Abstract
We consider the numerical analysis of the time discretization of Feynman–Kac semigroups associated with diffusion processes. These semigroups naturally appear in several fields, such as large deviation theory, Diffusion Monte Carlo or non-linear filtering. We present error estimates à la Talay–Tubaro on their invariant measures when the underlying continuous stochastic differential equation is discretized; as well as on the leading eigenvalue of the generator of the dynamics, which corresponds to the rate of creation of probability. This provides criteria to construct efficient integration schemes of Feynman–Kac dynamics, as well as a mathematical justification of numerical results already observed in the Diffusion Monte Carlo community. Our analysis is illustrated by numerical simulations.
Similar content being viewed by others
References
Abdulle, A., Cohen, D., Vilmart, G., Zygalakis, K.C.: High weak order methods for stochastic differential equations based on modified equations. SIAM J. Sci. Comput. 34(3), A1800–A1823 (2012)
Abdulle, A., Vilmart, G., Zygalakis, K.C.: High order numerical approximation of the invariant measure of ergodic SDEs. SIAM J. Numer. Anal. 52(4), 1600–1622 (2014)
Abdulle, A., Vilmart, G., Zygalakis, K.C.: Long time accuracy of Lie–Trotter splitting methods for Langevin dynamics. SIAM J. Numer. Anal. 53(1), 1–16 (2015)
Anderson, J.B.: A random-walk simulation of the Schrödinger equation: H\(^+_3\). J. Chem. Phys. 63(4), 1499–1503 (1975)
Bou-Rabee, N., Owhadi, H.: Long-run accuracy of variational integrators in the stochastic context. SIAM J. Numer. Anal. 48(1), 278–297 (2010)
Ceperley, D.M., Alder, B.: Ground state of the electron gas by a stochastic method. Phys. Rev. Lett. 45(7), 566 (1980)
Dautray, R., Lions, J.-L.: Mathematical Analysis and Numerical Methods for Science and Technology, Volume 5: Evolution Problems I. Springer, Berlin (2000)
Debussche, A., Faou, E.: Weak backward error analysis for SDEs. SIAM J. Numer. Anal. 50(3), 1735–1752 (2012)
Del Moral, P.: Feynman–Kac Formulae. Springer, Berlin (2004)
Del Moral, P., Guionnet, A.: On the stability of interacting processes with applications to filtering and genetic algorithms. Annales de l’IHP Probabilités et Statistiques 37(2), 155–194 (2001)
Del Moral, P., Miclo, L.: Branching and interacting particle systems approximations of Feynman–Kac formulae with applications to non-linear filtering. In: Séminaire de probabilités XXXIV, pp. 1–145. Springer (2000)
Del Moral, P., Miclo, L.: Particle approximations of Lyapunov exponents connected to Schrödinger operators and Feynman–Kac semigroups. ESAIM Probab. Stat. 7, 171–208 (2003)
Dembo, A., Zeitouni, O.: Large Deviations Techniques and Applications. Stochastic Modelling and Applied Probability, vol. 38. Springer, Berlin (2010)
den Hollander, F.: Large Deviations, Volume 14 of Fields Institute Monographs. American Mathematical Society (2000)
Dobrushin, R.L.: Central limit theorem for nonstationary Markov chains. I. Theory Probab. Appl. 1(1), 65–80 (1956)
Dobrushin, R.L.: Central limit theorem for nonstationary Markov chains. II. Theory Probab. Appl. 1(4), 329–383 (1956)
Donsker, M.D., Varadhan, S.R.S.: On a variational formula for the principal eigenvalue for operators with maximum principle. Proc. Natl. Acad. Sci. 72(3), 780–783 (1975)
Douc, R., Moulines, E.: Limit theorems for weighted samples with applications to sequential Monte Carlo methods. ESAIM Proc. 19, 101–107 (2007)
Douc, R., Moulines, E., Olsson, J.: Long-term stability of sequential Monte Carlo methods under verifiable conditions. Ann. Appl. Probab. 24(5), 1767–1802 (2014)
Doucet, A., de Freitas, N., Gordon, N.: Sequential Monte Carlo Methods in Practice. Springer, Cham (2001)
Du, Y.: Order Structure and Topological Methods in Nonlinear Partial Differential Equations, Volume 1: Maximum Principles and Applications. World Scientific, Singapore (2006)
El Makrini, M., Jourdain, B., Lelièvre, T.: Diffusion Monte Carlo method: numerical analysis in a simple case. ESAIM Math. Model. Numer. Anal. 41(2), 189–213 (2007)
Evans, L.C.: Partial Differential Equations, Volume 19 of Graduate Studies in Mathematics. American Mathematical Society, New York (2010)
Fathi, M., Stoltz, G.: Improving dynamical properties of stabilized discretizations of overdamped Langevin dynamics. Numer. Math. 136(2), 545–602 (2017)
Ferré, G.: Large Deviations Theory in Statistical Physics: Some Theoretical and Numerical Aspects. Ph.D. thesis, Université Paris-Est (2019)
Ferré, G., Rousset, M., Stoltz, G.: More on the long time stability of Feynman–Kac semigroups arXiv:1807.00390 (2018)
Foulkes, W., Mitas, L., Needs, R., Rajagopal, G.: Quantum Monte Carlo simulations of solids. Rev. Mod. Phys. 73(1), 33 (2001)
Gärtner, J.: On large deviations from the invariant measure. Theory Probab. Appl. 22(1), 24–39 (1977)
Giardina, C., Kurchan, J., Peliti, L.: Direct evaluation of large deviation functions. Phys. Rev. Lett. 96(12), 120603 (2006)
Grimm, R., Storer, R.: Monte Carlo solution of Schrödinger’s equation. J. Comput. Phys. 7(1), 134–156 (1971)
Hairer, E., Lubich, C., Wanner, G.: Geometric Numerical Integration: Structure-Preserving Algorithms for Ordinary Differential Equations. Volume 31 of Springer Series in Computational Mathematics. Springer, Berlin (2006)
Hairer, M., Mattingly, J.C.: Yet another look at Harris’ ergodic theorem for Markov chains. In: Seminar on Stochastic Analysis, Random Fields and Applications VI, pp. 109–117. Springer (2011)
Hairer, M., Weare, J.: Improved diffusion Monte Carlo. Commun. Pure Appl. Math. 67(12), 1995–2021 (2014)
Hairer, M., Weare, J.: The Brownian fan. Commun. Pure Appl. Math. 68(1), 1–60 (2015)
Jarzynski, C.: Equilibrium free-energy differences from nonequilibrium measurements: a master-equation approach. Phys. Rev. E 56(5), 5018–5035 (1997)
Jarzynski, C.: Nonequilibrium equality for free energy differences. Phys. Rev. Lett. 78(14), 2690–2693 (1997)
Jasra, A.: On the behaviour of the backward interpretation of Feynman–Kac formulae under verifiable conditions. J. Appl. Probab. 52(02), 339–359 (2015)
Karatzas, I., Shreve, S.: Brownian Motion and Stochastic Calculus, Volume 113 of Graduate Texts in Mathematics. Springer, Berlin (2012)
Kopec, M.: Weak backward error analysis for overdamped Langevin processes. IMA J. Numer. Anal. 35(2), 583–614 (2014)
Kopec, M.: Weak backward error analysis for Langevin process. BIT Numer. Math. 55(4), 1057–1103 (2015)
Lecomte, V., Tailleur, J.: A numerical approach to large deviations in continuous time. J. Stat. Mech. Theory Exp. 2007(03), P03004 (2007)
Leimkuhler, B., Matthews, C., Stoltz, G.: The computation of averages from equilibrium and nonequilibrium Langevin molecular dynamics. IMA J. Numer. Anal. 36(1), 13–79 (2016)
Lelièvre, T., Stoltz, G.: Partial differential equations and stochastic methods in molecular dynamics. Acta Numer. 25, 681–880 (2016)
Lelièvre, T., Stoltz, G., Rousset, M.: Free Energy Computations: A Mathematical Perspective. World Scientific, Singapore (2010)
Lim, L.H., Weare, J.: Fast randomized iteration: diffusion Monte Carlo through the lens of numerical linear algebra. SIAM Rev. 59(3), 547–587 (2017)
Makri, N., Miller, W.H.: Exponential power series expansion for the quantum time evolution operator. J. Chem. Phys. 90(2), 904–911 (1989)
Mattingly, J.C., Stuart, A.M., Tretyakov, M.V.: Convergence of numerical time-averaging and stationary measures via Poisson equations. SIAM J. Numer. Anal. 48(2), 552–577 (2010)
Mella, M., Morosi, G., Bressanini, D.: Time step bias improvement in diffusion Monte Carlo simulations. Phys. Rev. E 61(2), 2050 (2000)
Milstein, G.N., Tretyakov, M.V.: Stochastic Numerics for Mathematical Physics. Scientific Computation. Springer, Berlin (2004)
Nemoto, T., Bouchet, F., Jack, R.L., Lecomte, V.: Population-dynamics method with a multicanonical feedback control. Phys. Rev. E 93(6), 062123 (2016)
Nemoto, T., Hidalgo, E.G., Lecomte, V.: Finite-time and finite-size scalings in the evaluation of large-deviation functions: analytical study using a birth–death process. Phys. Rev. E 95(1), 012102 (2017)
Pazy, A.: Semigroups of Linear Operators and Applications to Partial Differential Equations. Applied Mathematical Sciences, vol. 44. Springer, New York (1983)
Reed, M., Simon, B.: Modern Methods in Mathematical Physics. Analysis of Operators, vol. 4. Academic Press, New York (1978)
Rey-Bellet, L.: Ergodic properties of Markov processes. In: Attal, S., Joye, A., Pillet, C.A. (eds.) Open Quantum Systems II, pp. 1–39. Springer, Berlin, Heidelberg (2006)
Rousset, M.: Continuous Time Population Monte Carlo and Computational Physics. Ph.D. thesis, Universitié Paul Sabatier Toulouse (2006)
Rousset, M.: On the control of an interacting particle estimation of Schrödinger ground states. SIAM J. Math. Anal. 38(3), 824–844 (2006)
Rousset, M., Stoltz, G.: Equilibrium sampling from nonequilibrium dynamics. J. Stat. Phys. 123(6), 1251–1272 (2006)
Sarsa, A., Boronat, J., Casulleras, J.: Quadratic diffusion Monte Carlo and pure estimators for atoms. J. Chem. Phys. 116(14), 5956–5962 (2002)
Schaefer, H.H.: Banach Lattices and Positive Operators, volume 215 of Grundlehren der Mathematischen Wissenschaften. Springer (1974)
Tailleur, J., Lecomte, V.: Simulation of large deviation functions using population dynamics. AIP Conf. Proc. 1091(1), 212–219 (2009)
Talay, D.: Second-order discretization schemes of stochastic differential systems for the computation of the invariant law. Stochastics 29(1), 13–36 (1990)
Talay, D., Tubaro, L.: Expansion of the global error for numerical schemes solving stochastic differential equations. Stoch. Anal. Appl. 8(4), 483–509 (1990)
Touchette, H.: The large deviation approach to statistical mechanics. Phys. Rep. 478(1), 1–69 (2009)
Trstanova, Z.: Mathematical and Algorithmic Analysis of Modified Langevin Dynamics. Ph.D. thesis, Université Grenoble Alpes (2016)
Umrigar, C., Nightingale, M., Runge, K.: A diffusion Monte Carlo algorithm with very small time-step errors. J. Chem. Phys. 99(4), 2865–2890 (1993)
Varadhan, S.R.S.: Large Deviations and Applications. SIAM, Philadelphia (1984)
Zygalakis, K.: On the existence and the applications of modified equations for stochastic differential equations. SIAM J. Sci. Comput. 33(1), 102–130 (2011)
Acknowledgements
The authors are grateful to Mathias Rousset for his help in understanding Feynman–Kac models. We also thank Frédéric Cérou, Jonathan Mattingly, Julien Roussel, Hugo Touchette and Jonathan Weare for fruitful discussions, and the anonymous referees for their useful comments. The Ph.D. fellowship of Grégoire Ferré is partly funded by the Bézout Labex, funded by ANR, reference ANR-10-LABX-58. The work of Gabriel Stoltz was funded in part by the Agence Nationale de la Recherche, under Grant ANR-14-CE23-0012 (COSMOS) and by the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement No. 614492. We also benefited from the scientific environment of the Laboratoire International Associé between the Centre National de la Recherche Scientifique and the University of Illinois at Urbana-Champaign.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Markov contractions and Dobrushin coefficients
Appendix A: Markov contractions and Dobrushin coefficients
Denoting by \(\mathcal {M}(\mathcal {D})\) is the set of measures over \(\mathcal {D}\), we define \(\mathcal {M}_0(\mathcal {D})=\{\eta \in \mathcal {M}(\mathcal {D})\, |\, \eta (\mathcal {D})=0\}\) the set of (unsigned) measures with zero mass. The contraction norm of a Markov operator \(Q:\mathcal {P}(\mathcal {D})\rightarrow \mathcal {P}(\mathcal {D})\) is
the second equality coming from the fact that all elements in \(\mathcal {M}_0(\mathcal {D})\) are proportional to the difference of two probability measures. In particular,
A fundamental tool [9,10,11] for the study of Feynman–Kac type semigroups (15) and introduced by Dobrushin [15, 16] is the so-called Dobrushin ergodic coefficient, which can be defined for a Markov operator Q as:
where the infimum in the last equality runs over points \(q,q'\in \mathcal {D}\) and all partitions \((A_i)_{i=1}^m\) of \(\mathcal {D}\). If we interpret \(Q(q,A_i)\) as the probability of going from q into the set \(A_i\), we see that this coefficient provides information on the mixing properties of the operator Q. The link between this coefficient and the contraction properties of Q is made precise by the following relationship [15, 16]:
As a result, a minorization condition on Q translates into a contraction of the operator through its ergodic coefficient \(\alpha (Q)\). Relation (89) is essentially obtained by a Hahn decomposition of measures of zero mass, as made precise in [15, 16].
Rights and permissions
About this article
Cite this article
Ferré, G., Stoltz, G. Error estimates on ergodic properties of discretized Feynman–Kac semigroups. Numer. Math. 143, 261–313 (2019). https://doi.org/10.1007/s00211-019-01059-1
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00211-019-01059-1