, Cambridge.
  • [11] Duane, S., Kennedy, A. D., Pendleton, B. J. and Roweth, D. (1987). Hybrid Monte-Carlo. Phys. Lett. B 195 216–222.
  • [12] E, W. and Li, D. (2008). The Andersen thermostat in molecular dynamics. Comm. Pure Appl. Math. 61 96–136.
  • [13] Fang, Y., Sanz-Serna, J. M. and Skeel, R. D. (2014). Compressible generalized hybrid Monte Carlo. J. Chem. Phys. 140 174108.
  • [14] Goodman, J. Acor, statistical analysis of a time series. Available at http://www.math.nyu.edu/faculty/goodman/software/acor/. Accessed 06–18–06–2015.
  • [15] Hairer, M. and Mattingly, J. C. (2011). Yet another look at Harris’ ergodic theorem for Markov chains. In Seminar on Stochastic Analysis, Random Fields and Applications VI 109–117. Springer, Basel.
  • [16] Hastings, W. K. (1970). Monte Carlo sampling methods using Markov chains and their applications. Biometrika 57 97–109.
  • [17] Homan, M. D. and Gelman, A. (2014). The no-u-turn sampler: Adaptively setting path lengths in Hamiltonian Monte Carlo. J. Mach. Learn. Res. 15 1593–1623.
  • [18] Horowitz, A. M. (1991). A generalized guided Monte-Carlo algorithm. Phys. Lett. A 268 247–252.
  • [19] Kennedy, A. D. and Pendleton, B. (2001). Cost of the generalised hybrid Monte Carlo algorithm for free field theory. Nuclear Phys. B 607 456–510.
  • [20] Kushner, H. J. and Dupuis, P. (2001). Numerical Methods for Stochastic Control Problems in Continuous Time, 2nd ed. Applications of Mathematics (New York) 24. Springer, New York.
  • [21] Li, D. (2007). On the rate of convergence to equilibrium of the Andersen thermostat in molecular dynamics. J. Stat. Phys. 129 265–287.
  • [22] Liu, J. S. (2008). Monte Carlo Strategies in Scientific Computing, 2nd ed. Springer, New York.
  • [23] Mackenzie, P. B. (1989). An improved hybrid Monte Carlo method. Phys. Lett. B 226 369–371.
  • [24] Marsden, J. E. and Ratiu, T. S. (1999). Introduction to Mechanics and Symmetry, 2nd ed. Texts in Applied Mathematics 17. Springer, New York.
  • [25] Marsden, J. E. and West, M. (2001). Discrete mechanics and variational integrators. Acta Numer. 10 357–514.
  • [26] Martin, M. G. and Siepmann, J. I. (1998). Transferable potentials for phase equilibria. 1. United-atom description of n-alkanes. J. Phys. Chem., B 102 2569–2577.
  • [27] Mattingly, J. C., Stuart, A. M. and Higham, D. J. (2002). Ergodicity for SDEs and approximations: Locally Lipschitz vector fields and degenerate noise. Stochastic Process. Appl. 101 185–232.
  • [28] Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H. and Teller, E. (1953). Equations of state calculations by fast computing machines. J. Chem. Phys. 21 1087–1092.
  • [29] Meyn, S. and Tweedie, R. (1993). Stability of Markovian processes III: Foster–Lyapunov criteria for continuous-time processes. Adv. in Appl. Probab. 25 518–548.
  • [30] Meyn, S. and Tweedie, R. L. (2009). Markov Chains and Stochastic Stability, 2nd ed. Cambridge Univ. Press, Cambridge.
  • [31] Neal, R. M. (2011). MCMC using Hamiltonian dynamics. In Handbook of Markov Chain Monte Carlo. Chapman & Hall/CRC Handb. Mod. Stat. Methods 113–162. CRC Press, Boca Raton, FL.
  • [32] Protter, P. E. (2004). Stochastic Integration and Differential Equations, 2nd ed. Applications of Mathematics (New York) 21. Springer, Berlin.
  • [33] Sanz-Serna, J. M. (2014). Markov Chain Monte Carlo and numerical differential equations. In Current Challenges in Stability Issues for Numerical Differential Equations (L. Dieci and N. Guglielmi, eds.) 2082 39–88. Springer, Berlin.
  • [34] Sanz-Serna, J. M. and Calvo, M. P. (1994). Numerical Hamiltonian Problems. Applied Mathematics and Mathematical Computation 7. Chapman & Hall, London.
  • [35] Sanz-Serna, J. M. and Stuart, A. M. (1999). Ergodicity of dissipative differential equations subject to random impulses. J. Differential Equations 155 262–284.
  • [36] Sokal, A. (1997). Monte Carlo methods in statistical mechanics: Foundations and new algorithms. In Functional Integration (Cargèse, 1996). NATO Adv. Sci. Inst. Ser. B Phys. 361 131–192. Plenum, New York.
  • [37] Talay, D. (2002). Stochastic Hamiltonian systems: Exponential convergence to the invariant measure, and discretization by the implicit Euler scheme. Markov Processes Related Fields 8 1–36.