Asbstract
By casting stochastic optimal estimation of time series in path integral form, one can apply analytical and computational techniques of equilibrium statistical mechanics. In particular, one can use standard or accelerated Monte Carlo methods for smoothing, filtering and/or prediction. Here we demonstrate the applicability and efficiency of generalized (nonlocal) hybrid Monte Carlo and multigrid methods applied to optimal estimation, specifically smoothing. We test these methods on a stochastic diffusion dynamics in a bistable potential. This particular problem has been chosen to illustrate the speedup due to the nonlocal sampling technique, and because there is an available optimal solution which can be used to validate the solution via the hybrid Monte Carlo strategy. In addition to showing that the nonlocal hybrid Monte Carlo is statistically accurate, we demonstrate a significant speedup compared with other strategies, thus making it a practical alternative to smoothing/filtering and data assimilation on problems with state vectors of fairly large dimensions, as well as a large total number of time steps.
Similar content being viewed by others
References
B. Eraker (2001) ArticleTitleMCMC analysis of diffusion processes with application to finance J. Bus. Econ. Stati. 19 IssueID2 177–191 Occurrence Handle10.1198/073500101316970403
J.C. Hargreaves JD. Annan (2002) ArticleTitleAssimilation of paleo-data in a simple earth system model Clim. Dynam. 19 371–381 Occurrence Handle10.1007/s00382-002-0241-0
C. Wunsch (1996) The Ocean Circulation Inverse Problem Cambridge University Press Cambridge, UK
E. Kalnay (2003) Atmospheric Modeling, Data Assimilation and Predictability Cambridge University Press Cambridge
J. Liu C. Sabatti (1998) ArticleTitleSimulated sintering: MCMC with spaces of varying dimensions Bayesian Stat 6 386–413
J. Liu C. Sabatti (2000) ArticleTitleGeneralized Gibbs sampler and multigrid Monte Carlo for Bayesian computation Biometrika 87 353–369 Occurrence Handle10.1093/biomet/87.2.353
H. Sorensen (2004) ArticleTitleParametric inference for diffusion processes observed at discrete points in time: A survey Internat. Statist. Rev. 72 IssueID3 337–354
WR. Gilks S. Richardson D.J. Spiegelhalter (Eds) (1996) Markov Chain Monte Carlo in Practice Chapman and Hall New York
Chen, Bayesian filtering: From Kalman filters to particle filters and beyond, McMaster University Technical Report, 2003
JS. Liu (2001) Monte Carlo Strategies in Scientific Computing Springer New York
ChenM-H. Q-M. Shao JG. Ibrahim (2000) Monte Carlo Methods in Bayesian Computation Springer-Verlag New York, Berlin
B. Pendleton S. Duane A.D. Kennedy D. Roweth (1987) ArticleTitleHybrid Monte Carlo Phys. Lett. B. 195 216–222 Occurrence Handle10.1016/0370-2693(87)91197-X
Neal RM., Probabilistic inference using Markov chain Monte Carlo methods, Technical Report CRG-TR-93-1, Dept. of Computer Science, University of Toronto, 1993
R. Toral AL. Ferreira (1994) ArticleTitleA general class of hybrid Monte Carlo methods Proceedings of Physics Computing. 94 265–268
A.D. Kennedy B. Pendelton (2001) ArticleTitleCost of the generalised hybrid Monte Carlo algorithm for free field theory Nucl. Phys. B. B 607 456–510 Occurrence Handle10.1016/S0550-3213(01)00129-8
R. Salazar R. Toral (1997) ArticleTitleSimulated annealing using hybrid Monte Carlo J. Stat. Phys. 89 1047–1060
R. Salazar R. Toral (1997) ArticleTitleHybrid simulated annealing J. Stat. Phys. 89 1047
A. Gelb (1974) Applied Optimal Estimation MIT Press Cambridge, MA
H. Tanizaki (1996) Nonlinear Filters: Estimation and Applications Springer-Verlag New York
P. Del Moral (2004) Feynman-Kac Formulae Springer-Verlag New York
A. Doucet N. deFreitas N. Gordon (2002) Sequential Monte Carlo Methods in Practice Springer-Verlag New York
RL. Stratonovich (1960) ArticleTitleConditional Markov processes Theor. Prob. Appl. 5 156–178 Occurrence Handle10.1137/1105015
HJ. Kushner (1962) ArticleTitleOn the differential equations satisfied by conditional probability densities of Markov processes, with applications J. SIAM Control Ser.A. 2 106–119 Occurrence Handle10.1137/0302009
HJ. Kushner (1967) ArticleTitleDynamical equations for optimal nonlinear filtering J. Diff. Eq. 3 179–190A Occurrence Handle10.1016/0022-0396(67)90023-X
E. Pardoux (1982) ArticleTitle’Equations du filtrage non linéla préet du lissage Stochastics 6 193–231
GL. Eyink JM. Restrepo FJ. Alexander (2004) A mean field approximation in data assimilation for nonlinear dynamics Physica D 194 347–368
P. Kloeden E. Platen (1992) Numerical Solution of Stochastic Differential Equations Springer-Verlag Berlin
F. Langouche D. Roeckaerts E. Tirapegui (1978) ArticleTitleOn the most probable path for diffusion processes J. Phys. A 11 L263–L268
R. Graham (1977) ArticleTitlePath integral formulation of general diffusion processes Zeitschrift fur Physik 26 281–290 Occurrence Handle10.1007/BF01312935
F. Langouche D. Roeckaerts E. Tirapegui (1979) ArticleTitleFunctional integral methods for stochastic fields Physica 95 252–274
A. Thomas Severini (2001) Likelihood Methods in Statistics Oxford University Press New York
O. Talagrand P. Courtier (1987) ArticleTitleVariational assimilation of meteorological observations with the adjoint vorticity equation. I: Theory Quart. J. Roy. Meteor. Soc. 113 1311–1328 Occurrence Handle10.1256/smsqj.47811
J. Goodman AD. Sokal (1989) ArticleTitleMultigrid Monte Carlo method. Conceptual foundations Phys. Rev. D 40 2035–2071 Occurrence Handle10.1103/PhysRevD.40.2035
R.H. Swendsen JS. Wang (1987) ArticleTitleNonuniversal critical dynamics in Monte Carlo simulations Phys. Rev. Lett. 58 86–88 Occurrence Handle10.1103/PhysRevLett.58.86 Occurrence Handle10034599
E. Domany D. Kandel (1991) ArticleTitleGeneral cluster Monte Carlo dynamics Phys. Rev. B 43 8539 Occurrence Handle10.1103/PhysRevB.43.8539
GG. Batrouni GR. Katz AS. Kronfeld GP. Lepage B. Svetitsky KG. Wilson (1985) ArticleTitleLangevin simulations of lattice field theories Phys. Rev. D 32 2736–2747 Occurrence Handle10.1103/PhysRevD.32.2736
Alexander FJ., Boghosian BM., Brower RC., Kimura SR. (2001). Fourier acceleration of langevin molecular dynamics. Phys. Rev. E 066704
S. Caterall S. Karamov (2002) ArticleTitleTesting a Fourier accelerated hybrid Monte Carlo algorithm Phys. Lett. B B 528 301–305 Occurrence Handle10.1016/S0370-2693(02)01217-0
R. Toral A. Ferreira (1993) ArticleTitleHybrid Monte Carlo method for conserved-order-parameter systems Phys. Rev. E 47 3848–3851 Occurrence Handle10.1103/PhysRevE.47.4240
RN. Miller M.Ghil P. Gauthiez (1994) ArticleTitleAdvanced data assimilation in strongly non-linear dynamical systems J. Atmos. Sci. 51 1037–1056 Occurrence Handle10.1175/1520-0469(1994)051<1037:ADAISN>2.0.CO;2
G.L. Eyink JR. Restrepo (2000) ArticleTitleMost probable histories for nonlinear dynamics: Tracking climate transitions J. Stat. Phys. 101 459–472 Occurrence Handle10.1023/A:1026437432570
Godsill S., Doucet A., West M. (2001). Monte carlo smoothing for non-linear time series. URL: citeseer.ist.psu.edu/godsill01monte.html
Eyink GL., Restrepo JM., and Alexander FJ., Reducing computational complexity using closures in a mean field approach in data assimilation, submitted (2002)
A. Bennett (2002) Inverse Modeling of the Ocean and Atmosphere Cambridge University Press Cambridge, UK
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Alexander, F.J., Eyink, G.L. & Restrepo, J.M. Accelerated Monte Carlo for Optimal Estimation of Time Series. J Stat Phys 119, 1331–1345 (2005). https://doi.org/10.1007/s10955-005-3770-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10955-005-3770-1